Coupled dimensionality reduction and classification for supervised and semi-supervised multilabel learning.

PubMed

Gönen, Mehmet

2014-03-01

Coupled training of dimensionality reduction and classification is proposed previously to improve the prediction performance for single-label problems. Following this line of research, in this paper, we first introduce a novel Bayesian method that combines linear dimensionality reduction with linear binary classification for supervised multilabel learning and present a deterministic variational approximation algorithm to learn the proposed probabilistic model. We then extend the proposed method to find intrinsic dimensionality of the projected subspace using automatic relevance determination and to handle semi-supervised learning using a low-density assumption. We perform supervised learning experiments on four benchmark multilabel learning data sets by comparing our method with baseline linear dimensionality reduction algorithms. These experiments show that the proposed approach achieves good performance values in terms of hamming loss, average AUC, macro F 1 , and micro F 1 on held-out test data. The low-dimensional embeddings obtained by our method are also very useful for exploratory data analysis. We also show the effectiveness of our approach in finding intrinsic subspace dimensionality and semi-supervised learning tasks.

Improving M-SBL for Joint Sparse Recovery Using a Subspace Penalty

NASA Astrophysics Data System (ADS)

Ye, Jong Chul; Kim, Jong Min; Bresler, Yoram

2015-12-01

The multiple measurement vector problem (MMV) is a generalization of the compressed sensing problem that addresses the recovery of a set of jointly sparse signal vectors. One of the important contributions of this paper is to reveal that the seemingly least related state-of-art MMV joint sparse recovery algorithms - M-SBL (multiple sparse Bayesian learning) and subspace-based hybrid greedy algorithms - have a very important link. More specifically, we show that replacing the $\\log\\det(\\cdot)$ term in M-SBL by a rank proxy that exploits the spark reduction property discovered in subspace-based joint sparse recovery algorithms, provides significant improvements. In particular, if we use the Schatten-$p$ quasi-norm as the corresponding rank proxy, the global minimiser of the proposed algorithm becomes identical to the true solution as $p \\rightarrow 0$. Furthermore, under the same regularity conditions, we show that the convergence to a local minimiser is guaranteed using an alternating minimization algorithm that has closed form expressions for each of the minimization steps, which are convex. Numerical simulations under a variety of scenarios in terms of SNR, and condition number of the signal amplitude matrix demonstrate that the proposed algorithm consistently outperforms M-SBL and other state-of-the art algorithms.

Similarity preserving low-rank representation for enhanced data representation and effective subspace learning.

PubMed

Zhang, Zhao; Yan, Shuicheng; Zhao, Mingbo

2014-05-01

Latent Low-Rank Representation (LatLRR) delivers robust and promising results for subspace recovery and feature extraction through mining the so-called hidden effects, but the locality of both similar principal and salient features cannot be preserved in the optimizations. To solve this issue for achieving enhanced performance, a boosted version of LatLRR, referred to as Regularized Low-Rank Representation (rLRR), is proposed through explicitly including an appropriate Laplacian regularization that can maximally preserve the similarity among local features. Resembling LatLRR, rLRR decomposes given data matrix from two directions by seeking a pair of low-rank matrices. But the similarities of principal and salient features can be effectively preserved by rLRR. As a result, the correlated features are well grouped and the robustness of representations is also enhanced. Based on the outputted bi-directional low-rank codes by rLRR, an unsupervised subspace learning framework termed Low-rank Similarity Preserving Projections (LSPP) is also derived for feature learning. The supervised extension of LSPP is also discussed for discriminant subspace learning. The validity of rLRR is examined by robust representation and decomposition of real images. Results demonstrated the superiority of our rLRR and LSPP in comparison to other related state-of-the-art algorithms. Copyright © 2014 Elsevier Ltd. All rights reserved.

Dual signal subspace projection (DSSP): a novel algorithm for removing large interference in biomagnetic measurements

NASA Astrophysics Data System (ADS)

Sekihara, Kensuke; Kawabata, Yuya; Ushio, Shuta; Sumiya, Satoshi; Kawabata, Shigenori; Adachi, Yoshiaki; Nagarajan, Srikantan S.

2016-06-01

Objective. In functional electrophysiological imaging, signals are often contaminated by interference that can be of considerable magnitude compared to the signals of interest. This paper proposes a novel algorithm for removing such interferences that does not require separate noise measurements. Approach. The algorithm is based on a dual definition of the signal subspace in the spatial- and time-domains. Since the algorithm makes use of this duality, it is named the dual signal subspace projection (DSSP). The DSSP algorithm first projects the columns of the measured data matrix onto the inside and outside of the spatial-domain signal subspace, creating a set of two preprocessed data matrices. The intersection of the row spans of these two matrices is estimated as the time-domain interference subspace. The original data matrix is projected onto the subspace that is orthogonal to this interference subspace. Main results. The DSSP algorithm is validated by using the computer simulation, and using two sets of real biomagnetic data: spinal cord evoked field data measured from a healthy volunteer and magnetoencephalography data from a patient with a vagus nerve stimulator. Significance. The proposed DSSP algorithm is effective for removing overlapped interference in a wide variety of biomagnetic measurements.

BI-sparsity pursuit for robust subspace recovery

DOE PAGES

Bian, Xiao; Krim, Hamid

2015-09-01

Here, the success of sparse models in computer vision and machine learning in many real-world applications, may be attributed in large part, to the fact that many high dimensional data are distributed in a union of low dimensional subspaces. The underlying structure may, however, be adversely affected by sparse errors, thus inducing additional complexity in recovering it. In this paper, we propose a bi-sparse model as a framework to investigate and analyze this problem, and provide as a result , a novel algorithm to recover the union of subspaces in presence of sparse corruptions. We additionally demonstrate the effectiveness ofmore » our method by experiments on real-world vision data.« less

Structured Kernel Subspace Learning for Autonomous Robot Navigation.

PubMed

Kim, Eunwoo; Choi, Sungjoon; Oh, Songhwai

2018-02-14

This paper considers two important problems for autonomous robot navigation in a dynamic environment, where the goal is to predict pedestrian motion and control a robot with the prediction for safe navigation. While there are several methods for predicting the motion of a pedestrian and controlling a robot to avoid incoming pedestrians, it is still difficult to safely navigate in a dynamic environment due to challenges, such as the varying quality and complexity of training data with unwanted noises. This paper addresses these challenges simultaneously by proposing a robust kernel subspace learning algorithm based on the recent advances in nuclear-norm and l 1 -norm minimization. We model the motion of a pedestrian and the robot controller using Gaussian processes. The proposed method efficiently approximates a kernel matrix used in Gaussian process regression by learning low-rank structured matrix (with symmetric positive semi-definiteness) to find an orthogonal basis, which eliminates the effects of erroneous and inconsistent data. Based on structured kernel subspace learning, we propose a robust motion model and motion controller for safe navigation in dynamic environments. We evaluate the proposed robust kernel learning in various tasks, including regression, motion prediction, and motion control problems, and demonstrate that the proposed learning-based systems are robust against outliers and outperform existing regression and navigation methods.

Reweighted mass center based object-oriented sparse subspace clustering for hyperspectral images

NASA Astrophysics Data System (ADS)

Zhai, Han; Zhang, Hongyan; Zhang, Liangpei; Li, Pingxiang

2016-10-01

Considering the inevitable obstacles faced by the pixel-based clustering methods, such as salt-and-pepper noise, high computational complexity, and the lack of spatial information, a reweighted mass center based object-oriented sparse subspace clustering (RMC-OOSSC) algorithm for hyperspectral images (HSIs) is proposed. First, the mean-shift segmentation method is utilized to oversegment the HSI to obtain meaningful objects. Second, a distance reweighted mass center learning model is presented to extract the representative and discriminative features for each object. Third, assuming that all the objects are sampled from a union of subspaces, it is natural to apply the SSC algorithm to the HSI. Faced with the high correlation among the hyperspectral objects, a weighting scheme is adopted to ensure that the highly correlated objects are preferred in the procedure of sparse representation, to reduce the representation errors. Two widely used hyperspectral datasets were utilized to test the performance of the proposed RMC-OOSSC algorithm, obtaining high clustering accuracies (overall accuracy) of 71.98% and 89.57%, respectively. The experimental results show that the proposed method clearly improves the clustering performance with respect to the other state-of-the-art clustering methods, and it significantly reduces the computational time.

A hyperspectral imagery anomaly detection algorithm based on local three-dimensional orthogonal subspace projection

NASA Astrophysics Data System (ADS)

Zhang, Xing; Wen, Gongjian

2015-10-01

Anomaly detection (AD) becomes increasingly important in hyperspectral imagery analysis with many practical applications. Local orthogonal subspace projection (LOSP) detector is a popular anomaly detector which exploits local endmembers/eigenvectors around the pixel under test (PUT) to construct background subspace. However, this subspace only takes advantage of the spectral information, but the spatial correlat ion of the background clutter is neglected, which leads to the anomaly detection result sensitive to the accuracy of the estimated subspace. In this paper, a local three dimensional orthogonal subspace projection (3D-LOSP) algorithm is proposed. Firstly, under the jointly use of both spectral and spatial information, three directional background subspaces are created along the image height direction, the image width direction and the spectral direction, respectively. Then, the three corresponding orthogonal subspaces are calculated. After that, each vector along three direction of the local cube is projected onto the corresponding orthogonal subspace. Finally, a composite score is given through the three direction operators. In 3D-LOSP, the anomalies are redefined as the target not only spectrally different to the background, but also spatially distinct. Thanks to the addition of the spatial information, the robustness of the anomaly detection result has been improved greatly by the proposed 3D-LOSP algorithm. It is noteworthy that the proposed algorithm is an expansion of LOSP and this ideology can inspire many other spectral-based anomaly detection methods. Experiments with real hyperspectral images have proved the stability of the detection result.

Identifying Optimal Measurement Subspace for the Ensemble Kalman Filter

DOE Office of Scientific and Technical Information (OSTI.GOV)

Zhou, Ning; Huang, Zhenyu; Welch, Greg

2012-05-24

To reduce the computational load of the ensemble Kalman filter while maintaining its efficacy, an optimization algorithm based on the generalized eigenvalue decomposition method is proposed for identifying the most informative measurement subspace. When the number of measurements is large, the proposed algorithm can be used to make an effective tradeoff between computational complexity and estimation accuracy. This algorithm also can be extended to other Kalman filters for measurement subspace selection.

Dimension Reduction With Extreme Learning Machine.

PubMed

Kasun, Liyanaarachchi Lekamalage Chamara; Yang, Yan; Huang, Guang-Bin; Zhang, Zhengyou

2016-08-01

Data may often contain noise or irrelevant information, which negatively affect the generalization capability of machine learning algorithms. The objective of dimension reduction algorithms, such as principal component analysis (PCA), non-negative matrix factorization (NMF), random projection (RP), and auto-encoder (AE), is to reduce the noise or irrelevant information of the data. The features of PCA (eigenvectors) and linear AE are not able to represent data as parts (e.g. nose in a face image). On the other hand, NMF and non-linear AE are maimed by slow learning speed and RP only represents a subspace of original data. This paper introduces a dimension reduction framework which to some extend represents data as parts, has fast learning speed, and learns the between-class scatter subspace. To this end, this paper investigates a linear and non-linear dimension reduction framework referred to as extreme learning machine AE (ELM-AE) and sparse ELM-AE (SELM-AE). In contrast to tied weight AE, the hidden neurons in ELM-AE and SELM-AE need not be tuned, and their parameters (e.g, input weights in additive neurons) are initialized using orthogonal and sparse random weights, respectively. Experimental results on USPS handwritten digit recognition data set, CIFAR-10 object recognition, and NORB object recognition data set show the efficacy of linear and non-linear ELM-AE and SELM-AE in terms of discriminative capability, sparsity, training time, and normalized mean square error.

Linear Subspace Ranking Hashing for Cross-Modal Retrieval.

PubMed

Li, Kai; Qi, Guo-Jun; Ye, Jun; Hua, Kien A

2017-09-01

Hashing has attracted a great deal of research in recent years due to its effectiveness for the retrieval and indexing of large-scale high-dimensional multimedia data. In this paper, we propose a novel ranking-based hashing framework that maps data from different modalities into a common Hamming space where the cross-modal similarity can be measured using Hamming distance. Unlike existing cross-modal hashing algorithms where the learned hash functions are binary space partitioning functions, such as the sign and threshold function, the proposed hashing scheme takes advantage of a new class of hash functions closely related to rank correlation measures which are known to be scale-invariant, numerically stable, and highly nonlinear. Specifically, we jointly learn two groups of linear subspaces, one for each modality, so that features' ranking orders in different linear subspaces maximally preserve the cross-modal similarities. We show that the ranking-based hash function has a natural probabilistic approximation which transforms the original highly discontinuous optimization problem into one that can be efficiently solved using simple gradient descent algorithms. The proposed hashing framework is also flexible in the sense that the optimization procedures are not tied up to any specific form of loss function, which is typical for existing cross-modal hashing methods, but rather we can flexibly accommodate different loss functions with minimal changes to the learning steps. We demonstrate through extensive experiments on four widely-used real-world multimodal datasets that the proposed cross-modal hashing method can achieve competitive performance against several state-of-the-arts with only moderate training and testing time.

CLAss-Specific Subspace Kernel Representations and Adaptive Margin Slack Minimization for Large Scale Classification.

PubMed

Yu, Yinan; Diamantaras, Konstantinos I; McKelvey, Tomas; Kung, Sun-Yuan

2018-02-01

In kernel-based classification models, given limited computational power and storage capacity, operations over the full kernel matrix becomes prohibitive. In this paper, we propose a new supervised learning framework using kernel models for sequential data processing. The framework is based on two components that both aim at enhancing the classification capability with a subset selection scheme. The first part is a subspace projection technique in the reproducing kernel Hilbert space using a CLAss-specific Subspace Kernel representation for kernel approximation. In the second part, we propose a novel structural risk minimization algorithm called the adaptive margin slack minimization to iteratively improve the classification accuracy by an adaptive data selection. We motivate each part separately, and then integrate them into learning frameworks for large scale data. We propose two such frameworks: the memory efficient sequential processing for sequential data processing and the parallelized sequential processing for distributed computing with sequential data acquisition. We test our methods on several benchmark data sets and compared with the state-of-the-art techniques to verify the validity of the proposed techniques.

Mining subspace clusters from DNA microarray data using large itemset techniques.

PubMed

Chang, Ye-In; Chen, Jiun-Rung; Tsai, Yueh-Chi

2009-05-01

Mining subspace clusters from the DNA microarrays could help researchers identify those genes which commonly contribute to a disease, where a subspace cluster indicates a subset of genes whose expression levels are similar under a subset of conditions. Since in a DNA microarray, the number of genes is far larger than the number of conditions, those previous proposed algorithms which compute the maximum dimension sets (MDSs) for any two genes will take a long time to mine subspace clusters. In this article, we propose the Large Itemset-Based Clustering (LISC) algorithm for mining subspace clusters. Instead of constructing MDSs for any two genes, we construct only MDSs for any two conditions. Then, we transform the task of finding the maximal possible gene sets into the problem of mining large itemsets from the condition-pair MDSs. Since we are only interested in those subspace clusters with gene sets as large as possible, it is desirable to pay attention to those gene sets which have reasonable large support values in the condition-pair MDSs. From our simulation results, we show that the proposed algorithm needs shorter processing time than those previous proposed algorithms which need to construct gene-pair MDSs.

On iterative processes in the Krylov-Sonneveld subspaces

NASA Astrophysics Data System (ADS)

Ilin, Valery P.

2016-10-01

The iterative Induced Dimension Reduction (IDR) methods are considered for solving large systems of linear algebraic equations (SLAEs) with nonsingular nonsymmetric matrices. These approaches are investigated by many authors and are charachterized sometimes as the alternative to the classical processes of Krylov type. The key moments of the IDR algorithms consist in the construction of the embedded Sonneveld subspaces, which have the decreasing dimensions and use the orthogonalization to some fixed subspace. Other independent approaches for research and optimization of the iterations are based on the augmented and modified Krylov subspaces by using the aggregation and deflation procedures with present various low rank approximations of the original matrices. The goal of this paper is to show, that IDR method in Sonneveld subspaces present an original interpretation of the modified algorithms in the Krylov subspaces. In particular, such description is given for the multi-preconditioned semi-conjugate direction methods which are actual for the parallel algebraic domain decomposition approaches.

Conjunctive patches subspace learning with side information for collaborative image retrieval.

PubMed

Zhang, Lining; Wang, Lipo; Lin, Weisi

2012-08-01

Content-Based Image Retrieval (CBIR) has attracted substantial attention during the past few years for its potential practical applications to image management. A variety of Relevance Feedback (RF) schemes have been designed to bridge the semantic gap between the low-level visual features and the high-level semantic concepts for an image retrieval task. Various Collaborative Image Retrieval (CIR) schemes aim to utilize the user historical feedback log data with similar and dissimilar pairwise constraints to improve the performance of a CBIR system. However, existing subspace learning approaches with explicit label information cannot be applied for a CIR task, although the subspace learning techniques play a key role in various computer vision tasks, e.g., face recognition and image classification. In this paper, we propose a novel subspace learning framework, i.e., Conjunctive Patches Subspace Learning (CPSL) with side information, for learning an effective semantic subspace by exploiting the user historical feedback log data for a CIR task. The CPSL can effectively integrate the discriminative information of labeled log images, the geometrical information of labeled log images and the weakly similar information of unlabeled images together to learn a reliable subspace. We formally formulate this problem into a constrained optimization problem and then present a new subspace learning technique to exploit the user historical feedback log data. Extensive experiments on both synthetic data sets and a real-world image database demonstrate the effectiveness of the proposed scheme in improving the performance of a CBIR system by exploiting the user historical feedback log data.

New Parallel Algorithms for Structural Analysis and Design of Aerospace Structures

NASA Technical Reports Server (NTRS)

Nguyen, Duc T.

1998-01-01

Subspace and Lanczos iterations have been developed, well documented, and widely accepted as efficient methods for obtaining p-lowest eigen-pair solutions of large-scale, practical engineering problems. The focus of this paper is to incorporate recent developments in vectorized sparse technologies in conjunction with Subspace and Lanczos iterative algorithms for computational enhancements. Numerical performance, in terms of accuracy and efficiency of the proposed sparse strategies for Subspace and Lanczos algorithm, is demonstrated by solving for the lowest frequencies and mode shapes of structural problems on the IBM-R6000/590 and SunSparc 20 workstations.

Multiple Kernel Sparse Representation based Orthogonal Discriminative Projection and Its Cost-Sensitive Extension.

PubMed

Zhang, Guoqing; Sun, Huaijiang; Xia, Guiyu; Sun, Quansen

2016-07-07

Sparse representation based classification (SRC) has been developed and shown great potential for real-world application. Based on SRC, Yang et al. [10] devised a SRC steered discriminative projection (SRC-DP) method. However, as a linear algorithm, SRC-DP cannot handle the data with highly nonlinear distribution. Kernel sparse representation-based classifier (KSRC) is a non-linear extension of SRC and can remedy the drawback of SRC. KSRC requires the use of a predetermined kernel function and selection of the kernel function and its parameters is difficult. Recently, multiple kernel learning for SRC (MKL-SRC) [22] has been proposed to learn a kernel from a set of base kernels. However, MKL-SRC only considers the within-class reconstruction residual while ignoring the between-class relationship, when learning the kernel weights. In this paper, we propose a novel multiple kernel sparse representation-based classifier (MKSRC), and then we use it as a criterion to design a multiple kernel sparse representation based orthogonal discriminative projection method (MK-SR-ODP). The proposed algorithm aims at learning a projection matrix and a corresponding kernel from the given base kernels such that in the low dimension subspace the between-class reconstruction residual is maximized and the within-class reconstruction residual is minimized. Furthermore, to achieve a minimum overall loss by performing recognition in the learned low-dimensional subspace, we introduce cost information into the dimensionality reduction method. The solutions for the proposed method can be efficiently found based on trace ratio optimization method [33]. Extensive experimental results demonstrate the superiority of the proposed algorithm when compared with the state-of-the-art methods.

Multiple Source DF (Direction Finding) Signal Processing: An Experimental System,

DTIC Science & Technology

The MUltiple SIgnal Characterization ( MUSIC ) algorithm is an implementation of the Signal Subspace Approach to provide parameter estimates of...the signal subspace (obtained from the received data) and the array manifold (obtained via array calibration). The MUSIC algorithm has been

Subspace-based interference removal methods for a multichannel biomagnetic sensor array.

PubMed

Sekihara, Kensuke; Nagarajan, Srikantan S

2017-10-01

In biomagnetic signal processing, the theory of the signal subspace has been applied to removing interfering magnetic fields, and a representative algorithm is the signal space projection algorithm, in which the signal/interference subspace is defined in the spatial domain as the span of signal/interference-source lead field vectors. This paper extends the notion of this conventional (spatial domain) signal subspace by introducing a new definition of signal subspace in the time domain. It defines the time-domain signal subspace as the span of row vectors that contain the source time course values. This definition leads to symmetric relationships between the time-domain and the conventional (spatial-domain) signal subspaces. As a review, this article shows that the notion of the time-domain signal subspace provides useful insights over existing interference removal methods from a unified perspective. Main results and significance. Using the time-domain signal subspace, it is possible to interpret a number of interference removal methods as the time domain signal space projection. Such methods include adaptive noise canceling, sensor noise suppression, the common temporal subspace projection, the spatio-temporal signal space separation, and the recently-proposed dual signal subspace projection. Our analysis using the notion of the time domain signal space projection reveals implicit assumptions these methods rely on, and shows that the difference between these methods results only from the manner of deriving the interference subspace. Numerical examples that illustrate the results of our arguments are provided.

Subspace-based interference removal methods for a multichannel biomagnetic sensor array

NASA Astrophysics Data System (ADS)

Sekihara, Kensuke; Nagarajan, Srikantan S.

2017-10-01

Objective. In biomagnetic signal processing, the theory of the signal subspace has been applied to removing interfering magnetic fields, and a representative algorithm is the signal space projection algorithm, in which the signal/interference subspace is defined in the spatial domain as the span of signal/interference-source lead field vectors. This paper extends the notion of this conventional (spatial domain) signal subspace by introducing a new definition of signal subspace in the time domain. Approach. It defines the time-domain signal subspace as the span of row vectors that contain the source time course values. This definition leads to symmetric relationships between the time-domain and the conventional (spatial-domain) signal subspaces. As a review, this article shows that the notion of the time-domain signal subspace provides useful insights over existing interference removal methods from a unified perspective. Main results and significance. Using the time-domain signal subspace, it is possible to interpret a number of interference removal methods as the time domain signal space projection. Such methods include adaptive noise canceling, sensor noise suppression, the common temporal subspace projection, the spatio-temporal signal space separation, and the recently-proposed dual signal subspace projection. Our analysis using the notion of the time domain signal space projection reveals implicit assumptions these methods rely on, and shows that the difference between these methods results only from the manner of deriving the interference subspace. Numerical examples that illustrate the results of our arguments are provided.

Low-rank plus sparse decomposition for exoplanet detection in direct-imaging ADI sequences. The LLSG algorithm

NASA Astrophysics Data System (ADS)

Gomez Gonzalez, C. A.; Absil, O.; Absil, P.-A.; Van Droogenbroeck, M.; Mawet, D.; Surdej, J.

2016-05-01

Context. Data processing constitutes a critical component of high-contrast exoplanet imaging. Its role is almost as important as the choice of a coronagraph or a wavefront control system, and it is intertwined with the chosen observing strategy. Among the data processing techniques for angular differential imaging (ADI), the most recent is the family of principal component analysis (PCA) based algorithms. It is a widely used statistical tool developed during the first half of the past century. PCA serves, in this case, as a subspace projection technique for constructing a reference point spread function (PSF) that can be subtracted from the science data for boosting the detectability of potential companions present in the data. Unfortunately, when building this reference PSF from the science data itself, PCA comes with certain limitations such as the sensitivity of the lower dimensional orthogonal subspace to non-Gaussian noise. Aims: Inspired by recent advances in machine learning algorithms such as robust PCA, we aim to propose a localized subspace projection technique that surpasses current PCA-based post-processing algorithms in terms of the detectability of companions at near real-time speed, a quality that will be useful for future direct imaging surveys. Methods: We used randomized low-rank approximation methods recently proposed in the machine learning literature, coupled with entry-wise thresholding to decompose an ADI image sequence locally into low-rank, sparse, and Gaussian noise components (LLSG). This local three-term decomposition separates the starlight and the associated speckle noise from the planetary signal, which mostly remains in the sparse term. We tested the performance of our new algorithm on a long ADI sequence obtained on β Pictoris with VLT/NACO. Results: Compared to a standard PCA approach, LLSG decomposition reaches a higher signal-to-noise ratio and has an overall better performance in the receiver operating characteristic space. This three-term decomposition brings a detectability boost compared to the full-frame standard PCA approach, especially in the small inner working angle region where complex speckle noise prevents PCA from discerning true companions from noise.

Sparse subspace clustering for data with missing entries and high-rank matrix completion.

PubMed

Fan, Jicong; Chow, Tommy W S

2017-09-01

Many methods have recently been proposed for subspace clustering, but they are often unable to handle incomplete data because of missing entries. Using matrix completion methods to recover missing entries is a common way to solve the problem. Conventional matrix completion methods require that the matrix should be of low-rank intrinsically, but most matrices are of high-rank or even full-rank in practice, especially when the number of subspaces is large. In this paper, a new method called Sparse Representation with Missing Entries and Matrix Completion is proposed to solve the problems of incomplete-data subspace clustering and high-rank matrix completion. The proposed algorithm alternately computes the matrix of sparse representation coefficients and recovers the missing entries of a data matrix. The proposed algorithm recovers missing entries through minimizing the representation coefficients, representation errors, and matrix rank. Thorough experimental study and comparative analysis based on synthetic data and natural images were conducted. The presented results demonstrate that the proposed algorithm is more effective in subspace clustering and matrix completion compared with other existing methods. Copyright © 2017 Elsevier Ltd. All rights reserved.

Adaptive low-rank subspace learning with online optimization for robust visual tracking.

PubMed

Liu, Risheng; Wang, Di; Han, Yuzhuo; Fan, Xin; Luo, Zhongxuan

2017-04-01

In recent years, sparse and low-rank models have been widely used to formulate appearance subspace for visual tracking. However, most existing methods only consider the sparsity or low-rankness of the coefficients, which is not sufficient enough for appearance subspace learning on complex video sequences. Moreover, as both the low-rank and the column sparse measures are tightly related to all the samples in the sequences, it is challenging to incrementally solve optimization problems with both nuclear norm and column sparse norm on sequentially obtained video data. To address above limitations, this paper develops a novel low-rank subspace learning with adaptive penalization (LSAP) framework for subspace based robust visual tracking. Different from previous work, which often simply decomposes observations as low-rank features and sparse errors, LSAP simultaneously learns the subspace basis, low-rank coefficients and column sparse errors to formulate appearance subspace. Within LSAP framework, we introduce a Hadamard production based regularization to incorporate rich generative/discriminative structure constraints to adaptively penalize the coefficients for subspace learning. It is shown that such adaptive penalization can significantly improve the robustness of LSAP on severely corrupted dataset. To utilize LSAP for online visual tracking, we also develop an efficient incremental optimization scheme for nuclear norm and column sparse norm minimizations. Experiments on 50 challenging video sequences demonstrate that our tracker outperforms other state-of-the-art methods. Copyright © 2017 Elsevier Ltd. All rights reserved.

EEG and MEG source localization using recursively applied (RAP) MUSIC

DOE Office of Scientific and Technical Information (OSTI.GOV)

Mosher, J.C.; Leahy, R.M.

1996-12-31

The multiple signal characterization (MUSIC) algorithm locates multiple asynchronous dipolar sources from electroencephalography (EEG) and magnetoencephalography (MEG) data. A signal subspace is estimated from the data, then the algorithm scans a single dipole model through a three-dimensional head volume and computes projections onto this subspace. To locate the sources, the user must search the head volume for local peaks in the projection metric. Here we describe a novel extension of this approach which we refer to as RAP (Recursively APplied) MUSIC. This new procedure automatically extracts the locations of the sources through a recursive use of subspace projections, which usesmore » the metric of principal correlations as a multidimensional form of correlation analysis between the model subspace and the data subspace. The dipolar orientations, a form of `diverse polarization,` are easily extracted using the associated principal vectors.« less

Limited Memory Block Krylov Subspace Optimization for Computing Dominant Singular Value Decompositions

DTIC Science & Technology

2012-03-22

with performance profiles, Math. Program., 91 (2002), pp. 201–213. [6] P. DRINEAS, R. KANNAN, AND M. W. MAHONEY , Fast Monte Carlo algorithms for matrices...computing invariant subspaces of non-Hermitian matri- ces, Numer. Math., 25 ( 1975 /76), pp. 123–136. [25] , Matrix algorithms Vol. II: Eigensystems

Joint fMRI analysis and subject clustering using sparse dictionary learning

NASA Astrophysics Data System (ADS)

Kim, Seung-Jun; Dontaraju, Krishna K.

2017-08-01

Multi-subject fMRI data analysis methods based on sparse dictionary learning are proposed. In addition to identifying the component spatial maps by exploiting the sparsity of the maps, clusters of the subjects are learned by postulating that the fMRI volumes admit a subspace clustering structure. Furthermore, in order to tune the associated hyper-parameters systematically, a cross-validation strategy is developed based on entry-wise sampling of the fMRI dataset. Efficient algorithms for solving the proposed constrained dictionary learning formulations are developed. Numerical tests performed on synthetic fMRI data show promising results and provides insights into the proposed technique.

Gene selection for microarray data classification via subspace learning and manifold regularization.

PubMed

Tang, Chang; Cao, Lijuan; Zheng, Xiao; Wang, Minhui

2017-12-19

With the rapid development of DNA microarray technology, large amount of genomic data has been generated. Classification of these microarray data is a challenge task since gene expression data are often with thousands of genes but a small number of samples. In this paper, an effective gene selection method is proposed to select the best subset of genes for microarray data with the irrelevant and redundant genes removed. Compared with original data, the selected gene subset can benefit the classification task. We formulate the gene selection task as a manifold regularized subspace learning problem. In detail, a projection matrix is used to project the original high dimensional microarray data into a lower dimensional subspace, with the constraint that the original genes can be well represented by the selected genes. Meanwhile, the local manifold structure of original data is preserved by a Laplacian graph regularization term on the low-dimensional data space. The projection matrix can serve as an importance indicator of different genes. An iterative update algorithm is developed for solving the problem. Experimental results on six publicly available microarray datasets and one clinical dataset demonstrate that the proposed method performs better when compared with other state-of-the-art methods in terms of microarray data classification. Graphical Abstract The graphical abstract of this work.

Bi Sparsity Pursuit: A Paradigm for Robust Subspace Recovery

DTIC Science & Technology

2016-09-27

16. SECURITY CLASSIFICATION OF: The success of sparse models in computer vision and machine learning is due to the fact that, high dimensional data...Office P.O. Box 12211 Research Triangle Park, NC 27709-2211 Signal recovery, Sparse learning , Subspace modeling REPORT DOCUMENTATION PAGE 11...vision and machine learning is due to the fact that, high dimensional data is distributed in a union of low dimensional subspaces in many real-world

Multiple-instance ensemble learning for hyperspectral images

NASA Astrophysics Data System (ADS)

Ergul, Ugur; Bilgin, Gokhan

2017-10-01

An ensemble framework for multiple-instance (MI) learning (MIL) is introduced for use in hyperspectral images (HSIs) by inspiring the bagging (bootstrap aggregation) method in ensemble learning. Ensemble-based bagging is performed by a small percentage of training samples, and MI bags are formed by a local windowing process with variable window sizes on selected instances. In addition to bootstrap aggregation, random subspace is another method used to diversify base classifiers. The proposed method is implemented using four MIL classification algorithms. The classifier model learning phase is carried out with MI bags, and the estimation phase is performed over single-test instances. In the experimental part of the study, two different HSIs that have ground-truth information are used, and comparative results are demonstrated with state-of-the-art classification methods. In general, the MI ensemble approach produces more compact results in terms of both diversity and error compared to equipollent non-MIL algorithms.

Application of higher order SVD to vibration-based system identification and damage detection

NASA Astrophysics Data System (ADS)

Chao, Shu-Hsien; Loh, Chin-Hsiung; Weng, Jian-Huang

2012-04-01

Singular value decomposition (SVD) is a powerful linear algebra tool. It is widely used in many different signal processing methods, such principal component analysis (PCA), singular spectrum analysis (SSA), frequency domain decomposition (FDD), subspace identification and stochastic subspace identification method ( SI and SSI ). In each case, the data is arranged appropriately in matrix form and SVD is used to extract the feature of the data set. In this study three different algorithms on signal processing and system identification are proposed: SSA, SSI-COV and SSI-DATA. Based on the extracted subspace and null-space from SVD of data matrix, damage detection algorithms can be developed. The proposed algorithm is used to process the shaking table test data of the 6-story steel frame. Features contained in the vibration data are extracted by the proposed method. Damage detection can then be investigated from the test data of the frame structure through subspace-based and nullspace-based damage indices.

Subarray Processing for Projection-based RFI Mitigation in Radio Astronomical Interferometers

NASA Astrophysics Data System (ADS)

Burnett, Mitchell C.; Jeffs, Brian D.; Black, Richard A.; Warnick, Karl F.

2018-04-01

Radio Frequency Interference (RFI) is a major problem for observations in Radio Astronomy (RA). Adaptive spatial filtering techniques such as subspace projection are promising candidates for RFI mitigation; however, for radio interferometric imaging arrays, these have primarily been used in engineering demonstration experiments rather than mainstream scientific observations. This paper considers one reason that adoption of such algorithms is limited: RFI decorrelates across the interferometric array because of long baseline lengths. This occurs when the relative RFI time delay along a baseline is large compared to the frequency channel inverse bandwidth used in the processing chain. Maximum achievable excision of the RFI is limited by covariance matrix estimation error when identifying interference subspace parameters, and decorrelation of the RFI introduces errors that corrupt the subspace estimate, rendering subspace projection ineffective over the entire array. In this work, we present an algorithm that overcomes this challenge of decorrelation by applying subspace projection via subarray processing (SP-SAP). Each subarray is designed to have a set of elements with high mutual correlation in the interferer for better estimation of subspace parameters. In an RFI simulation scenario for the proposed ngVLA interferometric imaging array with 15 kHz channel bandwidth for correlator processing, we show that compared to the former approach of applying subspace projection on the full array, SP-SAP improves mitigation of the RFI on the order of 9 dB. An example of improved image synthesis and reduced RFI artifacts for a simulated image “phantom” using the SP-SAP algorithm is presented.

Mode Shape Estimation Algorithms Under Ambient Conditions: A Comparative Review

DOE Office of Scientific and Technical Information (OSTI.GOV)

Dosiek, Luke; Zhou, Ning; Pierre, John W.

Abstract—This paper provides a comparative review of five existing ambient electromechanical mode shape estimation algorithms, i.e., the Transfer Function (TF), Spectral, Frequency Domain Decomposition (FDD), Channel Matching, and Subspace Methods. It is also shown that the TF Method is a general approach to estimating mode shape and that the Spectral, FDD, and Channel Matching Methods are actually special cases of it. Additionally, some of the variations of the Subspace Method are reviewed and the Numerical algorithm for Subspace State Space System IDentification (N4SID) is implemented. The five algorithms are then compared using data simulated from a 17-machine model of themore » Western Electricity Coordinating Council (WECC) under ambient conditions with both low and high damping, as well as during the case where ambient data is disrupted by an oscillatory ringdown. The performance of the algorithms is compared using the statistics from Monte Carlo Simulations and results from measured WECC data, and a discussion of the practical issues surrounding their implementation, including cases where power system probing is an option, is provided. The paper concludes with some recommendations as to the appropriate use of the various techniques. Index Terms—Electromechanical mode shape, small-signal stability, phasor measurement units (PMU), system identification, N4SID, subspace.« less

Globally convergent techniques in nonlinear Newton-Krylov

NASA Technical Reports Server (NTRS)

Brown, Peter N.; Saad, Youcef

1989-01-01

Some convergence theory is presented for nonlinear Krylov subspace methods. The basic idea of these methods is to use variants of Newton's iteration in conjunction with a Krylov subspace method for solving the Jacobian linear systems. These methods are variants of inexact Newton methods where the approximate Newton direction is taken from a subspace of small dimensions. The main focus is to analyze these methods when they are combined with global strategies such as linesearch techniques and model trust region algorithms. Most of the convergence results are formulated for projection onto general subspaces rather than just Krylov subspaces.

Anomaly Detection in Moving-Camera Video Sequences Using Principal Subspace Analysis

DOE PAGES

Thomaz, Lucas A.; Jardim, Eric; da Silva, Allan F.; ...

2017-10-16

This study presents a family of algorithms based on sparse decompositions that detect anomalies in video sequences obtained from slow moving cameras. These algorithms start by computing the union of subspaces that best represents all the frames from a reference (anomaly free) video as a low-rank projection plus a sparse residue. Then, they perform a low-rank representation of a target (possibly anomalous) video by taking advantage of both the union of subspaces and the sparse residue computed from the reference video. Such algorithms provide good detection results while at the same time obviating the need for previous video synchronization. However,more » this is obtained at the cost of a large computational complexity, which hinders their applicability. Another contribution of this paper approaches this problem by using intrinsic properties of the obtained data representation in order to restrict the search space to the most relevant subspaces, providing computational complexity gains of up to two orders of magnitude. The developed algorithms are shown to cope well with videos acquired in challenging scenarios, as verified by the analysis of 59 videos from the VDAO database that comprises videos with abandoned objects in a cluttered industrial scenario.« less

Anomaly Detection in Moving-Camera Video Sequences Using Principal Subspace Analysis

DOE Office of Scientific and Technical Information (OSTI.GOV)

Thomaz, Lucas A.; Jardim, Eric; da Silva, Allan F.

This study presents a family of algorithms based on sparse decompositions that detect anomalies in video sequences obtained from slow moving cameras. These algorithms start by computing the union of subspaces that best represents all the frames from a reference (anomaly free) video as a low-rank projection plus a sparse residue. Then, they perform a low-rank representation of a target (possibly anomalous) video by taking advantage of both the union of subspaces and the sparse residue computed from the reference video. Such algorithms provide good detection results while at the same time obviating the need for previous video synchronization. However,more » this is obtained at the cost of a large computational complexity, which hinders their applicability. Another contribution of this paper approaches this problem by using intrinsic properties of the obtained data representation in order to restrict the search space to the most relevant subspaces, providing computational complexity gains of up to two orders of magnitude. The developed algorithms are shown to cope well with videos acquired in challenging scenarios, as verified by the analysis of 59 videos from the VDAO database that comprises videos with abandoned objects in a cluttered industrial scenario.« less

Adaptive bearing estimation and tracking of multiple targets in a realistic passive sonar scenario

NASA Astrophysics Data System (ADS)

Rajagopal, R.; Challa, Subhash; Faruqi, Farhan A.; Rao, P. R.

1997-06-01

In a realistic passive sonar environment, the received signal consists of multipath arrivals from closely separated moving targets. The signals are contaminated by spatially correlated noise. The differential MUSIC has been proposed to estimate the DOAs in such a scenario. This method estimates the 'noise subspace' in order to estimate the DOAs. However, the 'noise subspace' estimate has to be updated as and when new data become available. In order to save the computational costs, a new adaptive noise subspace estimation algorithm is proposed in this paper. The salient features of the proposed algorithm are: (1) Noise subspace estimation is done by QR decomposition of the difference matrix which is formed from the data covariance matrix. Thus, as compared to standard eigen-decomposition based methods which require O(N3) computations, the proposed method requires only O(N2) computations. (2) Noise subspace is updated by updating the QR decomposition. (3) The proposed algorithm works in a realistic sonar environment. In the second part of the paper, the estimated bearing values are used to track multiple targets. In order to achieve this, the nonlinear system/linear measurement extended Kalman filtering proposed is applied. Computer simulation results are also presented to support the theory.

A General Algorithm for Reusing Krylov Subspace Information. I. Unsteady Navier-Stokes

NASA Technical Reports Server (NTRS)

Carpenter, Mark H.; Vuik, C.; Lucas, Peter; vanGijzen, Martin; Bijl, Hester

2010-01-01

A general algorithm is developed that reuses available information to accelerate the iterative convergence of linear systems with multiple right-hand sides A x = b (sup i), which are commonly encountered in steady or unsteady simulations of nonlinear equations. The algorithm is based on the classical GMRES algorithm with eigenvector enrichment but also includes a Galerkin projection preprocessing step and several novel Krylov subspace reuse strategies. The new approach is applied to a set of test problems, including an unsteady turbulent airfoil, and is shown in some cases to provide significant improvement in computational efficiency relative to baseline approaches.

Subspace-Aware Index Codes

DOE Office of Scientific and Technical Information (OSTI.GOV)

Kailkhura, Bhavya; Theagarajan, Lakshmi Narasimhan; Varshney, Pramod K.

In this paper, we generalize the well-known index coding problem to exploit the structure in the source-data to improve system throughput. In many applications (e.g., multimedia), the data to be transmitted may lie (or can be well approximated) in a low-dimensional subspace. We exploit this low-dimensional structure of the data using an algebraic framework to solve the index coding problem (referred to as subspace-aware index coding) as opposed to the traditional index coding problem which is subspace-unaware. Also, we propose an efficient algorithm based on the alternating minimization approach to obtain near optimal index codes for both subspace-aware and -unawaremore » cases. In conclusion, our simulations indicate that under certain conditions, a significant throughput gain (about 90%) can be achieved by subspace-aware index codes over conventional subspace-unaware index codes.« less

Subspace-Aware Index Codes

DOE PAGES

Kailkhura, Bhavya; Theagarajan, Lakshmi Narasimhan; Varshney, Pramod K.

2017-04-12

In this paper, we generalize the well-known index coding problem to exploit the structure in the source-data to improve system throughput. In many applications (e.g., multimedia), the data to be transmitted may lie (or can be well approximated) in a low-dimensional subspace. We exploit this low-dimensional structure of the data using an algebraic framework to solve the index coding problem (referred to as subspace-aware index coding) as opposed to the traditional index coding problem which is subspace-unaware. Also, we propose an efficient algorithm based on the alternating minimization approach to obtain near optimal index codes for both subspace-aware and -unawaremore » cases. In conclusion, our simulations indicate that under certain conditions, a significant throughput gain (about 90%) can be achieved by subspace-aware index codes over conventional subspace-unaware index codes.« less

DOE Office of Scientific and Technical Information (OSTI.GOV)

Mosher, J.C.; Leahy, R.M.

A new method for source localization is described that is based on a modification of the well known multiple signal classification (MUSIC) algorithm. In classical MUSIC, the array manifold vector is projected onto an estimate of the signal subspace, but errors in the estimate can make location of multiple sources difficult. Recursively applied and projected (RAP) MUSIC uses each successively located source to form an intermediate array gain matrix, and projects both the array manifold and the signal subspace estimate into its orthogonal complement. The MUSIC projection is then performed in this reduced subspace. Using the metric of principal angles,more » the authors describe a general form of the RAP-MUSIC algorithm for the case of diversely polarized sources. Through a uniform linear array simulation, the authors demonstrate the improved Monte Carlo performance of RAP-MUSIC relative to MUSIC and two other sequential subspace methods, S and IES-MUSIC.« less

Fast Low-Rank Shared Dictionary Learning for Image Classification.

PubMed

Tiep Huu Vu; Monga, Vishal

2017-11-01

Despite the fact that different objects possess distinct class-specific features, they also usually share common patterns. This observation has been exploited partially in a recently proposed dictionary learning framework by separating the particularity and the commonality (COPAR). Inspired by this, we propose a novel method to explicitly and simultaneously learn a set of common patterns as well as class-specific features for classification with more intuitive constraints. Our dictionary learning framework is hence characterized by both a shared dictionary and particular (class-specific) dictionaries. For the shared dictionary, we enforce a low-rank constraint, i.e., claim that its spanning subspace should have low dimension and the coefficients corresponding to this dictionary should be similar. For the particular dictionaries, we impose on them the well-known constraints stated in the Fisher discrimination dictionary learning (FDDL). Furthermore, we develop new fast and accurate algorithms to solve the subproblems in the learning step, accelerating its convergence. The said algorithms could also be applied to FDDL and its extensions. The efficiencies of these algorithms are theoretically and experimentally verified by comparing their complexities and running time with those of other well-known dictionary learning methods. Experimental results on widely used image data sets establish the advantages of our method over the state-of-the-art dictionary learning methods.

Fast Low-Rank Shared Dictionary Learning for Image Classification

NASA Astrophysics Data System (ADS)

Vu, Tiep Huu; Monga, Vishal

2017-11-01

Despite the fact that different objects possess distinct class-specific features, they also usually share common patterns. This observation has been exploited partially in a recently proposed dictionary learning framework by separating the particularity and the commonality (COPAR). Inspired by this, we propose a novel method to explicitly and simultaneously learn a set of common patterns as well as class-specific features for classification with more intuitive constraints. Our dictionary learning framework is hence characterized by both a shared dictionary and particular (class-specific) dictionaries. For the shared dictionary, we enforce a low-rank constraint, i.e. claim that its spanning subspace should have low dimension and the coefficients corresponding to this dictionary should be similar. For the particular dictionaries, we impose on them the well-known constraints stated in the Fisher discrimination dictionary learning (FDDL). Further, we develop new fast and accurate algorithms to solve the subproblems in the learning step, accelerating its convergence. The said algorithms could also be applied to FDDL and its extensions. The efficiencies of these algorithms are theoretically and experimentally verified by comparing their complexities and running time with those of other well-known dictionary learning methods. Experimental results on widely used image datasets establish the advantages of our method over state-of-the-art dictionary learning methods.

ASCS online fault detection and isolation based on an improved MPCA

NASA Astrophysics Data System (ADS)

Peng, Jianxin; Liu, Haiou; Hu, Yuhui; Xi, Junqiang; Chen, Huiyan

2014-09-01

Multi-way principal component analysis (MPCA) has received considerable attention and been widely used in process monitoring. A traditional MPCA algorithm unfolds multiple batches of historical data into a two-dimensional matrix and cut the matrix along the time axis to form subspaces. However, low efficiency of subspaces and difficult fault isolation are the common disadvantages for the principal component model. This paper presents a new subspace construction method based on kernel density estimation function that can effectively reduce the storage amount of the subspace information. The MPCA model and the knowledge base are built based on the new subspace. Then, fault detection and isolation with the squared prediction error (SPE) statistic and the Hotelling ( T 2) statistic are also realized in process monitoring. When a fault occurs, fault isolation based on the SPE statistic is achieved by residual contribution analysis of different variables. For fault isolation of subspace based on the T 2 statistic, the relationship between the statistic indicator and state variables is constructed, and the constraint conditions are presented to check the validity of fault isolation. Then, to improve the robustness of fault isolation to unexpected disturbances, the statistic method is adopted to set the relation between single subspace and multiple subspaces to increase the corrective rate of fault isolation. Finally fault detection and isolation based on the improved MPCA is used to monitor the automatic shift control system (ASCS) to prove the correctness and effectiveness of the algorithm. The research proposes a new subspace construction method to reduce the required storage capacity and to prove the robustness of the principal component model, and sets the relationship between the state variables and fault detection indicators for fault isolation.

Effective learning strategies for real-time image-guided adaptive control of multiple-source hyperthermia applicators.

PubMed

Cheng, Kung-Shan; Dewhirst, Mark W; Stauffer, Paul R; Das, Shiva

2010-03-01

This paper investigates overall theoretical requirements for reducing the times required for the iterative learning of a real-time image-guided adaptive control routine for multiple-source heat applicators, as used in hyperthermia and thermal ablative therapy for cancer. Methods for partial reconstruction of the physical system with and without model reduction to find solutions within a clinically practical timeframe were analyzed. A mathematical analysis based on the Fredholm alternative theorem (FAT) was used to compactly analyze the existence and uniqueness of the optimal heating vector under two fundamental situations: (1) noiseless partial reconstruction and (2) noisy partial reconstruction. These results were coupled with a method for further acceleration of the solution using virtual source (VS) model reduction. The matrix approximation theorem (MAT) was used to choose the optimal vectors spanning the reduced-order subspace to reduce the time for system reconstruction and to determine the associated approximation error. Numerical simulations of the adaptive control of hyperthermia using VS were also performed to test the predictions derived from the theoretical analysis. A thigh sarcoma patient model surrounded by a ten-antenna phased-array applicator was retained for this purpose. The impacts of the convective cooling from blood flow and the presence of sudden increase of perfusion in muscle and tumor were also simulated. By FAT, partial system reconstruction directly conducted in the full space of the physical variables such as phases and magnitudes of the heat sources cannot guarantee reconstructing the optimal system to determine the global optimal setting of the heat sources. A remedy for this limitation is to conduct the partial reconstruction within a reduced-order subspace spanned by the first few maximum eigenvectors of the true system matrix. By MAT, this VS subspace is the optimal one when the goal is to maximize the average tumor temperature. When more than 6 sources present, the steps required for a nonlinear learning scheme is theoretically fewer than that of a linear one, however, finite number of iterative corrections is necessary for a single learning step of a nonlinear algorithm. Thus, the actual computational workload for a nonlinear algorithm is not necessarily less than that required by a linear algorithm. Based on the analysis presented herein, obtaining a unique global optimal heating vector for a multiple-source applicator within the constraints of real-time clinical hyperthermia treatments and thermal ablative therapies appears attainable using partial reconstruction with minimum norm least-squares method with supplemental equations. One way to supplement equations is the inclusion of a method of model reduction.

Principal component and clustering analysis on molecular dynamics data of the ribosomal L11·23S subdomain.

PubMed

Wolf, Antje; Kirschner, Karl N

2013-02-01

With improvements in computer speed and algorithm efficiency, MD simulations are sampling larger amounts of molecular and biomolecular conformations. Being able to qualitatively and quantitatively sift these conformations into meaningful groups is a difficult and important task, especially when considering the structure-activity paradigm. Here we present a study that combines two popular techniques, principal component (PC) analysis and clustering, for revealing major conformational changes that occur in molecular dynamics (MD) simulations. Specifically, we explored how clustering different PC subspaces effects the resulting clusters versus clustering the complete trajectory data. As a case example, we used the trajectory data from an explicitly solvated simulation of a bacteria's L11·23S ribosomal subdomain, which is a target of thiopeptide antibiotics. Clustering was performed, using K-means and average-linkage algorithms, on data involving the first two to the first five PC subspace dimensions. For the average-linkage algorithm we found that data-point membership, cluster shape, and cluster size depended on the selected PC subspace data. In contrast, K-means provided very consistent results regardless of the selected subspace. Since we present results on a single model system, generalization concerning the clustering of different PC subspaces of other molecular systems is currently premature. However, our hope is that this study illustrates a) the complexities in selecting the appropriate clustering algorithm, b) the complexities in interpreting and validating their results, and c) by combining PC analysis with subsequent clustering valuable dynamic and conformational information can be obtained.

Hyperspectral Super-Resolution of Locally Low Rank Images From Complementary Multisource Data.

PubMed

Veganzones, Miguel A; Simoes, Miguel; Licciardi, Giorgio; Yokoya, Naoto; Bioucas-Dias, Jose M; Chanussot, Jocelyn

2016-01-01

Remote sensing hyperspectral images (HSIs) are quite often low rank, in the sense that the data belong to a low dimensional subspace/manifold. This has been recently exploited for the fusion of low spatial resolution HSI with high spatial resolution multispectral images in order to obtain super-resolution HSI. Most approaches adopt an unmixing or a matrix factorization perspective. The derived methods have led to state-of-the-art results when the spectral information lies in a low-dimensional subspace/manifold. However, if the subspace/manifold dimensionality spanned by the complete data set is large, i.e., larger than the number of multispectral bands, the performance of these methods mainly decreases because the underlying sparse regression problem is severely ill-posed. In this paper, we propose a local approach to cope with this difficulty. Fundamentally, we exploit the fact that real world HSIs are locally low rank, that is, pixels acquired from a given spatial neighborhood span a very low-dimensional subspace/manifold, i.e., lower or equal than the number of multispectral bands. Thus, we propose to partition the image into patches and solve the data fusion problem independently for each patch. This way, in each patch the subspace/manifold dimensionality is low enough, such that the problem is not ill-posed anymore. We propose two alternative approaches to define the hyperspectral super-resolution through local dictionary learning using endmember induction algorithms. We also explore two alternatives to define the local regions, using sliding windows and binary partition trees. The effectiveness of the proposed approaches is illustrated with synthetic and semi real data.

Machine learning strategy for accelerated design of polymer dielectrics

DOE PAGES

Mannodi-Kanakkithodi, Arun; Pilania, Ghanshyam; Huan, Tran Doan; ...

2016-02-15

The ability to efficiently design new and advanced dielectric polymers is hampered by the lack of sufficient, reliable data on wide polymer chemical spaces, and the difficulty of generating such data given time and computational/experimental constraints. Here, we address the issue of accelerating polymer dielectrics design by extracting learning models from data generated by accurate state-of-the-art first principles computations for polymers occupying an important part of the chemical subspace. The polymers are ‘fingerprinted’ as simple, easily attainable numerical representations, which are mapped to the properties of interest using a machine learning algorithm to develop an on-demand property prediction model. Further,more » a genetic algorithm is utilised to optimise polymer constituent blocks in an evolutionary manner, thus directly leading to the design of polymers with given target properties. Furthermore, while this philosophy of learning to make instant predictions and design is demonstrated here for the example of polymer dielectrics, it is equally applicable to other classes of materials as well.« less

Learning Multirobot Hose Transportation and Deployment by Distributed Round-Robin Q-Learning.

PubMed

Fernandez-Gauna, Borja; Etxeberria-Agiriano, Ismael; Graña, Manuel

2015-01-01

Multi-Agent Reinforcement Learning (MARL) algorithms face two main difficulties: the curse of dimensionality, and environment non-stationarity due to the independent learning processes carried out by the agents concurrently. In this paper we formalize and prove the convergence of a Distributed Round Robin Q-learning (D-RR-QL) algorithm for cooperative systems. The computational complexity of this algorithm increases linearly with the number of agents. Moreover, it eliminates environment non sta tionarity by carrying a round-robin scheduling of the action selection and execution. That this learning scheme allows the implementation of Modular State-Action Vetoes (MSAV) in cooperative multi-agent systems, which speeds up learning convergence in over-constrained systems by vetoing state-action pairs which lead to undesired termination states (UTS) in the relevant state-action subspace. Each agent's local state-action value function learning is an independent process, including the MSAV policies. Coordination of locally optimal policies to obtain the global optimal joint policy is achieved by a greedy selection procedure using message passing. We show that D-RR-QL improves over state-of-the-art approaches, such as Distributed Q-Learning, Team Q-Learning and Coordinated Reinforcement Learning in a paradigmatic Linked Multi-Component Robotic System (L-MCRS) control problem: the hose transportation task. L-MCRS are over-constrained systems with many UTS induced by the interaction of the passive linking element and the active mobile robots.

Visual tracking based on the sparse representation of the PCA subspace

NASA Astrophysics Data System (ADS)

Chen, Dian-bing; Zhu, Ming; Wang, Hui-li

2017-09-01

We construct a collaborative model of the sparse representation and the subspace representation. First, we represent the tracking target in the principle component analysis (PCA) subspace, and then we employ an L 1 regularization to restrict the sparsity of the residual term, an L 2 regularization term to restrict the sparsity of the representation coefficients, and an L 2 norm to restrict the distance between the reconstruction and the target. Then we implement the algorithm in the particle filter framework. Furthermore, an iterative method is presented to get the global minimum of the residual and the coefficients. Finally, an alternative template update scheme is adopted to avoid the tracking drift which is caused by the inaccurate update. In the experiment, we test the algorithm on 9 sequences, and compare the results with 5 state-of-art methods. According to the results, we can conclude that our algorithm is more robust than the other methods.

Synthetic Modeling of Autonomous Learning with a Chaotic Neural Network

NASA Astrophysics Data System (ADS)

Funabashi, Masatoshi

We investigate the possible role of intermittent chaotic dynamics called chaotic itinerancy, in interaction with nonsupervised learnings that reinforce and weaken the neural connection depending on the dynamics itself. We first performed hierarchical stability analysis of the Chaotic Neural Network model (CNN) according to the structure of invariant subspaces. Irregular transition between two attractor ruins with positive maximum Lyapunov exponent was triggered by the blowout bifurcation of the attractor spaces, and was associated with riddled basins structure. We secondly modeled two autonomous learnings, Hebbian learning and spike-timing-dependent plasticity (STDP) rule, and simulated the effect on the chaotic itinerancy state of CNN. Hebbian learning increased the residence time on attractor ruins, and produced novel attractors in the minimum higher-dimensional subspace. It also augmented the neuronal synchrony and established the uniform modularity in chaotic itinerancy. STDP rule reduced the residence time on attractor ruins, and brought a wide range of periodicity in emerged attractors, possibly including strange attractors. Both learning rules selectively destroyed and preserved the specific invariant subspaces, depending on the neuron synchrony of the subspace where the orbits are situated. Computational rationale of the autonomous learning is discussed in connectionist perspective.

Primary decomposition of zero-dimensional ideals over finite fields

NASA Astrophysics Data System (ADS)

Gao, Shuhong; Wan, Daqing; Wang, Mingsheng

2009-03-01

A new algorithm is presented for computing primary decomposition of zero-dimensional ideals over finite fields. Like Berlekamp's algorithm for univariate polynomials, the new method is based on the invariant subspace of the Frobenius map acting on the quotient algebra. The dimension of the invariant subspace equals the number of primary components, and a basis of the invariant subspace yields a complete decomposition. Unlike previous approaches for decomposing multivariate polynomial systems, the new method does not need primality testing nor any generic projection, instead it reduces the general decomposition problem directly to root finding of univariate polynomials over the ground field. Also, it is shown how Groebner basis structure can be used to get partial primary decomposition without any root finding.

Krylov subspace methods - Theory, algorithms, and applications

NASA Technical Reports Server (NTRS)

Sad, Youcef

1990-01-01

Projection methods based on Krylov subspaces for solving various types of scientific problems are reviewed. The main idea of this class of methods when applied to a linear system Ax = b, is to generate in some manner an approximate solution to the original problem from the so-called Krylov subspace span. Thus, the original problem of size N is approximated by one of dimension m, typically much smaller than N. Krylov subspace methods have been very successful in solving linear systems and eigenvalue problems and are now becoming popular for solving nonlinear equations. The main ideas in Krylov subspace methods are shown and their use in solving linear systems, eigenvalue problems, parabolic partial differential equations, Liapunov matrix equations, and nonlinear system of equations are discussed.

Computing interior eigenvalues of nonsymmetric matrices: application to three-dimensional metamaterial composites.

PubMed

Terao, Takamichi

2010-08-01

We propose a numerical method to calculate interior eigenvalues and corresponding eigenvectors for nonsymmetric matrices. Based on the subspace projection technique onto expanded Ritz subspace, it becomes possible to obtain eigenvalues and eigenvectors with sufficiently high precision. This method overcomes the difficulties of the traditional nonsymmetric Lanczos algorithm, and improves the accuracy of the obtained interior eigenvalues and eigenvectors. Using this algorithm, we investigate three-dimensional metamaterial composites consisting of positive and negative refractive index materials, and it is demonstrated that the finite-difference frequency-domain algorithm is applicable to analyze these metamaterial composites.

An analysis of value function learning with piecewise linear control

NASA Astrophysics Data System (ADS)

Tutsoy, Onder; Brown, Martin

2016-05-01

Reinforcement learning (RL) algorithms attempt to learn optimal control actions by iteratively estimating a long-term measure of system performance, the so-called value function. For example, RL algorithms have been applied to walking robots to examine the connection between robot motion and the brain, which is known as embodied cognition. In this paper, RL algorithms are analysed using an exemplar test problem. A closed form solution for the value function is calculated and this is represented in terms of a set of basis functions and parameters, which is used to investigate parameter convergence. The value function expression is shown to have a polynomial form where the polynomial terms depend on the plant's parameters and the value function's discount factor. It is shown that the temporal difference error introduces a null space for the differenced higher order basis associated with the effects of controller switching (saturated to linear control or terminating an experiment) apart from the time of the switch. This leads to slow convergence in the relevant subspace. It is also shown that badly conditioned learning problems can occur, and this is a function of the value function discount factor and the controller switching points. Finally, a comparison is performed between the residual gradient and TD(0) learning algorithms, and it is shown that the former has a faster rate of convergence for this test problem.

Low-Rank Tensor Subspace Learning for RGB-D Action Recognition.

PubMed

Jia, Chengcheng; Fu, Yun

2016-07-09

Since RGB-D action data inherently equip with extra depth information compared with RGB data, recently many works employ RGB-D data in a third-order tensor representation containing spatio-temporal structure to find a subspace for action recognition. However, there are two main challenges of these methods. First, the dimension of subspace is usually fixed manually. Second, preserving local information by finding intraclass and inter-class neighbors from a manifold is highly timeconsuming. In this paper, we learn a tensor subspace, whose dimension is learned automatically by low-rank learning, for RGB-D action recognition. Particularly, the tensor samples are factorized to obtain three Projection Matrices (PMs) by Tucker Decomposition, where all the PMs are performed by nuclear norm in a close-form to obtain the tensor ranks which are used as tensor subspace dimension. Additionally, we extract the discriminant and local information from a manifold using a graph constraint. This graph preserves the local knowledge inherently, which is faster than the previous way by calculating both the intra-class and inter-class neighbors of each sample. We evaluate the proposed method on four widely used RGB-D action datasets including MSRDailyActivity3D, MSRActionPairs, MSRActionPairs skeleton and UTKinect-Action3D datasets, and the experimental results show higher accuracy and efficiency of the proposed method.

DOE Office of Scientific and Technical Information (OSTI.GOV)

Krause, Josua; Dasgupta, Aritra; Fekete, Jean-Daniel

Dealing with the curse of dimensionality is a key challenge in high-dimensional data visualization. We present SeekAView to address three main gaps in the existing research literature. First, automated methods like dimensionality reduction or clustering suffer from a lack of transparency in letting analysts interact with their outputs in real-time to suit their exploration strategies. The results often suffer from a lack of interpretability, especially for domain experts not trained in statistics and machine learning. Second, exploratory visualization techniques like scatter plots or parallel coordinates suffer from a lack of visual scalability: it is difficult to present a coherent overviewmore » of interesting combinations of dimensions. Third, the existing techniques do not provide a flexible workflow that allows for multiple perspectives into the analysis process by automatically detecting and suggesting potentially interesting subspaces. In SeekAView we address these issues using suggestion based visual exploration of interesting patterns for building and refining multidimensional subspaces. Compared to the state-of-the-art in subspace search and visualization methods, we achieve higher transparency in showing not only the results of the algorithms, but also interesting dimensions calibrated against different metrics. We integrate a visually scalable design space with an iterative workflow guiding the analysts by choosing the starting points and letting them slice and dice through the data to find interesting subspaces and detect correlations, clusters, and outliers. We present two usage scenarios for demonstrating how SeekAView can be applied in real-world data analysis scenarios.« less

Machine learning algorithms for the creation of clinical healthcare enterprise systems

NASA Astrophysics Data System (ADS)

Mandal, Indrajit

2017-10-01

Clinical recommender systems are increasingly becoming popular for improving modern healthcare systems. Enterprise systems are persuasively used for creating effective nurse care plans to provide nurse training, clinical recommendations and clinical quality control. A novel design of a reliable clinical recommender system based on multiple classifier system (MCS) is implemented. A hybrid machine learning (ML) ensemble based on random subspace method and random forest is presented. The performance accuracy and robustness of proposed enterprise architecture are quantitatively estimated to be above 99% and 97%, respectively (above 95% confidence interval). The study then extends to experimental analysis of the clinical recommender system with respect to the noisy data environment. The ranking of items in nurse care plan is demonstrated using machine learning algorithms (MLAs) to overcome the drawback of the traditional association rule method. The promising experimental results are compared against the sate-of-the-art approaches to highlight the advancement in recommendation technology. The proposed recommender system is experimentally validated using five benchmark clinical data to reinforce the research findings.

Subspace-based optimization method for inverse scattering problems with an inhomogeneous background medium

NASA Astrophysics Data System (ADS)

Chen, Xudong

2010-07-01

This paper proposes a version of the subspace-based optimization method to solve the inverse scattering problem with an inhomogeneous background medium where the known inhomogeneities are bounded in a finite domain. Although the background Green's function at each discrete point in the computational domain is not directly available in an inhomogeneous background scenario, the paper uses the finite element method to simultaneously obtain the Green's function at all discrete points. The essence of the subspace-based optimization method is that part of the contrast source is determined from the spectrum analysis without using any optimization, whereas the orthogonally complementary part is determined by solving a lower dimension optimization problem. This feature significantly speeds up the convergence of the algorithm and at the same time makes it robust against noise. Numerical simulations illustrate the efficacy of the proposed algorithm. The algorithm presented in this paper finds wide applications in nondestructive evaluation, such as through-wall imaging.

Recognition of multiple imbalanced cancer types based on DNA microarray data using ensemble classifiers.

PubMed

Yu, Hualong; Hong, Shufang; Yang, Xibei; Ni, Jun; Dan, Yuanyuan; Qin, Bin

2013-01-01

DNA microarray technology can measure the activities of tens of thousands of genes simultaneously, which provides an efficient way to diagnose cancer at the molecular level. Although this strategy has attracted significant research attention, most studies neglect an important problem, namely, that most DNA microarray datasets are skewed, which causes traditional learning algorithms to produce inaccurate results. Some studies have considered this problem, yet they merely focus on binary-class problem. In this paper, we dealt with multiclass imbalanced classification problem, as encountered in cancer DNA microarray, by using ensemble learning. We utilized one-against-all coding strategy to transform multiclass to multiple binary classes, each of them carrying out feature subspace, which is an evolving version of random subspace that generates multiple diverse training subsets. Next, we introduced one of two different correction technologies, namely, decision threshold adjustment or random undersampling, into each training subset to alleviate the damage of class imbalance. Specifically, support vector machine was used as base classifier, and a novel voting rule called counter voting was presented for making a final decision. Experimental results on eight skewed multiclass cancer microarray datasets indicate that unlike many traditional classification approaches, our methods are insensitive to class imbalance.

Discriminant projective non-negative matrix factorization.

PubMed

Guan, Naiyang; Zhang, Xiang; Luo, Zhigang; Tao, Dacheng; Yang, Xuejun

2013-01-01

Projective non-negative matrix factorization (PNMF) projects high-dimensional non-negative examples X onto a lower-dimensional subspace spanned by a non-negative basis W and considers W(T) X as their coefficients, i.e., X≈WW(T) X. Since PNMF learns the natural parts-based representation Wof X, it has been widely used in many fields such as pattern recognition and computer vision. However, PNMF does not perform well in classification tasks because it completely ignores the label information of the dataset. This paper proposes a Discriminant PNMF method (DPNMF) to overcome this deficiency. In particular, DPNMF exploits Fisher's criterion to PNMF for utilizing the label information. Similar to PNMF, DPNMF learns a single non-negative basis matrix and needs less computational burden than NMF. In contrast to PNMF, DPNMF maximizes the distance between centers of any two classes of examples meanwhile minimizes the distance between any two examples of the same class in the lower-dimensional subspace and thus has more discriminant power. We develop a multiplicative update rule to solve DPNMF and prove its convergence. Experimental results on four popular face image datasets confirm its effectiveness comparing with the representative NMF and PNMF algorithms.

Discriminant Projective Non-Negative Matrix Factorization

PubMed Central

Guan, Naiyang; Zhang, Xiang; Luo, Zhigang; Tao, Dacheng; Yang, Xuejun

2013-01-01

Projective non-negative matrix factorization (PNMF) projects high-dimensional non-negative examples X onto a lower-dimensional subspace spanned by a non-negative basis W and considers WT X as their coefficients, i.e., X≈WWT X. Since PNMF learns the natural parts-based representation Wof X, it has been widely used in many fields such as pattern recognition and computer vision. However, PNMF does not perform well in classification tasks because it completely ignores the label information of the dataset. This paper proposes a Discriminant PNMF method (DPNMF) to overcome this deficiency. In particular, DPNMF exploits Fisher's criterion to PNMF for utilizing the label information. Similar to PNMF, DPNMF learns a single non-negative basis matrix and needs less computational burden than NMF. In contrast to PNMF, DPNMF maximizes the distance between centers of any two classes of examples meanwhile minimizes the distance between any two examples of the same class in the lower-dimensional subspace and thus has more discriminant power. We develop a multiplicative update rule to solve DPNMF and prove its convergence. Experimental results on four popular face image datasets confirm its effectiveness comparing with the representative NMF and PNMF algorithms. PMID:24376680

Manifold regularized multitask learning for semi-supervised multilabel image classification.

PubMed

Luo, Yong; Tao, Dacheng; Geng, Bo; Xu, Chao; Maybank, Stephen J

2013-02-01

It is a significant challenge to classify images with multiple labels by using only a small number of labeled samples. One option is to learn a binary classifier for each label and use manifold regularization to improve the classification performance by exploring the underlying geometric structure of the data distribution. However, such an approach does not perform well in practice when images from multiple concepts are represented by high-dimensional visual features. Thus, manifold regularization is insufficient to control the model complexity. In this paper, we propose a manifold regularized multitask learning (MRMTL) algorithm. MRMTL learns a discriminative subspace shared by multiple classification tasks by exploiting the common structure of these tasks. It effectively controls the model complexity because different tasks limit one another's search volume, and the manifold regularization ensures that the functions in the shared hypothesis space are smooth along the data manifold. We conduct extensive experiments, on the PASCAL VOC'07 dataset with 20 classes and the MIR dataset with 38 classes, by comparing MRMTL with popular image classification algorithms. The results suggest that MRMTL is effective for image classification.

Structured sparse linear graph embedding.

PubMed

Wang, Haixian

2012-03-01

Subspace learning is a core issue in pattern recognition and machine learning. Linear graph embedding (LGE) is a general framework for subspace learning. In this paper, we propose a structured sparse extension to LGE (SSLGE) by introducing a structured sparsity-inducing norm into LGE. Specifically, SSLGE casts the projection bases learning into a regression-type optimization problem, and then the structured sparsity regularization is applied to the regression coefficients. The regularization selects a subset of features and meanwhile encodes high-order information reflecting a priori structure information of the data. The SSLGE technique provides a unified framework for discovering structured sparse subspace. Computationally, by using a variational equality and the Procrustes transformation, SSLGE is efficiently solved with closed-form updates. Experimental results on face image show the effectiveness of the proposed method. Copyright © 2011 Elsevier Ltd. All rights reserved.

Beyond union of subspaces: Subspace pursuit on Grassmann manifold for data representation

DOE PAGES

Shen, Xinyue; Krim, Hamid; Gu, Yuantao

2016-03-01

Discovering the underlying structure of a high-dimensional signal or big data has always been a challenging topic, and has become harder to tackle especially when the observations are exposed to arbitrary sparse perturbations. Here in this paper, built on the model of a union of subspaces (UoS) with sparse outliers and inspired by a basis pursuit strategy, we exploit the fundamental structure of a Grassmann manifold, and propose a new technique of pursuing the subspaces systematically by solving a non-convex optimization problem using the alternating direction method of multipliers. This problem as noted is further complicated by non-convex constraints onmore » the Grassmann manifold, as well as the bilinearity in the penalty caused by the subspace bases and coefficients. Nevertheless, numerical experiments verify that the proposed algorithm, which provides elegant solutions to the sub-problems in each step, is able to de-couple the subspaces and pursue each of them under time-efficient parallel computation.« less

Sampling algorithms for validation of supervised learning models for Ising-like systems

NASA Astrophysics Data System (ADS)

Portman, Nataliya; Tamblyn, Isaac

2017-12-01

In this paper, we build and explore supervised learning models of ferromagnetic system behavior, using Monte-Carlo sampling of the spin configuration space generated by the 2D Ising model. Given the enormous size of the space of all possible Ising model realizations, the question arises as to how to choose a reasonable number of samples that will form physically meaningful and non-intersecting training and testing datasets. Here, we propose a sampling technique called ;ID-MH; that uses the Metropolis-Hastings algorithm creating Markov process across energy levels within the predefined configuration subspace. We show that application of this method retains phase transitions in both training and testing datasets and serves the purpose of validation of a machine learning algorithm. For larger lattice dimensions, ID-MH is not feasible as it requires knowledge of the complete configuration space. As such, we develop a new ;block-ID; sampling strategy: it decomposes the given structure into square blocks with lattice dimension N ≤ 5 and uses ID-MH sampling of candidate blocks. Further comparison of the performance of commonly used machine learning methods such as random forests, decision trees, k nearest neighbors and artificial neural networks shows that the PCA-based Decision Tree regressor is the most accurate predictor of magnetizations of the Ising model. For energies, however, the accuracy of prediction is not satisfactory, highlighting the need to consider more algorithmically complex methods (e.g., deep learning).

Subspace-based analysis of the ERT inverse problem

NASA Astrophysics Data System (ADS)

Ben Hadj Miled, Mohamed Khames; Miller, Eric L.

2004-05-01

In a previous work, we proposed a source-type formulation to the electrical resistance tomography (ERT) problem. Specifically, we showed that inhomogeneities in the medium can be viewed as secondary sources embedded in the homogeneous background medium and located at positions associated with variation in electrical conductivity. Assuming a piecewise constant conductivity distribution, the support of equivalent sources is equal to the boundary of the inhomogeneity. The estimation of the anomaly shape takes the form of an inverse source-type problem. In this paper, we explore the use of subspace methods to localize the secondary equivalent sources associated with discontinuities in the conductivity distribution. Our first alternative is the multiple signal classification (MUSIC) algorithm which is commonly used in the localization of multiple sources. The idea is to project a finite collection of plausible pole (or dipole) sources onto an estimated signal subspace and select those with largest correlations. In ERT, secondary sources are excited simultaneously but in different ways, i.e. with distinct amplitude patterns, depending on the locations and amplitudes of primary sources. If the number of receivers is "large enough", different source configurations can lead to a set of observation vectors that span the data subspace. However, since sources that are spatially close to each other have highly correlated signatures, seperation of such signals becomes very difficult in the presence of noise. To overcome this problem we consider iterative MUSIC algorithms like R-MUSIC and RAP-MUSIC. These recursive algorithms pose a computational burden as they require multiple large combinatorial searches. Results obtained with these algorithms using simulated data of different conductivity patterns are presented.

Blind equalization and automatic modulation classification based on subspace for subcarrier MPSK optical communications

NASA Astrophysics Data System (ADS)

Chen, Dan; Guo, Lin-yuan; Wang, Chen-hao; Ke, Xi-zheng

2017-07-01

Equalization can compensate channel distortion caused by channel multipath effects, and effectively improve convergent of modulation constellation diagram in optical wireless system. In this paper, the subspace blind equalization algorithm is used to preprocess M-ary phase shift keying (MPSK) subcarrier modulation signal in receiver. Mountain clustering is adopted to get the clustering centers of MPSK modulation constellation diagram, and the modulation order is automatically identified through the k-nearest neighbor (KNN) classifier. The experiment has been done under four different weather conditions. Experimental results show that the convergent of constellation diagram is improved effectively after using the subspace blind equalization algorithm, which means that the accuracy of modulation recognition is increased. The correct recognition rate of 16PSK can be up to 85% in any kind of weather condition which is mentioned in paper. Meanwhile, the correct recognition rate is the highest in cloudy and the lowest in heavy rain condition.

Face verification with balanced thresholds.

PubMed

Yan, Shuicheng; Xu, Dong; Tang, Xiaoou

2007-01-01

The process of face verification is guided by a pre-learned global threshold, which, however, is often inconsistent with class-specific optimal thresholds. It is, hence, beneficial to pursue a balance of the class-specific thresholds in the model-learning stage. In this paper, we present a new dimensionality reduction algorithm tailored to the verification task that ensures threshold balance. This is achieved by the following aspects. First, feasibility is guaranteed by employing an affine transformation matrix, instead of the conventional projection matrix, for dimensionality reduction, and, hence, we call the proposed algorithm threshold balanced transformation (TBT). Then, the affine transformation matrix, constrained as the product of an orthogonal matrix and a diagonal matrix, is optimized to improve the threshold balance and classification capability in an iterative manner. Unlike most algorithms for face verification which are directly transplanted from face identification literature, TBT is specifically designed for face verification and clarifies the intrinsic distinction between these two tasks. Experiments on three benchmark face databases demonstrate that TBT significantly outperforms the state-of-the-art subspace techniques for face verification.

Characterizing L1-norm best-fit subspaces

NASA Astrophysics Data System (ADS)

Brooks, J. Paul; Dulá, José H.

2017-05-01

Fitting affine objects to data is the basis of many tools and methodologies in statistics, machine learning, and signal processing. The L1 norm is often employed to produce subspaces exhibiting a robustness to outliers and faulty observations. The L1-norm best-fit subspace problem is directly formulated as a nonlinear, nonconvex, and nondifferentiable optimization problem. The case when the subspace is a hyperplane can be solved to global optimality efficiently by solving a series of linear programs. The problem of finding the best-fit line has recently been shown to be NP-hard. We present necessary conditions for optimality for the best-fit subspace problem, and use them to characterize properties of optimal solutions.

[Orthogonal Vector Projection Algorithm for Spectral Unmixing].

PubMed

Song, Mei-ping; Xu, Xing-wei; Chang, Chein-I; An, Ju-bai; Yao, Li

2015-12-01

Spectrum unmixing is an important part of hyperspectral technologies, which is essential for material quantity analysis in hyperspectral imagery. Most linear unmixing algorithms require computations of matrix multiplication and matrix inversion or matrix determination. These are difficult for programming, especially hard for realization on hardware. At the same time, the computation costs of the algorithms increase significantly as the number of endmembers grows. Here, based on the traditional algorithm Orthogonal Subspace Projection, a new method called. Orthogonal Vector Projection is prompted using orthogonal principle. It simplifies this process by avoiding matrix multiplication and inversion. It firstly computes the final orthogonal vector via Gram-Schmidt process for each endmember spectrum. And then, these orthogonal vectors are used as projection vector for the pixel signature. The unconstrained abundance can be obtained directly by projecting the signature to the projection vectors, and computing the ratio of projected vector length and orthogonal vector length. Compared to the Orthogonal Subspace Projection and Least Squares Error algorithms, this method does not need matrix inversion, which is much computation costing and hard to implement on hardware. It just completes the orthogonalization process by repeated vector operations, easy for application on both parallel computation and hardware. The reasonability of the algorithm is proved by its relationship with Orthogonal Sub-space Projection and Least Squares Error algorithms. And its computational complexity is also compared with the other two algorithms', which is the lowest one. At last, the experimental results on synthetic image and real image are also provided, giving another evidence for effectiveness of the method.

An alternative subspace approach to EEG dipole source localization

NASA Astrophysics Data System (ADS)

Xu, Xiao-Liang; Xu, Bobby; He, Bin

2004-01-01

In the present study, we investigate a new approach to electroencephalography (EEG) three-dimensional (3D) dipole source localization by using a non-recursive subspace algorithm called FINES. In estimating source dipole locations, the present approach employs projections onto a subspace spanned by a small set of particular vectors (FINES vector set) in the estimated noise-only subspace instead of the entire estimated noise-only subspace in the case of classic MUSIC. The subspace spanned by this vector set is, in the sense of principal angle, closest to the subspace spanned by the array manifold associated with a particular brain region. By incorporating knowledge of the array manifold in identifying FINES vector sets in the estimated noise-only subspace for different brain regions, the present approach is able to estimate sources with enhanced accuracy and spatial resolution, thus enhancing the capability of resolving closely spaced sources and reducing estimation errors. The present computer simulations show, in EEG 3D dipole source localization, that compared to classic MUSIC, FINES has (1) better resolvability of two closely spaced dipolar sources and (2) better estimation accuracy of source locations. In comparison with RAP-MUSIC, FINES' performance is also better for the cases studied when the noise level is high and/or correlations among dipole sources exist.

Non-Convex Sparse and Low-Rank Based Robust Subspace Segmentation for Data Mining.

PubMed

Cheng, Wenlong; Zhao, Mingbo; Xiong, Naixue; Chui, Kwok Tai

2017-07-15

Parsimony, including sparsity and low-rank, has shown great importance for data mining in social networks, particularly in tasks such as segmentation and recognition. Traditionally, such modeling approaches rely on an iterative algorithm that minimizes an objective function with convex l ₁-norm or nuclear norm constraints. However, the obtained results by convex optimization are usually suboptimal to solutions of original sparse or low-rank problems. In this paper, a novel robust subspace segmentation algorithm has been proposed by integrating l p -norm and Schatten p -norm constraints. Our so-obtained affinity graph can better capture local geometrical structure and the global information of the data. As a consequence, our algorithm is more generative, discriminative and robust. An efficient linearized alternating direction method is derived to realize our model. Extensive segmentation experiments are conducted on public datasets. The proposed algorithm is revealed to be more effective and robust compared to five existing algorithms.

Heterogeneous Tensor Decomposition for Clustering via Manifold Optimization.

PubMed

Sun, Yanfeng; Gao, Junbin; Hong, Xia; Mishra, Bamdev; Yin, Baocai

2016-03-01

Tensor clustering is an important tool that exploits intrinsically rich structures in real-world multiarray or Tensor datasets. Often in dealing with those datasets, standard practice is to use subspace clustering that is based on vectorizing multiarray data. However, vectorization of tensorial data does not exploit complete structure information. In this paper, we propose a subspace clustering algorithm without adopting any vectorization process. Our approach is based on a novel heterogeneous Tucker decomposition model taking into account cluster membership information. We propose a new clustering algorithm that alternates between different modes of the proposed heterogeneous tensor model. All but the last mode have closed-form updates. Updating the last mode reduces to optimizing over the multinomial manifold for which we investigate second order Riemannian geometry and propose a trust-region algorithm. Numerical experiments show that our proposed algorithm compete effectively with state-of-the-art clustering algorithms that are based on tensor factorization.

Correlational Neural Networks.

PubMed

Chandar, Sarath; Khapra, Mitesh M; Larochelle, Hugo; Ravindran, Balaraman

2016-02-01

Common representation learning (CRL), wherein different descriptions (or views) of the data are embedded in a common subspace, has been receiving a lot of attention recently. Two popular paradigms here are canonical correlation analysis (CCA)-based approaches and autoencoder (AE)-based approaches. CCA-based approaches learn a joint representation by maximizing correlation of the views when projected to the common subspace. AE-based methods learn a common representation by minimizing the error of reconstructing the two views. Each of these approaches has its own advantages and disadvantages. For example, while CCA-based approaches outperform AE-based approaches for the task of transfer learning, they are not as scalable as the latter. In this work, we propose an AE-based approach, correlational neural network (CorrNet), that explicitly maximizes correlation among the views when projected to the common subspace. Through a series of experiments, we demonstrate that the proposed CorrNet is better than AE and CCA with respect to its ability to learn correlated common representations. We employ CorrNet for several cross-language tasks and show that the representations learned using it perform better than the ones learned using other state-of-the-art approaches.

SubspaceEM: A Fast Maximum-a-posteriori Algorithm for Cryo-EM Single Particle Reconstruction

PubMed Central

Dvornek, Nicha C.; Sigworth, Fred J.; Tagare, Hemant D.

2015-01-01

Single particle reconstruction methods based on the maximum-likelihood principle and the expectation-maximization (E–M) algorithm are popular because of their ability to produce high resolution structures. However, these algorithms are computationally very expensive, requiring a network of computational servers. To overcome this computational bottleneck, we propose a new mathematical framework for accelerating maximum-likelihood reconstructions. The speedup is by orders of magnitude and the proposed algorithm produces similar quality reconstructions compared to the standard maximum-likelihood formulation. Our approach uses subspace approximations of the cryo-electron microscopy (cryo-EM) data and projection images, greatly reducing the number of image transformations and comparisons that are computed. Experiments using simulated and actual cryo-EM data show that speedup in overall execution time compared to traditional maximum-likelihood reconstruction reaches factors of over 300. PMID:25839831

Constrained dictionary learning and probabilistic hypergraph ranking for person re-identification

NASA Astrophysics Data System (ADS)

He, You; Wu, Song; Pu, Nan; Qian, Li; Xiao, Guoqiang

2018-04-01

Person re-identification is a fundamental and inevitable task in public security. In this paper, we propose a novel framework to improve the performance of this task. First, two different types of descriptors are extracted to represent a pedestrian: (1) appearance-based superpixel features, which are constituted mainly by conventional color features and extracted from the supepixel rather than a whole picture and (2) due to the limitation of discrimination of appearance features, the deep features extracted by feature fusion Network are also used. Second, a view invariant subspace is learned by dictionary learning constrained by the minimum negative sample (termed as DL-cMN) to reduce the noise in appearance-based superpixel feature domain. Then, we use deep features and sparse codes transformed by appearancebased features to establish the hyperedges respectively by k-nearest neighbor, rather than jointing different features simply. Finally, a final ranking is performed by probabilistic hypergraph ranking algorithm. Extensive experiments on three challenging datasets (VIPeR, PRID450S and CUHK01) demonstrate the advantages and effectiveness of our proposed algorithm.

Interactive machine learning for health informatics: when do we need the human-in-the-loop?

PubMed

Holzinger, Andreas

2016-06-01

Machine learning (ML) is the fastest growing field in computer science, and health informatics is among the greatest challenges. The goal of ML is to develop algorithms which can learn and improve over time and can be used for predictions. Most ML researchers concentrate on automatic machine learning (aML), where great advances have been made, for example, in speech recognition, recommender systems, or autonomous vehicles. Automatic approaches greatly benefit from big data with many training sets. However, in the health domain, sometimes we are confronted with a small number of data sets or rare events, where aML-approaches suffer of insufficient training samples. Here interactive machine learning (iML) may be of help, having its roots in reinforcement learning, preference learning, and active learning. The term iML is not yet well used, so we define it as "algorithms that can interact with agents and can optimize their learning behavior through these interactions, where the agents can also be human." This "human-in-the-loop" can be beneficial in solving computationally hard problems, e.g., subspace clustering, protein folding, or k-anonymization of health data, where human expertise can help to reduce an exponential search space through heuristic selection of samples. Therefore, what would otherwise be an NP-hard problem, reduces greatly in complexity through the input and the assistance of a human agent involved in the learning phase.

A projected preconditioned conjugate gradient algorithm for computing many extreme eigenpairs of a Hermitian matrix [A projected preconditioned conjugate gradient algorithm for computing a large eigenspace of a Hermitian matrix

DOE PAGES

Vecharynski, Eugene; Yang, Chao; Pask, John E.

2015-02-25

Here, we present an iterative algorithm for computing an invariant subspace associated with the algebraically smallest eigenvalues of a large sparse or structured Hermitian matrix A. We are interested in the case in which the dimension of the invariant subspace is large (e.g., over several hundreds or thousands) even though it may still be small relative to the dimension of A. These problems arise from, for example, density functional theory (DFT) based electronic structure calculations for complex materials. The key feature of our algorithm is that it performs fewer Rayleigh–Ritz calculations compared to existing algorithms such as the locally optimalmore » block preconditioned conjugate gradient or the Davidson algorithm. It is a block algorithm, and hence can take advantage of efficient BLAS3 operations and be implemented with multiple levels of concurrency. We discuss a number of practical issues that must be addressed in order to implement the algorithm efficiently on a high performance computer.« less

Differential evolution algorithm-based kernel parameter selection for Fukunaga-Koontz Transform subspaces construction

NASA Astrophysics Data System (ADS)

Binol, Hamidullah; Bal, Abdullah; Cukur, Huseyin

2015-10-01

The performance of the kernel based techniques depends on the selection of kernel parameters. That's why; suitable parameter selection is an important problem for many kernel based techniques. This article presents a novel technique to learn the kernel parameters in kernel Fukunaga-Koontz Transform based (KFKT) classifier. The proposed approach determines the appropriate values of kernel parameters through optimizing an objective function constructed based on discrimination ability of KFKT. For this purpose we have utilized differential evolution algorithm (DEA). The new technique overcomes some disadvantages such as high time consumption existing in the traditional cross-validation method, and it can be utilized in any type of data. The experiments for target detection applications on the hyperspectral images verify the effectiveness of the proposed method.

Koopman Invariant Subspaces and Finite Linear Representations of Nonlinear Dynamical Systems for Control.

PubMed

Brunton, Steven L; Brunton, Bingni W; Proctor, Joshua L; Kutz, J Nathan

2016-01-01

In this wIn this work, we explore finite-dimensional linear representations of nonlinear dynamical systems by restricting the Koopman operator to an invariant subspace spanned by specially chosen observable functions. The Koopman operator is an infinite-dimensional linear operator that evolves functions of the state of a dynamical system. Dominant terms in the Koopman expansion are typically computed using dynamic mode decomposition (DMD). DMD uses linear measurements of the state variables, and it has recently been shown that this may be too restrictive for nonlinear systems. Choosing the right nonlinear observable functions to form an invariant subspace where it is possible to obtain linear reduced-order models, especially those that are useful for control, is an open challenge. Here, we investigate the choice of observable functions for Koopman analysis that enable the use of optimal linear control techniques on nonlinear problems. First, to include a cost on the state of the system, as in linear quadratic regulator (LQR) control, it is helpful to include these states in the observable subspace, as in DMD. However, we find that this is only possible when there is a single isolated fixed point, as systems with multiple fixed points or more complicated attractors are not globally topologically conjugate to a finite-dimensional linear system, and cannot be represented by a finite-dimensional linear Koopman subspace that includes the state. We then present a data-driven strategy to identify relevant observable functions for Koopman analysis by leveraging a new algorithm to determine relevant terms in a dynamical system by ℓ1-regularized regression of the data in a nonlinear function space; we also show how this algorithm is related to DMD. Finally, we demonstrate the usefulness of nonlinear observable subspaces in the design of Koopman operator optimal control laws for fully nonlinear systems using techniques from linear optimal control.ork, we explore finite-dimensional linear representations of nonlinear dynamical systems by restricting the Koopman operator to an invariant subspace spanned by specially chosen observable functions. The Koopman operator is an infinite-dimensional linear operator that evolves functions of the state of a dynamical system. Dominant terms in the Koopman expansion are typically computed using dynamic mode decomposition (DMD). DMD uses linear measurements of the state variables, and it has recently been shown that this may be too restrictive for nonlinear systems. Choosing the right nonlinear observable functions to form an invariant subspace where it is possible to obtain linear reduced-order models, especially those that are useful for control, is an open challenge. Here, we investigate the choice of observable functions for Koopman analysis that enable the use of optimal linear control techniques on nonlinear problems. First, to include a cost on the state of the system, as in linear quadratic regulator (LQR) control, it is helpful to include these states in the observable subspace, as in DMD. However, we find that this is only possible when there is a single isolated fixed point, as systems with multiple fixed points or more complicated attractors are not globally topologically conjugate to a finite-dimensional linear system, and cannot be represented by a finite-dimensional linear Koopman subspace that includes the state. We then present a data-driven strategy to identify relevant observable functions for Koopman analysis by leveraging a new algorithm to determine relevant terms in a dynamical system by ℓ1-regularized regression of the data in a nonlinear function space; we also show how this algorithm is related to DMD. Finally, we demonstrate the usefulness of nonlinear observable subspaces in the design of Koopman operator optimal control laws for fully nonlinear systems using techniques from linear optimal control.

Towards automatic music transcription: note extraction based on independent subspace analysis

NASA Astrophysics Data System (ADS)

Wellhausen, Jens; Hoynck, Michael

2005-01-01

Due to the increasing amount of music available electronically the need of automatic search, retrieval and classification systems for music becomes more and more important. In this paper an algorithm for automatic transcription of polyphonic piano music into MIDI data is presented, which is a very interesting basis for database applications, music analysis and music classification. The first part of the algorithm performs a note accurate temporal audio segmentation. In the second part, the resulting segments are examined using Independent Subspace Analysis to extract sounding notes. Finally, the results are used to build a MIDI file as a new representation of the piece of music which is examined.

Towards automatic music transcription: note extraction based on independent subspace analysis

NASA Astrophysics Data System (ADS)

Wellhausen, Jens; Höynck, Michael

2004-12-01

Due to the increasing amount of music available electronically the need of automatic search, retrieval and classification systems for music becomes more and more important. In this paper an algorithm for automatic transcription of polyphonic piano music into MIDI data is presented, which is a very interesting basis for database applications, music analysis and music classification. The first part of the algorithm performs a note accurate temporal audio segmentation. In the second part, the resulting segments are examined using Independent Subspace Analysis to extract sounding notes. Finally, the results are used to build a MIDI file as a new representation of the piece of music which is examined.

Subspace aware recovery of low rank and jointly sparse signals

PubMed Central

Biswas, Sampurna; Dasgupta, Soura; Mudumbai, Raghuraman; Jacob, Mathews

2017-01-01

We consider the recovery of a matrix X, which is simultaneously low rank and joint sparse, from few measurements of its columns using a two-step algorithm. Each column of X is measured using a combination of two measurement matrices; one which is the same for every column, while the the second measurement matrix varies from column to column. The recovery proceeds by first estimating the row subspace vectors from the measurements corresponding to the common matrix. The estimated row subspace vectors are then used to recover X from all the measurements using a convex program of joint sparsity minimization. Our main contribution is to provide sufficient conditions on the measurement matrices that guarantee the recovery of such a matrix using the above two-step algorithm. The results demonstrate quite significant savings in number of measurements when compared to the standard multiple measurement vector (MMV) scheme, which assumes same time invariant measurement pattern for all the time frames. We illustrate the impact of the sampling pattern on reconstruction quality using breath held cardiac cine MRI and cardiac perfusion MRI data, while the utility of the algorithm to accelerate the acquisition is demonstrated on MR parameter mapping. PMID:28630889

Self-Taught Low-Rank Coding for Visual Learning.

PubMed

Li, Sheng; Li, Kang; Fu, Yun

2018-03-01

The lack of labeled data presents a common challenge in many computer vision and machine learning tasks. Semisupervised learning and transfer learning methods have been developed to tackle this challenge by utilizing auxiliary samples from the same domain or from a different domain, respectively. Self-taught learning, which is a special type of transfer learning, has fewer restrictions on the choice of auxiliary data. It has shown promising performance in visual learning. However, existing self-taught learning methods usually ignore the structure information in data. In this paper, we focus on building a self-taught coding framework, which can effectively utilize the rich low-level pattern information abstracted from the auxiliary domain, in order to characterize the high-level structural information in the target domain. By leveraging a high quality dictionary learned across auxiliary and target domains, the proposed approach learns expressive codings for the samples in the target domain. Since many types of visual data have been proven to contain subspace structures, a low-rank constraint is introduced into the coding objective to better characterize the structure of the given target set. The proposed representation learning framework is called self-taught low-rank (S-Low) coding, which can be formulated as a nonconvex rank-minimization and dictionary learning problem. We devise an efficient majorization-minimization augmented Lagrange multiplier algorithm to solve it. Based on the proposed S-Low coding mechanism, both unsupervised and supervised visual learning algorithms are derived. Extensive experiments on five benchmark data sets demonstrate the effectiveness of our approach.

Pigments identification of paintings using subspace distance unmixing algorithm

NASA Astrophysics Data System (ADS)

Li, Bin; Lyu, Shuqiang; Zhang, Dafeng; Dong, Qinghao

2018-04-01

In the digital protection of the cultural relics, the identification of the pigment mixtures on the surface of the painting has been the research spot for many years. In this paper, as a hyperspectral unmixing algorithm, sub-space distance unmixing is introduced to solve the problem of recognition of pigments mixture in paintings. Firstly, some mixtures of different pigments are designed to measure their reflectance spectra using spectrometer. Moreover, the factors affecting the unmixing accuracy of pigments' mixtures are discussed. The unmixing results of two cases with and without rice paper and its underlay as endmembers are compared. The experiment results show that the algorithm is able to unmixing the pigments effectively and the unmixing accuracy can be improved after considering the influence of spectra of the rich paper and the underlaying material.

Dictionary Pair Learning on Grassmann Manifolds for Image Denoising.

PubMed

Zeng, Xianhua; Bian, Wei; Liu, Wei; Shen, Jialie; Tao, Dacheng

2015-11-01

Image denoising is a fundamental problem in computer vision and image processing that holds considerable practical importance for real-world applications. The traditional patch-based and sparse coding-driven image denoising methods convert 2D image patches into 1D vectors for further processing. Thus, these methods inevitably break down the inherent 2D geometric structure of natural images. To overcome this limitation pertaining to the previous image denoising methods, we propose a 2D image denoising model, namely, the dictionary pair learning (DPL) model, and we design a corresponding algorithm called the DPL on the Grassmann-manifold (DPLG) algorithm. The DPLG algorithm first learns an initial dictionary pair (i.e., the left and right dictionaries) by employing a subspace partition technique on the Grassmann manifold, wherein the refined dictionary pair is obtained through a sub-dictionary pair merging. The DPLG obtains a sparse representation by encoding each image patch only with the selected sub-dictionary pair. The non-zero elements of the sparse representation are further smoothed by the graph Laplacian operator to remove the noise. Consequently, the DPLG algorithm not only preserves the inherent 2D geometric structure of natural images but also performs manifold smoothing in the 2D sparse coding space. We demonstrate that the DPLG algorithm also improves the structural SIMilarity values of the perceptual visual quality for denoised images using the experimental evaluations on the benchmark images and Berkeley segmentation data sets. Moreover, the DPLG also produces the competitive peak signal-to-noise ratio values from popular image denoising algorithms.

A Class of Manifold Regularized Multiplicative Update Algorithms for Image Clustering.

PubMed

Yang, Shangming; Yi, Zhang; He, Xiaofei; Li, Xuelong

2015-12-01

Multiplicative update algorithms are important tools for information retrieval, image processing, and pattern recognition. However, when the graph regularization is added to the cost function, different classes of sample data may be mapped to the same subspace, which leads to the increase of data clustering error rate. In this paper, an improved nonnegative matrix factorization (NMF) cost function is introduced. Based on the cost function, a class of novel graph regularized NMF algorithms is developed, which results in a class of extended multiplicative update algorithms with manifold structure regularization. Analysis shows that in the learning, the proposed algorithms can efficiently minimize the rank of the data representation matrix. Theoretical results presented in this paper are confirmed by simulations. For different initializations and data sets, variation curves of cost functions and decomposition data are presented to show the convergence features of the proposed update rules. Basis images, reconstructed images, and clustering results are utilized to present the efficiency of the new algorithms. Last, the clustering accuracies of different algorithms are also investigated, which shows that the proposed algorithms can achieve state-of-the-art performance in applications of image clustering.

Semi-Supervised Projective Non-Negative Matrix Factorization for Cancer Classification.

PubMed

Zhang, Xiang; Guan, Naiyang; Jia, Zhilong; Qiu, Xiaogang; Luo, Zhigang

2015-01-01

Advances in DNA microarray technologies have made gene expression profiles a significant candidate in identifying different types of cancers. Traditional learning-based cancer identification methods utilize labeled samples to train a classifier, but they are inconvenient for practical application because labels are quite expensive in the clinical cancer research community. This paper proposes a semi-supervised projective non-negative matrix factorization method (Semi-PNMF) to learn an effective classifier from both labeled and unlabeled samples, thus boosting subsequent cancer classification performance. In particular, Semi-PNMF jointly learns a non-negative subspace from concatenated labeled and unlabeled samples and indicates classes by the positions of the maximum entries of their coefficients. Because Semi-PNMF incorporates statistical information from the large volume of unlabeled samples in the learned subspace, it can learn more representative subspaces and boost classification performance. We developed a multiplicative update rule (MUR) to optimize Semi-PNMF and proved its convergence. The experimental results of cancer classification for two multiclass cancer gene expression profile datasets show that Semi-PNMF outperforms the representative methods.

Current harmonics elimination control method for six-phase PM synchronous motor drives.

PubMed

Yuan, Lei; Chen, Ming-liang; Shen, Jian-qing; Xiao, Fei

2015-11-01

To reduce the undesired 5th and 7th stator harmonic current in the six-phase permanent magnet synchronous motor (PMSM), an improved vector control algorithm was proposed based on vector space decomposition (VSD) transformation method, which can control the fundamental and harmonic subspace separately. To improve the traditional VSD technology, a novel synchronous rotating coordinate transformation matrix was presented in this paper, and only using the traditional PI controller in d-q subspace can meet the non-static difference adjustment, the controller parameter design method is given by employing internal model principle. Moreover, the current PI controller parallel with resonant controller is employed in x-y subspace to realize the specific 5th and 7th harmonic component compensation. In addition, a new six-phase SVPWM algorithm based on VSD transformation theory is also proposed. Simulation and experimental results verify the effectiveness of current decoupling vector controller. Copyright © 2015 ISA. Published by Elsevier Ltd. All rights reserved.

Supervised orthogonal discriminant subspace projects learning for face recognition.

PubMed

Chen, Yu; Xu, Xiao-Hong

2014-02-01

In this paper, a new linear dimension reduction method called supervised orthogonal discriminant subspace projection (SODSP) is proposed, which addresses high-dimensionality of data and the small sample size problem. More specifically, given a set of data points in the ambient space, a novel weight matrix that describes the relationship between the data points is first built. And in order to model the manifold structure, the class information is incorporated into the weight matrix. Based on the novel weight matrix, the local scatter matrix as well as non-local scatter matrix is defined such that the neighborhood structure can be preserved. In order to enhance the recognition ability, we impose an orthogonal constraint into a graph-based maximum margin analysis, seeking to find a projection that maximizes the difference, rather than the ratio between the non-local scatter and the local scatter. In this way, SODSP naturally avoids the singularity problem. Further, we develop an efficient and stable algorithm for implementing SODSP, especially, on high-dimensional data set. Moreover, the theoretical analysis shows that LPP is a special instance of SODSP by imposing some constraints. Experiments on the ORL, Yale, Extended Yale face database B and FERET face database are performed to test and evaluate the proposed algorithm. The results demonstrate the effectiveness of SODSP. Copyright © 2013 Elsevier Ltd. All rights reserved.

Extending Correlation Filter-Based Visual Tracking by Tree-Structured Ensemble and Spatial Windowing.

PubMed

Gundogdu, Erhan; Ozkan, Huseyin; Alatan, A Aydin

2017-11-01

Correlation filters have been successfully used in visual tracking due to their modeling power and computational efficiency. However, the state-of-the-art correlation filter-based (CFB) tracking algorithms tend to quickly discard the previous poses of the target, since they consider only a single filter in their models. On the contrary, our approach is to register multiple CFB trackers for previous poses and exploit the registered knowledge when an appearance change occurs. To this end, we propose a novel tracking algorithm [of complexity O(D) ] based on a large ensemble of CFB trackers. The ensemble [of size O(2 D ) ] is organized over a binary tree (depth D ), and learns the target appearance subspaces such that each constituent tracker becomes an expert of a certain appearance. During tracking, the proposed algorithm combines only the appearance-aware relevant experts to produce boosted tracking decisions. Additionally, we propose a versatile spatial windowing technique to enhance the individual expert trackers. For this purpose, spatial windows are learned for target objects as well as the correlation filters and then the windowed regions are processed for more robust correlations. In our extensive experiments on benchmark datasets, we achieve a substantial performance increase by using the proposed tracking algorithm together with the spatial windowing.

Sparsity-aware tight frame learning with adaptive subspace recognition for multiple fault diagnosis

NASA Astrophysics Data System (ADS)

Zhang, Han; Chen, Xuefeng; Du, Zhaohui; Yang, Boyuan

2017-09-01

It is a challenging problem to design excellent dictionaries to sparsely represent diverse fault information and simultaneously discriminate different fault sources. Therefore, this paper describes and analyzes a novel multiple feature recognition framework which incorporates the tight frame learning technique with an adaptive subspace recognition strategy. The proposed framework consists of four stages. Firstly, by introducing the tight frame constraint into the popular dictionary learning model, the proposed tight frame learning model could be formulated as a nonconvex optimization problem which can be solved by alternatively implementing hard thresholding operation and singular value decomposition. Secondly, the noises are effectively eliminated through transform sparse coding techniques. Thirdly, the denoised signal is decoupled into discriminative feature subspaces by each tight frame filter. Finally, in guidance of elaborately designed fault related sensitive indexes, latent fault feature subspaces can be adaptively recognized and multiple faults are diagnosed simultaneously. Extensive numerical experiments are sequently implemented to investigate the sparsifying capability of the learned tight frame as well as its comprehensive denoising performance. Most importantly, the feasibility and superiority of the proposed framework is verified through performing multiple fault diagnosis of motor bearings. Compared with the state-of-the-art fault detection techniques, some important advantages have been observed: firstly, the proposed framework incorporates the physical prior with the data-driven strategy and naturally multiple fault feature with similar oscillation morphology can be adaptively decoupled. Secondly, the tight frame dictionary directly learned from the noisy observation can significantly promote the sparsity of fault features compared to analytical tight frames. Thirdly, a satisfactory complete signal space description property is guaranteed and thus weak feature leakage problem is avoided compared to typical learning methods.

Detecting coupled collective motions in protein by independent subspace analysis

NASA Astrophysics Data System (ADS)

Sakuraba, Shun; Joti, Yasumasa; Kitao, Akio

2010-11-01

Protein dynamics evolves in a high-dimensional space, comprising aharmonic, strongly correlated motional modes. Such correlation often plays an important role in analyzing protein function. In order to identify significantly correlated collective motions, here we employ independent subspace analysis based on the subspace joint approximate diagonalization of eigenmatrices algorithm for the analysis of molecular dynamics (MD) simulation trajectories. From the 100 ns MD simulation of T4 lysozyme, we extract several independent subspaces in each of which collective modes are significantly correlated, and identify the other modes as independent. This method successfully detects the modes along which long-tailed non-Gaussian probability distributions are obtained. Based on the time cross-correlation analysis, we identified a series of events among domain motions and more localized motions in the protein, indicating the connection between the functionally relevant phenomena which have been independently revealed by experiments.

Subspace projection method for unstructured searches with noisy quantum oracles using a signal-based quantum emulation device

NASA Astrophysics Data System (ADS)

La Cour, Brian R.; Ostrove, Corey I.

2017-01-01

This paper describes a novel approach to solving unstructured search problems using a classical, signal-based emulation of a quantum computer. The classical nature of the representation allows one to perform subspace projections in addition to the usual unitary gate operations. Although bandwidth requirements will limit the scale of problems that can be solved by this method, it can nevertheless provide a significant computational advantage for problems of limited size. In particular, we find that, for the same number of noisy oracle calls, the proposed subspace projection method provides a higher probability of success for finding a solution than does an single application of Grover's algorithm on the same device.

A comparative intelligibility study of single-microphone noise reduction algorithms.

PubMed

Hu, Yi; Loizou, Philipos C

2007-09-01

The evaluation of intelligibility of noise reduction algorithms is reported. IEEE sentences and consonants were corrupted by four types of noise including babble, car, street and train at two signal-to-noise ratio levels (0 and 5 dB), and then processed by eight speech enhancement methods encompassing four classes of algorithms: spectral subtractive, sub-space, statistical model based and Wiener-type algorithms. The enhanced speech was presented to normal-hearing listeners for identification. With the exception of a single noise condition, no algorithm produced significant improvements in speech intelligibility. Information transmission analysis of the consonant confusion matrices indicated that no algorithm improved significantly the place feature score, significantly, which is critically important for speech recognition. The algorithms which were found in previous studies to perform the best in terms of overall quality, were not the same algorithms that performed the best in terms of speech intelligibility. The subspace algorithm, for instance, was previously found to perform the worst in terms of overall quality, but performed well in the present study in terms of preserving speech intelligibility. Overall, the analysis of consonant confusion matrices suggests that in order for noise reduction algorithms to improve speech intelligibility, they need to improve the place and manner feature scores.

Collaboration space division in collaborative product development based on a genetic algorithm

NASA Astrophysics Data System (ADS)

Qian, Xueming; Ma, Yanqiao; Feng, Huan

2018-02-01

The advance in the global environment, rapidly changing markets, and information technology has created a new stage for design. In such an environment, one strategy for success is the Collaborative Product Development (CPD). Organizing people effectively is the goal of Collaborative Product Development, and it solves the problem with certain foreseeability. The development group activities are influenced not only by the methods and decisions available, but also by correlation among personnel. Grouping the personnel according to their correlation intensity is defined as collaboration space division (CSD). Upon establishment of a correlation matrix (CM) of personnel and an analysis of the collaboration space, the genetic algorithm (GA) and minimum description length (MDL) principle may be used as tools in optimizing collaboration space. The MDL principle is used in setting up an object function, and the GA is used as a methodology. The algorithm encodes spatial information as a chromosome in binary. After repetitious crossover, mutation, selection and multiplication, a robust chromosome is found, which can be decoded into an optimal collaboration space. This new method can calculate the members in sub-spaces and individual groupings within the staff. Furthermore, the intersection of sub-spaces and public persons belonging to all sub-spaces can be determined simultaneously.

Skin subspace color modeling for daytime and nighttime group activity recognition in confined operational spaces

NASA Astrophysics Data System (ADS)

Shirkhodaie, Amir; Poshtyar, Azin; Chan, Alex; Hu, Shuowen

2016-05-01

In many military and homeland security persistent surveillance applications, accurate detection of different skin colors in varying observability and illumination conditions is a valuable capability for video analytics. One of those applications is In-Vehicle Group Activity (IVGA) recognition, in which significant changes in observability and illumination may occur during the course of a specific human group activity of interest. Most of the existing skin color detection algorithms, however, are unable to perform satisfactorily in confined operational spaces with partial observability and occultation, as well as under diverse and changing levels of illumination intensity, reflection, and diffraction. In this paper, we investigate the salient features of ten popular color spaces for skin subspace color modeling. More specifically, we examine the advantages and disadvantages of each of these color spaces, as well as the stability and suitability of their features in differentiating skin colors under various illumination conditions. The salient features of different color subspaces are methodically discussed and graphically presented. Furthermore, we present robust and adaptive algorithms for skin color detection based on this analysis. Through examples, we demonstrate the efficiency and effectiveness of these new color skin detection algorithms and discuss their applicability for skin detection in IVGA recognition applications.

A Robust Sound Source Localization Approach for Microphone Array with Model Errors

NASA Astrophysics Data System (ADS)

Xiao, Hua; Shao, Huai-Zong; Peng, Qi-Cong

In this paper, a robust sound source localization approach is proposed. The approach retains good performance even when model errors exist. Compared with previous work in this field, the contributions of this paper are as follows. First, an improved broad-band and near-field array model is proposed. It takes array gain, phase perturbations into account and is based on the actual positions of the elements. It can be used in arbitrary planar geometry arrays. Second, a subspace model errors estimation algorithm and a Weighted 2-Dimension Multiple Signal Classification (W2D-MUSIC) algorithm are proposed. The subspace model errors estimation algorithm estimates unknown parameters of the array model, i. e., gain, phase perturbations, and positions of the elements, with high accuracy. The performance of this algorithm is improved with the increasing of SNR or number of snapshots. The W2D-MUSIC algorithm based on the improved array model is implemented to locate sound sources. These two algorithms compose the robust sound source approach. The more accurate steering vectors can be provided for further processing such as adaptive beamforming algorithm. Numerical examples confirm effectiveness of this proposed approach.

Koopman Invariant Subspaces and Finite Linear Representations of Nonlinear Dynamical Systems for Control

PubMed Central

Brunton, Steven L.; Brunton, Bingni W.; Proctor, Joshua L.; Kutz, J. Nathan

2016-01-01

In this work, we explore finite-dimensional linear representations of nonlinear dynamical systems by restricting the Koopman operator to an invariant subspace spanned by specially chosen observable functions. The Koopman operator is an infinite-dimensional linear operator that evolves functions of the state of a dynamical system. Dominant terms in the Koopman expansion are typically computed using dynamic mode decomposition (DMD). DMD uses linear measurements of the state variables, and it has recently been shown that this may be too restrictive for nonlinear systems. Choosing the right nonlinear observable functions to form an invariant subspace where it is possible to obtain linear reduced-order models, especially those that are useful for control, is an open challenge. Here, we investigate the choice of observable functions for Koopman analysis that enable the use of optimal linear control techniques on nonlinear problems. First, to include a cost on the state of the system, as in linear quadratic regulator (LQR) control, it is helpful to include these states in the observable subspace, as in DMD. However, we find that this is only possible when there is a single isolated fixed point, as systems with multiple fixed points or more complicated attractors are not globally topologically conjugate to a finite-dimensional linear system, and cannot be represented by a finite-dimensional linear Koopman subspace that includes the state. We then present a data-driven strategy to identify relevant observable functions for Koopman analysis by leveraging a new algorithm to determine relevant terms in a dynamical system by ℓ1-regularized regression of the data in a nonlinear function space; we also show how this algorithm is related to DMD. Finally, we demonstrate the usefulness of nonlinear observable subspaces in the design of Koopman operator optimal control laws for fully nonlinear systems using techniques from linear optimal control. PMID:26919740

Evaluating the utility of mid-infrared spectral subspaces for predicting soil properties.

PubMed

Sila, Andrew M; Shepherd, Keith D; Pokhariyal, Ganesh P

2016-04-15

We propose four methods for finding local subspaces in large spectral libraries. The proposed four methods include (a) cosine angle spectral matching; (b) hit quality index spectral matching; (c) self-organizing maps and (d) archetypal analysis methods. Then evaluate prediction accuracies for global and subspaces calibration models. These methods were tested on a mid-infrared spectral library containing 1907 soil samples collected from 19 different countries under the Africa Soil Information Service project. Calibration models for pH, Mehlich-3 Ca, Mehlich-3 Al, total carbon and clay soil properties were developed for the whole library and for the subspace. Root mean square error of prediction was used to evaluate predictive performance of subspace and global models. The root mean square error of prediction was computed using a one-third-holdout validation set. Effect of pretreating spectra with different methods was tested for 1st and 2nd derivative Savitzky-Golay algorithm, multiplicative scatter correction, standard normal variate and standard normal variate followed by detrending methods. In summary, the results show that global models outperformed the subspace models. We, therefore, conclude that global models are more accurate than the local models except in few cases. For instance, sand and clay root mean square error values from local models from archetypal analysis method were 50% poorer than the global models except for subspace models obtained using multiplicative scatter corrected spectra with which were 12% better. However, the subspace approach provides novel methods for discovering data pattern that may exist in large spectral libraries.

Non-Cooperative Target Recognition by Means of Singular Value Decomposition Applied to Radar High Resolution Range Profiles †

PubMed Central

López-Rodríguez, Patricia; Escot-Bocanegra, David; Fernández-Recio, Raúl; Bravo, Ignacio

2015-01-01

Radar high resolution range profiles are widely used among the target recognition community for the detection and identification of flying targets. In this paper, singular value decomposition is applied to extract the relevant information and to model each aircraft as a subspace. The identification algorithm is based on angle between subspaces and takes place in a transformed domain. In order to have a wide database of radar signatures and evaluate the performance, simulated range profiles are used as the recognition database while the test samples comprise data of actual range profiles collected in a measurement campaign. Thanks to the modeling of aircraft as subspaces only the valuable information of each target is used in the recognition process. Thus, one of the main advantages of using singular value decomposition, is that it helps to overcome the notable dissimilarities found in the shape and signal-to-noise ratio between actual and simulated profiles due to their difference in nature. Despite these differences, the recognition rates obtained with the algorithm are quite promising. PMID:25551484

A Tensor-Based Subspace Approach for Bistatic MIMO Radar in Spatial Colored Noise

PubMed Central

Wang, Xianpeng; Wang, Wei; Li, Xin; Wang, Junxiang

2014-01-01

In this paper, a new tensor-based subspace approach is proposed to estimate the direction of departure (DOD) and the direction of arrival (DOA) for bistatic multiple-input multiple-output (MIMO) radar in the presence of spatial colored noise. Firstly, the received signals can be packed into a third-order measurement tensor by exploiting the inherent structure of the matched filter. Then, the measurement tensor can be divided into two sub-tensors, and a cross-covariance tensor is formulated to eliminate the spatial colored noise. Finally, the signal subspace is constructed by utilizing the higher-order singular value decomposition (HOSVD) of the cross-covariance tensor, and the DOD and DOA can be obtained through the estimation of signal parameters via rotational invariance technique (ESPRIT) algorithm, which are paired automatically. Since the multidimensional inherent structure and the cross-covariance tensor technique are used, the proposed method provides better angle estimation performance than Chen's method, the ESPRIT algorithm and the multi-SVD method. Simulation results confirm the effectiveness and the advantage of the proposed method. PMID:24573313

A tensor-based subspace approach for bistatic MIMO radar in spatial colored noise.

PubMed

Wang, Xianpeng; Wang, Wei; Li, Xin; Wang, Junxiang

2014-02-25

In this paper, a new tensor-based subspace approach is proposed to estimate the direction of departure (DOD) and the direction of arrival (DOA) for bistatic multiple-input multiple-output (MIMO) radar in the presence of spatial colored noise. Firstly, the received signals can be packed into a third-order measurement tensor by exploiting the inherent structure of the matched filter. Then, the measurement tensor can be divided into two sub-tensors, and a cross-covariance tensor is formulated to eliminate the spatial colored noise. Finally, the signal subspace is constructed by utilizing the higher-order singular value decomposition (HOSVD) of the cross-covariance tensor, and the DOD and DOA can be obtained through the estimation of signal parameters via rotational invariance technique (ESPRIT) algorithm, which are paired automatically. Since the multidimensional inherent structure and the cross-covariance tensor technique are used, the proposed method provides better angle estimation performance than Chen's method, the ESPRIT algorithm and the multi-SVD method. Simulation results confirm the effectiveness and the advantage of the proposed method.

Inverse transport calculations in optical imaging with subspace optimization algorithms

DOE Office of Scientific and Technical Information (OSTI.GOV)

Ding, Tian, E-mail: tding@math.utexas.edu; Ren, Kui, E-mail: ren@math.utexas.edu

2014-09-15

Inverse boundary value problems for the radiative transport equation play an important role in optics-based medical imaging techniques such as diffuse optical tomography (DOT) and fluorescence optical tomography (FOT). Despite the rapid progress in the mathematical theory and numerical computation of these inverse problems in recent years, developing robust and efficient reconstruction algorithms remains a challenging task and an active research topic. We propose here a robust reconstruction method that is based on subspace minimization techniques. The method splits the unknown transport solution (or a functional of it) into low-frequency and high-frequency components, and uses singular value decomposition to analyticallymore » recover part of low-frequency information. Minimization is then applied to recover part of the high-frequency components of the unknowns. We present some numerical simulations with synthetic data to demonstrate the performance of the proposed algorithm.« less

Automated computation of autonomous spectral submanifolds for nonlinear modal analysis

NASA Astrophysics Data System (ADS)

Ponsioen, Sten; Pedergnana, Tiemo; Haller, George

2018-04-01

We discuss an automated computational methodology for computing two-dimensional spectral submanifolds (SSMs) in autonomous nonlinear mechanical systems of arbitrary degrees of freedom. In our algorithm, SSMs, the smoothest nonlinear continuations of modal subspaces of the linearized system, are constructed up to arbitrary orders of accuracy, using the parameterization method. An advantage of this approach is that the construction of the SSMs does not break down when the SSM folds over its underlying spectral subspace. A further advantage is an automated a posteriori error estimation feature that enables a systematic increase in the orders of the SSM computation until the required accuracy is reached. We find that the present algorithm provides a major speed-up, relative to numerical continuation methods, in the computation of backbone curves, especially in higher-dimensional problems. We illustrate the accuracy and speed of the automated SSM algorithm on lower- and higher-dimensional mechanical systems.

A multifaceted independent performance analysis of facial subspace recognition algorithms.

PubMed

Bajwa, Usama Ijaz; Taj, Imtiaz Ahmad; Anwar, Muhammad Waqas; Wang, Xuan

2013-01-01

Face recognition has emerged as the fastest growing biometric technology and has expanded a lot in the last few years. Many new algorithms and commercial systems have been proposed and developed. Most of them use Principal Component Analysis (PCA) as a base for their techniques. Different and even conflicting results have been reported by researchers comparing these algorithms. The purpose of this study is to have an independent comparative analysis considering both performance and computational complexity of six appearance based face recognition algorithms namely PCA, 2DPCA, A2DPCA, (2D)(2)PCA, LPP and 2DLPP under equal working conditions. This study was motivated due to the lack of unbiased comprehensive comparative analysis of some recent subspace methods with diverse distance metric combinations. For comparison with other studies, FERET, ORL and YALE databases have been used with evaluation criteria as of FERET evaluations which closely simulate real life scenarios. A comparison of results with previous studies is performed and anomalies are reported. An important contribution of this study is that it presents the suitable performance conditions for each of the algorithms under consideration.

An implementation of the QMR method based on coupled two-term recurrences

NASA Technical Reports Server (NTRS)

Freund, Roland W.; Nachtigal, Noeel M.

1992-01-01

The authors have proposed a new Krylov subspace iteration, the quasi-minimal residual algorithm (QMR), for solving non-Hermitian linear systems. In the original implementation of the QMR method, the Lanczos process with look-ahead is used to generate basis vectors for the underlying Krylov subspaces. In the Lanczos algorithm, these basis vectors are computed by means of three-term recurrences. It has been observed that, in finite precision arithmetic, vector iterations based on three-term recursions are usually less robust than mathematically equivalent coupled two-term vector recurrences. This paper presents a look-ahead algorithm that constructs the Lanczos basis vectors by means of coupled two-term recursions. Implementation details are given, and the look-ahead strategy is described. A new implementation of the QMR method, based on this coupled two-term algorithm, is described. A simplified version of the QMR algorithm without look-ahead is also presented, and the special case of QMR for complex symmetric linear systems is considered. Results of numerical experiments comparing the original and the new implementations of the QMR method are reported.

Performance Analysis of Blind Subspace-Based Signature Estimation Algorithms for DS-CDMA Systems with Unknown Correlated Noise

NASA Astrophysics Data System (ADS)

Zarifi, Keyvan; Gershman, Alex B.

2006-12-01

We analyze the performance of two popular blind subspace-based signature waveform estimation techniques proposed by Wang and Poor and Buzzi and Poor for direct-sequence code division multiple-access (DS-CDMA) systems with unknown correlated noise. Using the first-order perturbation theory, analytical expressions for the mean-square error (MSE) of these algorithms are derived. We also obtain simple high SNR approximations of the MSE expressions which explicitly clarify how the performance of these techniques depends on the environmental parameters and how it is related to that of the conventional techniques that are based on the standard white noise assumption. Numerical examples further verify the consistency of the obtained analytical results with simulation results.

Evolutionary-inspired probabilistic search for enhancing sampling of local minima in the protein energy surface

PubMed Central

2012-01-01

Background Despite computational challenges, elucidating conformations that a protein system assumes under physiologic conditions for the purpose of biological activity is a central problem in computational structural biology. While these conformations are associated with low energies in the energy surface that underlies the protein conformational space, few existing conformational search algorithms focus on explicitly sampling low-energy local minima in the protein energy surface. Methods This work proposes a novel probabilistic search framework, PLOW, that explicitly samples low-energy local minima in the protein energy surface. The framework combines algorithmic ingredients from evolutionary computation and computational structural biology to effectively explore the subspace of local minima. A greedy local search maps a conformation sampled in conformational space to a nearby local minimum. A perturbation move jumps out of a local minimum to obtain a new starting conformation for the greedy local search. The process repeats in an iterative fashion, resulting in a trajectory-based exploration of the subspace of local minima. Results and conclusions The analysis of PLOW's performance shows that, by navigating only the subspace of local minima, PLOW is able to sample conformations near a protein's native structure, either more effectively or as well as state-of-the-art methods that focus on reproducing the native structure for a protein system. Analysis of the actual subspace of local minima shows that PLOW samples this subspace more effectively that a naive sampling approach. Additional theoretical analysis reveals that the perturbation function employed by PLOW is key to its ability to sample a diverse set of low-energy conformations. This analysis also suggests directions for further research and novel applications for the proposed framework. PMID:22759582

Tensor Rank Preserving Discriminant Analysis for Facial Recognition.

PubMed

Tao, Dapeng; Guo, Yanan; Li, Yaotang; Gao, Xinbo

2017-10-12

Facial recognition, one of the basic topics in computer vision and pattern recognition, has received substantial attention in recent years. However, for those traditional facial recognition algorithms, the facial images are reshaped to a long vector, thereby losing part of the original spatial constraints of each pixel. In this paper, a new tensor-based feature extraction algorithm termed tensor rank preserving discriminant analysis (TRPDA) for facial image recognition is proposed; the proposed method involves two stages: in the first stage, the low-dimensional tensor subspace of the original input tensor samples was obtained; in the second stage, discriminative locality alignment was utilized to obtain the ultimate vector feature representation for subsequent facial recognition. On the one hand, the proposed TRPDA algorithm fully utilizes the natural structure of the input samples, and it applies an optimization criterion that can directly handle the tensor spectral analysis problem, thereby decreasing the computation cost compared those traditional tensor-based feature selection algorithms. On the other hand, the proposed TRPDA algorithm extracts feature by finding a tensor subspace that preserves most of the rank order information of the intra-class input samples. Experiments on the three facial databases are performed here to determine the effectiveness of the proposed TRPDA algorithm.

Optimal Wavelengths Selection Using Hierarchical Evolutionary Algorithm for Prediction of Firmness and Soluble Solids Content in Apples

USDA-ARS?s Scientific Manuscript database

Hyperspectral scattering is a promising technique for rapid and noninvasive measurement of multiple quality attributes of apple fruit. A hierarchical evolutionary algorithm (HEA) approach, in combination with subspace decomposition and partial least squares (PLS) regression, was proposed to select o...

NASA Astrophysics Data System (ADS)

2017-10-01

Different possible selections of the weighting matrices W1 and W2 will lead to different methods of subspace identification. In this study, the robust combined algorithm proposed by Van Overschee and De Moor [24] has been employed. In this algorithm, the weighting matrixes are chosen to be W1 = I and W2 =ΠUf⊥.

Multiple fault separation and detection by joint subspace learning for the health assessment of wind turbine gearboxes

NASA Astrophysics Data System (ADS)

Du, Zhaohui; Chen, Xuefeng; Zhang, Han; Zi, Yanyang; Yan, Ruqiang

2017-09-01

The gearbox of a wind turbine (WT) has dominant failure rates and highest downtime loss among all WT subsystems. Thus, gearbox health assessment for maintenance cost reduction is of paramount importance. The concurrence of multiple faults in gearbox components is a common phenomenon due to fault induction mechanism. This problem should be considered before planning to replace the components of the WT gearbox. Therefore, the key fault patterns should be reliably identified from noisy observation data for the development of an effective maintenance strategy. However, most of the existing studies focusing on multiple fault diagnosis always suffer from inappropriate division of fault information in order to satisfy various rigorous decomposition principles or statistical assumptions, such as the smooth envelope principle of ensemble empirical mode decomposition and the mutual independence assumption of independent component analysis. Thus, this paper presents a joint subspace learning-based multiple fault detection (JSL-MFD) technique to construct different subspaces adaptively for different fault patterns. Its main advantage is its capability to learn multiple fault subspaces directly from the observation signal itself. It can also sparsely concentrate the feature information into a few dominant subspace coefficients. Furthermore, it can eliminate noise by simply performing coefficient shrinkage operations. Consequently, multiple fault patterns are reliably identified by utilizing the maximum fault information criterion. The superiority of JSL-MFD in multiple fault separation and detection is comprehensively investigated and verified by the analysis of a data set of a 750 kW WT gearbox. Results show that JSL-MFD is superior to a state-of-the-art technique in detecting hidden fault patterns and enhancing detection accuracy.

Inverse regression-based uncertainty quantification algorithms for high-dimensional models: Theory and practice

NASA Astrophysics Data System (ADS)

Li, Weixuan; Lin, Guang; Li, Bing

2016-09-01

Many uncertainty quantification (UQ) approaches suffer from the curse of dimensionality, that is, their computational costs become intractable for problems involving a large number of uncertainty parameters. In these situations, the classic Monte Carlo often remains the preferred method of choice because its convergence rate O (n - 1 / 2), where n is the required number of model simulations, does not depend on the dimension of the problem. However, many high-dimensional UQ problems are intrinsically low-dimensional, because the variation of the quantity of interest (QoI) is often caused by only a few latent parameters varying within a low-dimensional subspace, known as the sufficient dimension reduction (SDR) subspace in the statistics literature. Motivated by this observation, we propose two inverse regression-based UQ algorithms (IRUQ) for high-dimensional problems. Both algorithms use inverse regression to convert the original high-dimensional problem to a low-dimensional one, which is then efficiently solved by building a response surface for the reduced model, for example via the polynomial chaos expansion. The first algorithm, which is for the situations where an exact SDR subspace exists, is proved to converge at rate O (n-1), hence much faster than MC. The second algorithm, which doesn't require an exact SDR, employs the reduced model as a control variate to reduce the error of the MC estimate. The accuracy gain could still be significant, depending on how well the reduced model approximates the original high-dimensional one. IRUQ also provides several additional practical advantages: it is non-intrusive; it does not require computing the high-dimensional gradient of the QoI; and it reports an error bar so the user knows how reliable the result is.

Blind source separation and localization using microphone arrays

NASA Astrophysics Data System (ADS)

Sun, Longji

The blind source separation and localization problem for audio signals is studied using microphone arrays. Pure delay mixtures of source signals typically encountered in outdoor environments are considered. Our proposed approach utilizes the subspace methods, including multiple signal classification (MUSIC) and estimation of signal parameters via rotational invariance techniques (ESPRIT) algorithms, to estimate the directions of arrival (DOAs) of the sources from the collected mixtures. Since audio signals are generally considered broadband, the DOA estimates at frequencies with the large sum of squared amplitude values are combined to obtain the final DOA estimates. Using the estimated DOAs, the corresponding mixing and demixing matrices are computed, and the source signals are recovered using the inverse short time Fourier transform. Subspace methods take advantage of the spatial covariance matrix of the collected mixtures to achieve robustness to noise. While the subspace methods have been studied for localizing radio frequency signals, audio signals have their special properties. For instance, they are nonstationary, naturally broadband and analog. All of these make the separation and localization for the audio signals more challenging. Moreover, our algorithm is essentially equivalent to the beamforming technique, which suppresses the signals in unwanted directions and only recovers the signals in the estimated DOAs. Several crucial issues related to our algorithm and their solutions have been discussed, including source number estimation, spatial aliasing, artifact filtering, different ways of mixture generation, and source coordinate estimation using multiple arrays. Additionally, comprehensive simulations and experiments have been conducted to examine various aspects of the algorithm. Unlike the existing blind source separation and localization methods, which are generally time consuming, our algorithm needs signal mixtures of only a short duration and therefore supports real-time implementation.

Randomized subspace-based robust principal component analysis for hyperspectral anomaly detection

NASA Astrophysics Data System (ADS)

Sun, Weiwei; Yang, Gang; Li, Jialin; Zhang, Dianfa

2018-01-01

A randomized subspace-based robust principal component analysis (RSRPCA) method for anomaly detection in hyperspectral imagery (HSI) is proposed. The RSRPCA combines advantages of randomized column subspace and robust principal component analysis (RPCA). It assumes that the background has low-rank properties, and the anomalies are sparse and do not lie in the column subspace of the background. First, RSRPCA implements random sampling to sketch the original HSI dataset from columns and to construct a randomized column subspace of the background. Structured random projections are also adopted to sketch the HSI dataset from rows. Sketching from columns and rows could greatly reduce the computational requirements of RSRPCA. Second, the RSRPCA adopts the columnwise RPCA (CWRPCA) to eliminate negative effects of sampled anomaly pixels and that purifies the previous randomized column subspace by removing sampled anomaly columns. The CWRPCA decomposes the submatrix of the HSI data into a low-rank matrix (i.e., background component), a noisy matrix (i.e., noise component), and a sparse anomaly matrix (i.e., anomaly component) with only a small proportion of nonzero columns. The algorithm of inexact augmented Lagrange multiplier is utilized to optimize the CWRPCA problem and estimate the sparse matrix. Nonzero columns of the sparse anomaly matrix point to sampled anomaly columns in the submatrix. Third, all the pixels are projected onto the complemental subspace of the purified randomized column subspace of the background and the anomaly pixels in the original HSI data are finally exactly located. Several experiments on three real hyperspectral images are carefully designed to investigate the detection performance of RSRPCA, and the results are compared with four state-of-the-art methods. Experimental results show that the proposed RSRPCA outperforms four comparison methods both in detection performance and in computational time.

Subspace methods for identification of human ankle joint stiffness.

PubMed

Zhao, Y; Westwick, D T; Kearney, R E

2011-11-01

Joint stiffness, the dynamic relationship between the angular position of a joint and the torque acting about it, describes the dynamic, mechanical behavior of a joint during posture and movement. Joint stiffness arises from both intrinsic and reflex mechanisms, but the torques due to these mechanisms cannot be measured separately experimentally, since they appear and change together. Therefore, the direct estimation of the intrinsic and reflex stiffnesses is difficult. In this paper, we present a new, two-step procedure to estimate the intrinsic and reflex components of ankle stiffness. In the first step, a discrete-time, subspace-based method is used to estimate a state-space model for overall stiffness from the measured overall torque and then predict the intrinsic and reflex torques. In the second step, continuous-time models for the intrinsic and reflex stiffnesses are estimated from the predicted intrinsic and reflex torques. Simulations and experimental results demonstrate that the algorithm estimates the intrinsic and reflex stiffnesses accurately. The new subspace-based algorithm has three advantages over previous algorithms: 1) It does not require iteration, and therefore, will always converge to an optimal solution; 2) it provides better estimates for data with high noise or short sample lengths; and 3) it provides much more accurate results for data acquired under the closed-loop conditions, that prevail when subjects interact with compliant loads.

A computationally efficient parallel Levenberg-Marquardt algorithm for highly parameterized inverse model analyses

NASA Astrophysics Data System (ADS)

Lin, Youzuo; O'Malley, Daniel; Vesselinov, Velimir V.

2016-09-01

Inverse modeling seeks model parameters given a set of observations. However, for practical problems because the number of measurements is often large and the model parameters are also numerous, conventional methods for inverse modeling can be computationally expensive. We have developed a new, computationally efficient parallel Levenberg-Marquardt method for solving inverse modeling problems with a highly parameterized model space. Levenberg-Marquardt methods require the solution of a linear system of equations which can be prohibitively expensive to compute for moderate to large-scale problems. Our novel method projects the original linear problem down to a Krylov subspace such that the dimensionality of the problem can be significantly reduced. Furthermore, we store the Krylov subspace computed when using the first damping parameter and recycle the subspace for the subsequent damping parameters. The efficiency of our new inverse modeling algorithm is significantly improved using these computational techniques. We apply this new inverse modeling method to invert for random transmissivity fields in 2-D and a random hydraulic conductivity field in 3-D. Our algorithm is fast enough to solve for the distributed model parameters (transmissivity) in the model domain. The algorithm is coded in Julia and implemented in the MADS computational framework (http://mads.lanl.gov). By comparing with Levenberg-Marquardt methods using standard linear inversion techniques such as QR or SVD methods, our Levenberg-Marquardt method yields a speed-up ratio on the order of ˜101 to ˜102 in a multicore computational environment. Therefore, our new inverse modeling method is a powerful tool for characterizing subsurface heterogeneity for moderate to large-scale problems.

Geometric MCMC for infinite-dimensional inverse problems

NASA Astrophysics Data System (ADS)

Beskos, Alexandros; Girolami, Mark; Lan, Shiwei; Farrell, Patrick E.; Stuart, Andrew M.

2017-04-01

Bayesian inverse problems often involve sampling posterior distributions on infinite-dimensional function spaces. Traditional Markov chain Monte Carlo (MCMC) algorithms are characterized by deteriorating mixing times upon mesh-refinement, when the finite-dimensional approximations become more accurate. Such methods are typically forced to reduce step-sizes as the discretization gets finer, and thus are expensive as a function of dimension. Recently, a new class of MCMC methods with mesh-independent convergence times has emerged. However, few of them take into account the geometry of the posterior informed by the data. At the same time, recently developed geometric MCMC algorithms have been found to be powerful in exploring complicated distributions that deviate significantly from elliptic Gaussian laws, but are in general computationally intractable for models defined in infinite dimensions. In this work, we combine geometric methods on a finite-dimensional subspace with mesh-independent infinite-dimensional approaches. Our objective is to speed up MCMC mixing times, without significantly increasing the computational cost per step (for instance, in comparison with the vanilla preconditioned Crank-Nicolson (pCN) method). This is achieved by using ideas from geometric MCMC to probe the complex structure of an intrinsic finite-dimensional subspace where most data information concentrates, while retaining robust mixing times as the dimension grows by using pCN-like methods in the complementary subspace. The resulting algorithms are demonstrated in the context of three challenging inverse problems arising in subsurface flow, heat conduction and incompressible flow control. The algorithms exhibit up to two orders of magnitude improvement in sampling efficiency when compared with the pCN method.

Subspace Clustering via Learning an Adaptive Low-Rank Graph.

PubMed

Yin, Ming; Xie, Shengli; Wu, Zongze; Zhang, Yun; Gao, Junbin

2018-08-01

By using a sparse representation or low-rank representation of data, the graph-based subspace clustering has recently attracted considerable attention in computer vision, given its capability and efficiency in clustering data. However, the graph weights built using the representation coefficients are not the exact ones as the traditional definition is in a deterministic way. The two steps of representation and clustering are conducted in an independent manner, thus an overall optimal result cannot be guaranteed. Furthermore, it is unclear how the clustering performance will be affected by using this graph. For example, the graph parameters, i.e., the weights on edges, have to be artificially pre-specified while it is very difficult to choose the optimum. To this end, in this paper, a novel subspace clustering via learning an adaptive low-rank graph affinity matrix is proposed, where the affinity matrix and the representation coefficients are learned in a unified framework. As such, the pre-computed graph regularizer is effectively obviated and better performance can be achieved. Experimental results on several famous databases demonstrate that the proposed method performs better against the state-of-the-art approaches, in clustering.

Factor analysis of auto-associative neural networks with application in speaker verification.

PubMed

Garimella, Sri; Hermansky, Hynek

2013-04-01

Auto-associative neural network (AANN) is a fully connected feed-forward neural network, trained to reconstruct its input at its output through a hidden compression layer, which has fewer numbers of nodes than the dimensionality of input. AANNs are used to model speakers in speaker verification, where a speaker-specific AANN model is obtained by adapting (or retraining) the universal background model (UBM) AANN, an AANN trained on multiple held out speakers, using corresponding speaker data. When the amount of speaker data is limited, this adaptation procedure may lead to overfitting as all the parameters of UBM-AANN are adapted. In this paper, we introduce and develop the factor analysis theory of AANNs to alleviate this problem. We hypothesize that only the weight matrix connecting the last nonlinear hidden layer and the output layer is speaker-specific, and further restrict it to a common low-dimensional subspace during adaptation. The subspace is learned using large amounts of development data, and is held fixed during adaptation. Thus, only the coordinates in a subspace, also known as i-vector, need to be estimated using speaker-specific data. The update equations are derived for learning both the common low-dimensional subspace and the i-vectors corresponding to speakers in the subspace. The resultant i-vector representation is used as a feature for the probabilistic linear discriminant analysis model. The proposed system shows promising results on the NIST-08 speaker recognition evaluation (SRE), and yields a 23% relative improvement in equal error rate over the previously proposed weighted least squares-based subspace AANNs system. The experiments on NIST-10 SRE confirm that these improvements are consistent and generalize across datasets.

Robust Framework to Combine Diverse Classifiers Assigning Distributed Confidence to Individual Classifiers at Class Level

PubMed Central

Arshad, Sannia; Rho, Seungmin

2014-01-01

We have presented a classification framework that combines multiple heterogeneous classifiers in the presence of class label noise. An extension of m-Mediods based modeling is presented that generates model of various classes whilst identifying and filtering noisy training data. This noise free data is further used to learn model for other classifiers such as GMM and SVM. A weight learning method is then introduced to learn weights on each class for different classifiers to construct an ensemble. For this purpose, we applied genetic algorithm to search for an optimal weight vector on which classifier ensemble is expected to give the best accuracy. The proposed approach is evaluated on variety of real life datasets. It is also compared with existing standard ensemble techniques such as Adaboost, Bagging, and Random Subspace Methods. Experimental results show the superiority of proposed ensemble method as compared to its competitors, especially in the presence of class label noise and imbalance classes. PMID:25295302

Robust framework to combine diverse classifiers assigning distributed confidence to individual classifiers at class level.

PubMed

Khalid, Shehzad; Arshad, Sannia; Jabbar, Sohail; Rho, Seungmin

2014-01-01

We have presented a classification framework that combines multiple heterogeneous classifiers in the presence of class label noise. An extension of m-Mediods based modeling is presented that generates model of various classes whilst identifying and filtering noisy training data. This noise free data is further used to learn model for other classifiers such as GMM and SVM. A weight learning method is then introduced to learn weights on each class for different classifiers to construct an ensemble. For this purpose, we applied genetic algorithm to search for an optimal weight vector on which classifier ensemble is expected to give the best accuracy. The proposed approach is evaluated on variety of real life datasets. It is also compared with existing standard ensemble techniques such as Adaboost, Bagging, and Random Subspace Methods. Experimental results show the superiority of proposed ensemble method as compared to its competitors, especially in the presence of class label noise and imbalance classes.

Fault-tolerant optimised tracking control for unknown discrete-time linear systems using a combined reinforcement learning and residual compensation methodology

NASA Astrophysics Data System (ADS)

Han, Ke-Zhen; Feng, Jian; Cui, Xiaohong

2017-10-01

This paper considers the fault-tolerant optimised tracking control (FTOTC) problem for unknown discrete-time linear system. A research scheme is proposed on the basis of data-based parity space identification, reinforcement learning and residual compensation techniques. The main characteristic of this research scheme lies in the parity-space-identification-based simultaneous tracking control and residual compensation. The specific technical line consists of four main contents: apply subspace aided method to design observer-based residual generator; use reinforcement Q-learning approach to solve optimised tracking control policy; rely on robust H∞ theory to achieve noise attenuation; adopt fault estimation triggered by residual generator to perform fault compensation. To clarify the design and implementation procedures, an integrated algorithm is further constructed to link up these four functional units. The detailed analysis and proof are subsequently given to explain the guaranteed FTOTC performance of the proposed conclusions. Finally, a case simulation is provided to verify its effectiveness.

Spectral-Spatial Shared Linear Regression for Hyperspectral Image Classification.

PubMed

Haoliang Yuan; Yuan Yan Tang

2017-04-01

Classification of the pixels in hyperspectral image (HSI) is an important task and has been popularly applied in many practical applications. Its major challenge is the high-dimensional small-sized problem. To deal with this problem, lots of subspace learning (SL) methods are developed to reduce the dimension of the pixels while preserving the important discriminant information. Motivated by ridge linear regression (RLR) framework for SL, we propose a spectral-spatial shared linear regression method (SSSLR) for extracting the feature representation. Comparing with RLR, our proposed SSSLR has the following two advantages. First, we utilize a convex set to explore the spatial structure for computing the linear projection matrix. Second, we utilize a shared structure learning model, which is formed by original data space and a hidden feature space, to learn a more discriminant linear projection matrix for classification. To optimize our proposed method, an efficient iterative algorithm is proposed. Experimental results on two popular HSI data sets, i.e., Indian Pines and Salinas demonstrate that our proposed methods outperform many SL methods.

Fast multi-scale feature fusion for ECG heartbeat classification

NASA Astrophysics Data System (ADS)

Ai, Danni; Yang, Jian; Wang, Zeyu; Fan, Jingfan; Ai, Changbin; Wang, Yongtian

2015-12-01

Electrocardiogram (ECG) is conducted to monitor the electrical activity of the heart by presenting small amplitude and duration signals; as a result, hidden information present in ECG data is difficult to determine. However, this concealed information can be used to detect abnormalities. In our study, a fast feature-fusion method of ECG heartbeat classification based on multi-linear subspace learning is proposed. The method consists of four stages. First, baseline and high frequencies are removed to segment heartbeat. Second, as an extension of wavelets, wavelet-packet decomposition is conducted to extract features. With wavelet-packet decomposition, good time and frequency resolutions can be provided simultaneously. Third, decomposed confidences are arranged as a two-way tensor, in which feature fusion is directly implemented with generalized N dimensional ICA (GND-ICA). In this method, co-relationship among different data information is considered, and disadvantages of dimensionality are prevented; this method can also be used to reduce computing compared with linear subspace-learning methods (PCA). Finally, support vector machine (SVM) is considered as a classifier in heartbeat classification. In this study, ECG records are obtained from the MIT-BIT arrhythmia database. Four main heartbeat classes are used to examine the proposed algorithm. Based on the results of five measurements, sensitivity, positive predictivity, accuracy, average accuracy, and t-test, our conclusion is that a GND-ICA-based strategy can be used to provide enhanced ECG heartbeat classification. Furthermore, large redundant features are eliminated, and classification time is reduced.

Hyperspectral material identification on radiance data using single-atmosphere or multiple-atmosphere modeling

NASA Astrophysics Data System (ADS)

Mariano, Adrian V.; Grossmann, John M.

2010-11-01

Reflectance-domain methods convert hyperspectral data from radiance to reflectance using an atmospheric compensation model. Material detection and identification are performed by comparing the compensated data to target reflectance spectra. We introduce two radiance-domain approaches, Single atmosphere Adaptive Cosine Estimator (SACE) and Multiple atmosphere ACE (MACE) in which the target reflectance spectra are instead converted into sensor-reaching radiance using physics-based models. For SACE, known illumination and atmospheric conditions are incorporated in a single atmospheric model. For MACE the conditions are unknown so the algorithm uses many atmospheric models to cover the range of environmental variability, and it approximates the result using a subspace model. This approach is sometimes called the invariant method, and requires the choice of a subspace dimension for the model. We compare these two radiance-domain approaches to a Reflectance-domain ACE (RACE) approach on a HYDICE image featuring concealed materials. All three algorithms use the ACE detector, and all three techniques are able to detect most of the hidden materials in the imagery. For MACE we observe a strong dependence on the choice of the material subspace dimension. Increasing this value can lead to a decline in performance.

Wavelet subspace decomposition of thermal infrared images for defect detection in artworks

NASA Astrophysics Data System (ADS)

Ahmad, M. Z.; Khan, A. A.; Mezghani, S.; Perrin, E.; Mouhoubi, K.; Bodnar, J. L.; Vrabie, V.

2016-07-01

Health of ancient artworks must be routinely monitored for their adequate preservation. Faults in these artworks may develop over time and must be identified as precisely as possible. The classical acoustic testing techniques, being invasive, risk causing permanent damage during periodic inspections. Infrared thermometry offers a promising solution to map faults in artworks. It involves heating the artwork and recording its thermal response using infrared camera. A novel strategy based on pseudo-random binary excitation principle is used in this work to suppress the risks associated with prolonged heating. The objective of this work is to develop an automatic scheme for detecting faults in the captured images. An efficient scheme based on wavelet based subspace decomposition is developed which favors identification of, the otherwise invisible, weaker faults. Two major problems addressed in this work are the selection of the optimal wavelet basis and the subspace level selection. A novel criterion based on regional mutual information is proposed for the latter. The approach is successfully tested on a laboratory based sample as well as real artworks. A new contrast enhancement metric is developed to demonstrate the quantitative efficiency of the algorithm. The algorithm is successfully deployed for both laboratory based and real artworks.

Robust Adaptive Beamforming with Sensor Position Errors Using Weighted Subspace Fitting-Based Covariance Matrix Reconstruction.

PubMed

Chen, Peng; Yang, Yixin; Wang, Yong; Ma, Yuanliang

2018-05-08

When sensor position errors exist, the performance of recently proposed interference-plus-noise covariance matrix (INCM)-based adaptive beamformers may be severely degraded. In this paper, we propose a weighted subspace fitting-based INCM reconstruction algorithm to overcome sensor displacement for linear arrays. By estimating the rough signal directions, we construct a novel possible mismatched steering vector (SV) set. We analyze the proximity of the signal subspace from the sample covariance matrix (SCM) and the space spanned by the possible mismatched SV set. After solving an iterative optimization problem, we reconstruct the INCM using the estimated sensor position errors. Then we estimate the SV of the desired signal by solving an optimization problem with the reconstructed INCM. The main advantage of the proposed algorithm is its robustness against SV mismatches dominated by unknown sensor position errors. Numerical examples show that even if the position errors are up to half of the assumed sensor spacing, the output signal-to-interference-plus-noise ratio is only reduced by 4 dB. Beam patterns plotted using experiment data show that the interference suppression capability of the proposed beamformer outperforms other tested beamformers.

A real negative selection algorithm with evolutionary preference for anomaly detection

NASA Astrophysics Data System (ADS)

Yang, Tao; Chen, Wen; Li, Tao

2017-04-01

Traditional real negative selection algorithms (RNSAs) adopt the estimated coverage (c0) as the algorithm termination threshold, and generate detectors randomly. With increasing dimensions, the data samples could reside in the low-dimensional subspace, so that the traditional detectors cannot effectively distinguish these samples. Furthermore, in high-dimensional feature space, c0 cannot exactly reflect the detectors set coverage rate for the nonself space, and it could lead the algorithm to be terminated unexpectedly when the number of detectors is insufficient. These shortcomings make the traditional RNSAs to perform poorly in high-dimensional feature space. Based upon "evolutionary preference" theory in immunology, this paper presents a real negative selection algorithm with evolutionary preference (RNSAP). RNSAP utilizes the "unknown nonself space", "low-dimensional target subspace" and "known nonself feature" as the evolutionary preference to guide the generation of detectors, thus ensuring the detectors can cover the nonself space more effectively. Besides, RNSAP uses redundancy to replace c0 as the termination threshold, in this way RNSAP can generate adequate detectors under a proper convergence rate. The theoretical analysis and experimental result demonstrate that, compared to the classical RNSA (V-detector), RNSAP can achieve a higher detection rate, but with less detectors and computing cost.

Automated modal parameter estimation using correlation analysis and bootstrap sampling

NASA Astrophysics Data System (ADS)

Yaghoubi, Vahid; Vakilzadeh, Majid K.; Abrahamsson, Thomas J. S.

2018-02-01

The estimation of modal parameters from a set of noisy measured data is a highly judgmental task, with user expertise playing a significant role in distinguishing between estimated physical and noise modes of a test-piece. Various methods have been developed to automate this procedure. The common approach is to identify models with different orders and cluster similar modes together. However, most proposed methods based on this approach suffer from high-dimensional optimization problems in either the estimation or clustering step. To overcome this problem, this study presents an algorithm for autonomous modal parameter estimation in which the only required optimization is performed in a three-dimensional space. To this end, a subspace-based identification method is employed for the estimation and a non-iterative correlation-based method is used for the clustering. This clustering is at the heart of the paper. The keys to success are correlation metrics that are able to treat the problems of spatial eigenvector aliasing and nonunique eigenvectors of coalescent modes simultaneously. The algorithm commences by the identification of an excessively high-order model from frequency response function test data. The high number of modes of this model provides bases for two subspaces: one for likely physical modes of the tested system and one for its complement dubbed the subspace of noise modes. By employing the bootstrap resampling technique, several subsets are generated from the same basic dataset and for each of them a model is identified to form a set of models. Then, by correlation analysis with the two aforementioned subspaces, highly correlated modes of these models which appear repeatedly are clustered together and the noise modes are collected in a so-called Trashbox cluster. Stray noise modes attracted to the mode clusters are trimmed away in a second step by correlation analysis. The final step of the algorithm is a fuzzy c-means clustering procedure applied to a three-dimensional feature space to assign a degree of physicalness to each cluster. The proposed algorithm is applied to two case studies: one with synthetic data and one with real test data obtained from a hammer impact test. The results indicate that the algorithm successfully clusters similar modes and gives a reasonable quantification of the extent to which each cluster is physical.

A Composite Algorithm for Mixed Integer Constrained Nonlinear Optimization.

DTIC Science & Technology

1980-01-01

de Silva [141, and Weisman and Wood [76). A particular direct search algorithm, the simplex method, has been cited for having the potential for...spaced discrete points on a line which makes the direction suitable for an efficient integer search technique based on Fibonacci numbers. Two...defined by a subset of variables. The complex algorithm is particularly well suited for this subspace search for two reasons. First, the complex method

Development of Parallel Architectures for Sensor Array Processing. Volume 1

DTIC Science & Technology

1993-08-01

required for the DOA estimation [ 1-7]. The Multiple Signal Classification ( MUSIC ) [ 1] and the Estimation of Signal Parameters by Rotational...manifold and the estimated subspace. Although MUSIC is a high resolution algorithm, it has several drawbacks including the fact that complete knowledge of...thoroughly, MUSIC algorithm was selected to develop special purpose hardware for real time computation. Summary of the MUSIC algorithm is as follows

Geometric mean for subspace selection.

PubMed

Tao, Dacheng; Li, Xuelong; Wu, Xindong; Maybank, Stephen J

2009-02-01

Subspace selection approaches are powerful tools in pattern classification and data visualization. One of the most important subspace approaches is the linear dimensionality reduction step in the Fisher's linear discriminant analysis (FLDA), which has been successfully employed in many fields such as biometrics, bioinformatics, and multimedia information management. However, the linear dimensionality reduction step in FLDA has a critical drawback: for a classification task with c classes, if the dimension of the projected subspace is strictly lower than c - 1, the projection to a subspace tends to merge those classes, which are close together in the original feature space. If separate classes are sampled from Gaussian distributions, all with identical covariance matrices, then the linear dimensionality reduction step in FLDA maximizes the mean value of the Kullback-Leibler (KL) divergences between different classes. Based on this viewpoint, the geometric mean for subspace selection is studied in this paper. Three criteria are analyzed: 1) maximization of the geometric mean of the KL divergences, 2) maximization of the geometric mean of the normalized KL divergences, and 3) the combination of 1 and 2. Preliminary experimental results based on synthetic data, UCI Machine Learning Repository, and handwriting digits show that the third criterion is a potential discriminative subspace selection method, which significantly reduces the class separation problem in comparing with the linear dimensionality reduction step in FLDA and its several representative extensions.

Beamspace dual signal space projection (bDSSP): a method for selective detection of deep sources in MEG measurements.

PubMed

Sekihara, Kensuke; Adachi, Yoshiaki; Kubota, Hiroshi K; Cai, Chang; Nagarajan, Srikantan S

2018-06-01

Magnetoencephalography (MEG) has a well-recognized weakness at detecting deeper brain activities. This paper proposes a novel algorithm for selective detection of deep sources by suppressing interference signals from superficial sources in MEG measurements. The proposed algorithm combines the beamspace preprocessing method with the dual signal space projection (DSSP) interference suppression method. A prerequisite of the proposed algorithm is prior knowledge of the location of the deep sources. The proposed algorithm first derives the basis vectors that span a local region just covering the locations of the deep sources. It then estimates the time-domain signal subspace of the superficial sources by using the projector composed of these basis vectors. Signals from the deep sources are extracted by projecting the row space of the data matrix onto the direction orthogonal to the signal subspace of the superficial sources. Compared with the previously proposed beamspace signal space separation (SSS) method, the proposed algorithm is capable of suppressing much stronger interference from superficial sources. This capability is demonstrated in our computer simulation as well as experiments using phantom data. The proposed bDSSP algorithm can be a powerful tool in studies of physiological functions of midbrain and deep brain structures.

Beamspace dual signal space projection (bDSSP): a method for selective detection of deep sources in MEG measurements

NASA Astrophysics Data System (ADS)

Sekihara, Kensuke; Adachi, Yoshiaki; Kubota, Hiroshi K.; Cai, Chang; Nagarajan, Srikantan S.

2018-06-01

Objective. Magnetoencephalography (MEG) has a well-recognized weakness at detecting deeper brain activities. This paper proposes a novel algorithm for selective detection of deep sources by suppressing interference signals from superficial sources in MEG measurements. Approach. The proposed algorithm combines the beamspace preprocessing method with the dual signal space projection (DSSP) interference suppression method. A prerequisite of the proposed algorithm is prior knowledge of the location of the deep sources. The proposed algorithm first derives the basis vectors that span a local region just covering the locations of the deep sources. It then estimates the time-domain signal subspace of the superficial sources by using the projector composed of these basis vectors. Signals from the deep sources are extracted by projecting the row space of the data matrix onto the direction orthogonal to the signal subspace of the superficial sources. Main results. Compared with the previously proposed beamspace signal space separation (SSS) method, the proposed algorithm is capable of suppressing much stronger interference from superficial sources. This capability is demonstrated in our computer simulation as well as experiments using phantom data. Significance. The proposed bDSSP algorithm can be a powerful tool in studies of physiological functions of midbrain and deep brain structures.

Hybrid fuzzy cluster ensemble framework for tumor clustering from biomolecular data.

PubMed

Yu, Zhiwen; Chen, Hantao; You, Jane; Han, Guoqiang; Li, Le

2013-01-01

Cancer class discovery using biomolecular data is one of the most important tasks for cancer diagnosis and treatment. Tumor clustering from gene expression data provides a new way to perform cancer class discovery. Most of the existing research works adopt single-clustering algorithms to perform tumor clustering is from biomolecular data that lack robustness, stability, and accuracy. To further improve the performance of tumor clustering from biomolecular data, we introduce the fuzzy theory into the cluster ensemble framework for tumor clustering from biomolecular data, and propose four kinds of hybrid fuzzy cluster ensemble frameworks (HFCEF), named as HFCEF-I, HFCEF-II, HFCEF-III, and HFCEF-IV, respectively, to identify samples that belong to different types of cancers. The difference between HFCEF-I and HFCEF-II is that they adopt different ensemble generator approaches to generate a set of fuzzy matrices in the ensemble. Specifically, HFCEF-I applies the affinity propagation algorithm (AP) to perform clustering on the sample dimension and generates a set of fuzzy matrices in the ensemble based on the fuzzy membership function and base samples selected by AP. HFCEF-II adopts AP to perform clustering on the attribute dimension, generates a set of subspaces, and obtains a set of fuzzy matrices in the ensemble by performing fuzzy c-means on subspaces. Compared with HFCEF-I and HFCEF-II, HFCEF-III and HFCEF-IV consider the characteristics of HFCEF-I and HFCEF-II. HFCEF-III combines HFCEF-I and HFCEF-II in a serial way, while HFCEF-IV integrates HFCEF-I and HFCEF-II in a concurrent way. HFCEFs adopt suitable consensus functions, such as the fuzzy c-means algorithm or the normalized cut algorithm (Ncut), to summarize generated fuzzy matrices, and obtain the final results. The experiments on real data sets from UCI machine learning repository and cancer gene expression profiles illustrate that 1) the proposed hybrid fuzzy cluster ensemble frameworks work well on real data sets, especially biomolecular data, and 2) the proposed approaches are able to provide more robust, stable, and accurate results when compared with the state-of-the-art single clustering algorithms and traditional cluster ensemble approaches.

Generalizing the self-healing diffusion Monte Carlo approach to finite temperature: a path for the optimization of low-energy many-body bases.

PubMed

Reboredo, Fernando A; Kim, Jeongnim

2014-02-21

A statistical method is derived for the calculation of thermodynamic properties of many-body systems at low temperatures. This method is based on the self-healing diffusion Monte Carlo method for complex functions [F. A. Reboredo, J. Chem. Phys. 136, 204101 (2012)] and some ideas of the correlation function Monte Carlo approach [D. M. Ceperley and B. Bernu, J. Chem. Phys. 89, 6316 (1988)]. In order to allow the evolution in imaginary time to describe the density matrix, we remove the fixed-node restriction using complex antisymmetric guiding wave functions. In the process we obtain a parallel algorithm that optimizes a small subspace of the many-body Hilbert space to provide maximum overlap with the subspace spanned by the lowest-energy eigenstates of a many-body Hamiltonian. We show in a model system that the partition function is progressively maximized within this subspace. We show that the subspace spanned by the small basis systematically converges towards the subspace spanned by the lowest energy eigenstates. Possible applications of this method for calculating the thermodynamic properties of many-body systems near the ground state are discussed. The resulting basis can also be used to accelerate the calculation of the ground or excited states with quantum Monte Carlo.

Generalizing the self-healing diffusion Monte Carlo approach to finite temperature: A path for the optimization of low-energy many-body bases

NASA Astrophysics Data System (ADS)

Reboredo, Fernando A.; Kim, Jeongnim

2014-02-01

A statistical method is derived for the calculation of thermodynamic properties of many-body systems at low temperatures. This method is based on the self-healing diffusion Monte Carlo method for complex functions [F. A. Reboredo, J. Chem. Phys. 136, 204101 (2012)] and some ideas of the correlation function Monte Carlo approach [D. M. Ceperley and B. Bernu, J. Chem. Phys. 89, 6316 (1988)]. In order to allow the evolution in imaginary time to describe the density matrix, we remove the fixed-node restriction using complex antisymmetric guiding wave functions. In the process we obtain a parallel algorithm that optimizes a small subspace of the many-body Hilbert space to provide maximum overlap with the subspace spanned by the lowest-energy eigenstates of a many-body Hamiltonian. We show in a model system that the partition function is progressively maximized within this subspace. We show that the subspace spanned by the small basis systematically converges towards the subspace spanned by the lowest energy eigenstates. Possible applications of this method for calculating the thermodynamic properties of many-body systems near the ground state are discussed. The resulting basis can also be used to accelerate the calculation of the ground or excited states with quantum Monte Carlo.

Low-Rank Discriminant Embedding for Multiview Learning.

PubMed

Li, Jingjing; Wu, Yue; Zhao, Jidong; Lu, Ke

2017-11-01

This paper focuses on the specific problem of multiview learning where samples have the same feature set but different probability distributions, e.g., different viewpoints or different modalities. Since samples lying in different distributions cannot be compared directly, this paper aims to learn a latent subspace shared by multiple views assuming that the input views are generated from this latent subspace. Previous approaches usually learn the common subspace by either maximizing the empirical likelihood, or preserving the geometric structure. However, considering the complementarity between the two objectives, this paper proposes a novel approach, named low-rank discriminant embedding (LRDE), for multiview learning by taking full advantage of both sides. By further considering the duality between data points and features of multiview scene, i.e., data points can be grouped based on their distribution on features, while features can be grouped based on their distribution on the data points, LRDE not only deploys low-rank constraints on both sample level and feature level to dig out the shared factors across different views, but also preserves geometric information in both the ambient sample space and the embedding feature space by designing a novel graph structure under the framework of graph embedding. Finally, LRDE jointly optimizes low-rank representation and graph embedding in a unified framework. Comprehensive experiments in both multiview manner and pairwise manner demonstrate that LRDE performs much better than previous approaches proposed in recent literatures.

A computationally efficient parallel Levenberg-Marquardt algorithm for highly parameterized inverse model analyses

DOE Office of Scientific and Technical Information (OSTI.GOV)

Lin, Youzuo; O'Malley, Daniel; Vesselinov, Velimir V.

Inverse modeling seeks model parameters given a set of observations. However, for practical problems because the number of measurements is often large and the model parameters are also numerous, conventional methods for inverse modeling can be computationally expensive. We have developed a new, computationally-efficient parallel Levenberg-Marquardt method for solving inverse modeling problems with a highly parameterized model space. Levenberg-Marquardt methods require the solution of a linear system of equations which can be prohibitively expensive to compute for moderate to large-scale problems. Our novel method projects the original linear problem down to a Krylov subspace, such that the dimensionality of themore » problem can be significantly reduced. Furthermore, we store the Krylov subspace computed when using the first damping parameter and recycle the subspace for the subsequent damping parameters. The efficiency of our new inverse modeling algorithm is significantly improved using these computational techniques. We apply this new inverse modeling method to invert for random transmissivity fields in 2D and a random hydraulic conductivity field in 3D. Our algorithm is fast enough to solve for the distributed model parameters (transmissivity) in the model domain. The algorithm is coded in Julia and implemented in the MADS computational framework (http://mads.lanl.gov). By comparing with Levenberg-Marquardt methods using standard linear inversion techniques such as QR or SVD methods, our Levenberg-Marquardt method yields a speed-up ratio on the order of ~10 1 to ~10 2 in a multi-core computational environment. Furthermore, our new inverse modeling method is a powerful tool for characterizing subsurface heterogeneity for moderate- to large-scale problems.« less

Efficient statistically accurate algorithms for the Fokker-Planck equation in large dimensions

NASA Astrophysics Data System (ADS)

Chen, Nan; Majda, Andrew J.

2018-02-01

Solving the Fokker-Planck equation for high-dimensional complex turbulent dynamical systems is an important and practical issue. However, most traditional methods suffer from the curse of dimensionality and have difficulties in capturing the fat tailed highly intermittent probability density functions (PDFs) of complex systems in turbulence, neuroscience and excitable media. In this article, efficient statistically accurate algorithms are developed for solving both the transient and the equilibrium solutions of Fokker-Planck equations associated with high-dimensional nonlinear turbulent dynamical systems with conditional Gaussian structures. The algorithms involve a hybrid strategy that requires only a small number of ensembles. Here, a conditional Gaussian mixture in a high-dimensional subspace via an extremely efficient parametric method is combined with a judicious non-parametric Gaussian kernel density estimation in the remaining low-dimensional subspace. Particularly, the parametric method provides closed analytical formulae for determining the conditional Gaussian distributions in the high-dimensional subspace and is therefore computationally efficient and accurate. The full non-Gaussian PDF of the system is then given by a Gaussian mixture. Different from traditional particle methods, each conditional Gaussian distribution here covers a significant portion of the high-dimensional PDF. Therefore a small number of ensembles is sufficient to recover the full PDF, which overcomes the curse of dimensionality. Notably, the mixture distribution has significant skill in capturing the transient behavior with fat tails of the high-dimensional non-Gaussian PDFs, and this facilitates the algorithms in accurately describing the intermittency and extreme events in complex turbulent systems. It is shown in a stringent set of test problems that the method only requires an order of O (100) ensembles to successfully recover the highly non-Gaussian transient PDFs in up to 6 dimensions with only small errors.

A computationally efficient parallel Levenberg-Marquardt algorithm for highly parameterized inverse model analyses

DOE PAGES

Lin, Youzuo; O'Malley, Daniel; Vesselinov, Velimir V.

2016-09-01

Inverse modeling seeks model parameters given a set of observations. However, for practical problems because the number of measurements is often large and the model parameters are also numerous, conventional methods for inverse modeling can be computationally expensive. We have developed a new, computationally-efficient parallel Levenberg-Marquardt method for solving inverse modeling problems with a highly parameterized model space. Levenberg-Marquardt methods require the solution of a linear system of equations which can be prohibitively expensive to compute for moderate to large-scale problems. Our novel method projects the original linear problem down to a Krylov subspace, such that the dimensionality of themore » problem can be significantly reduced. Furthermore, we store the Krylov subspace computed when using the first damping parameter and recycle the subspace for the subsequent damping parameters. The efficiency of our new inverse modeling algorithm is significantly improved using these computational techniques. We apply this new inverse modeling method to invert for random transmissivity fields in 2D and a random hydraulic conductivity field in 3D. Our algorithm is fast enough to solve for the distributed model parameters (transmissivity) in the model domain. The algorithm is coded in Julia and implemented in the MADS computational framework (http://mads.lanl.gov). By comparing with Levenberg-Marquardt methods using standard linear inversion techniques such as QR or SVD methods, our Levenberg-Marquardt method yields a speed-up ratio on the order of ~10 1 to ~10 2 in a multi-core computational environment. Furthermore, our new inverse modeling method is a powerful tool for characterizing subsurface heterogeneity for moderate- to large-scale problems.« less

Reduced multiple empirical kernel learning machine.

PubMed

Wang, Zhe; Lu, MingZhe; Gao, Daqi

2015-02-01

Multiple kernel learning (MKL) is demonstrated to be flexible and effective in depicting heterogeneous data sources since MKL can introduce multiple kernels rather than a single fixed kernel into applications. However, MKL would get a high time and space complexity in contrast to single kernel learning, which is not expected in real-world applications. Meanwhile, it is known that the kernel mapping ways of MKL generally have two forms including implicit kernel mapping and empirical kernel mapping (EKM), where the latter is less attracted. In this paper, we focus on the MKL with the EKM, and propose a reduced multiple empirical kernel learning machine named RMEKLM for short. To the best of our knowledge, it is the first to reduce both time and space complexity of the MKL with EKM. Different from the existing MKL, the proposed RMEKLM adopts the Gauss Elimination technique to extract a set of feature vectors, which is validated that doing so does not lose much information of the original feature space. Then RMEKLM adopts the extracted feature vectors to span a reduced orthonormal subspace of the feature space, which is visualized in terms of the geometry structure. It can be demonstrated that the spanned subspace is isomorphic to the original feature space, which means that the dot product of two vectors in the original feature space is equal to that of the two corresponding vectors in the generated orthonormal subspace. More importantly, the proposed RMEKLM brings a simpler computation and meanwhile needs a less storage space, especially in the processing of testing. Finally, the experimental results show that RMEKLM owns a much efficient and effective performance in terms of both complexity and classification. The contributions of this paper can be given as follows: (1) by mapping the input space into an orthonormal subspace, the geometry of the generated subspace is visualized; (2) this paper first reduces both the time and space complexity of the EKM-based MKL; (3) this paper adopts the Gauss Elimination, one of the on-the-shelf techniques, to generate a basis of the original feature space, which is stable and efficient.

Anomaly detection in hyperspectral imagery: statistics vs. graph-based algorithms

NASA Astrophysics Data System (ADS)

Berkson, Emily E.; Messinger, David W.

2016-05-01

Anomaly detection (AD) algorithms are frequently applied to hyperspectral imagery, but different algorithms produce different outlier results depending on the image scene content and the assumed background model. This work provides the first comparison of anomaly score distributions between common statistics-based anomaly detection algorithms (RX and subspace-RX) and the graph-based Topological Anomaly Detector (TAD). Anomaly scores in statistical AD algorithms should theoretically approximate a chi-squared distribution; however, this is rarely the case with real hyperspectral imagery. The expected distribution of scores found with graph-based methods remains unclear. We also look for general trends in algorithm performance with varied scene content. Three separate scenes were extracted from the hyperspectral MegaScene image taken over downtown Rochester, NY with the VIS-NIR-SWIR ProSpecTIR instrument. In order of most to least cluttered, we study an urban, suburban, and rural scene. The three AD algorithms were applied to each scene, and the distributions of the most anomalous 5% of pixels were compared. We find that subspace-RX performs better than RX, because the data becomes more normal when the highest variance principal components are removed. We also see that compared to statistical detectors, anomalies detected by TAD are easier to separate from the background. Due to their different underlying assumptions, the statistical and graph-based algorithms highlighted different anomalies within the urban scene. These results will lead to a deeper understanding of these algorithms and their applicability across different types of imagery.

Relevance feedback-based building recognition

NASA Astrophysics Data System (ADS)

Li, Jing; Allinson, Nigel M.

2010-07-01

Building recognition is a nontrivial task in computer vision research which can be utilized in robot localization, mobile navigation, etc. However, existing building recognition systems usually encounter the following two problems: 1) extracted low level features cannot reveal the true semantic concepts; and 2) they usually involve high dimensional data which require heavy computational costs and memory. Relevance feedback (RF), widely applied in multimedia information retrieval, is able to bridge the gap between the low level visual features and high level concepts; while dimensionality reduction methods can mitigate the high-dimensional problem. In this paper, we propose a building recognition scheme which integrates the RF and subspace learning algorithms. Experimental results undertaken on our own building database show that the newly proposed scheme appreciably enhances the recognition accuracy.

Discrete Optimization of Electronic Hyperpolarizabilities in a Chemical Subspace

DTIC Science & Technology

2009-05-01

molecular design. Methods for optimization in discrete spaces have been studied extensively and recently reviewed ( 5). Optimization methods include...integer programming, as in branch-and-bound techniques (including dead-end elimination [ 6]), simulated annealing ( 7), and genetic algorithms ( 8...These algorithms have found renewed interest and application in molecular and materials design (9- 12) . Recently, new approaches have been

A model reduction approach to numerical inversion for a parabolic partial differential equation

NASA Astrophysics Data System (ADS)

Borcea, Liliana; Druskin, Vladimir; Mamonov, Alexander V.; Zaslavsky, Mikhail

2014-12-01

We propose a novel numerical inversion algorithm for the coefficients of parabolic partial differential equations, based on model reduction. The study is motivated by the application of controlled source electromagnetic exploration, where the unknown is the subsurface electrical resistivity and the data are time resolved surface measurements of the magnetic field. The algorithm presented in this paper considers inversion in one and two dimensions. The reduced model is obtained with rational interpolation in the frequency (Laplace) domain and a rational Krylov subspace projection method. It amounts to a nonlinear mapping from the function space of the unknown resistivity to the small dimensional space of the parameters of the reduced model. We use this mapping as a nonlinear preconditioner for the Gauss-Newton iterative solution of the inverse problem. The advantage of the inversion algorithm is twofold. First, the nonlinear preconditioner resolves most of the nonlinearity of the problem. Thus the iterations are less likely to get stuck in local minima and the convergence is fast. Second, the inversion is computationally efficient because it avoids repeated accurate simulations of the time-domain response. We study the stability of the inversion algorithm for various rational Krylov subspaces, and assess its performance with numerical experiments.

Subspace techniques to remove artifacts from EEG: a quantitative analysis.

PubMed

Teixeira, A R; Tome, A M; Lang, E W; Martins da Silva, A

2008-01-01

In this work we discuss and apply projective subspace techniques to both multichannel as well as single channel recordings. The single-channel approach is based on singular spectrum analysis(SSA) and the multichannel approach uses the extended infomax algorithm which is implemented in the opensource toolbox EEGLAB. Both approaches will be evaluated using artificial mixtures of a set of selected EEG signals. The latter were selected visually to contain as the dominant activity one of the characteristic bands of an electroencephalogram (EEG). The evaluation is performed both in the time and frequency domain by using correlation coefficients and coherence function, respectively.

Scalable Robust Principal Component Analysis Using Grassmann Averages.

PubMed

Hauberg, Sren; Feragen, Aasa; Enficiaud, Raffi; Black, Michael J

2016-11-01

In large datasets, manual data verification is impossible, and we must expect the number of outliers to increase with data size. While principal component analysis (PCA) can reduce data size, and scalable solutions exist, it is well-known that outliers can arbitrarily corrupt the results. Unfortunately, state-of-the-art approaches for robust PCA are not scalable. We note that in a zero-mean dataset, each observation spans a one-dimensional subspace, giving a point on the Grassmann manifold. We show that the average subspace corresponds to the leading principal component for Gaussian data. We provide a simple algorithm for computing this Grassmann Average ( GA), and show that the subspace estimate is less sensitive to outliers than PCA for general distributions. Because averages can be efficiently computed, we immediately gain scalability. We exploit robust averaging to formulate the Robust Grassmann Average (RGA) as a form of robust PCA. The resulting Trimmed Grassmann Average ( TGA) is appropriate for computer vision because it is robust to pixel outliers. The algorithm has linear computational complexity and minimal memory requirements. We demonstrate TGA for background modeling, video restoration, and shadow removal. We show scalability by performing robust PCA on the entire Star Wars IV movie; a task beyond any current method. Source code is available online.

Component-based subspace linear discriminant analysis method for face recognition with one training sample

NASA Astrophysics Data System (ADS)

Huang, Jian; Yuen, Pong C.; Chen, Wen-Sheng; Lai, J. H.

2005-05-01

Many face recognition algorithms/systems have been developed in the last decade and excellent performances have also been reported when there is a sufficient number of representative training samples. In many real-life applications such as passport identification, only one well-controlled frontal sample image is available for training. Under this situation, the performance of existing algorithms will degrade dramatically or may not even be implemented. We propose a component-based linear discriminant analysis (LDA) method to solve the one training sample problem. The basic idea of the proposed method is to construct local facial feature component bunches by moving each local feature region in four directions. In this way, we not only generate more samples with lower dimension than the original image, but also consider the face detection localization error while training. After that, we propose a subspace LDA method, which is tailor-made for a small number of training samples, for the local feature projection to maximize the discrimination power. Theoretical analysis and experiment results show that our proposed subspace LDA is efficient and overcomes the limitations in existing LDA methods. Finally, we combine the contributions of each local component bunch with a weighted combination scheme to draw the recognition decision. A FERET database is used for evaluating the proposed method and results are encouraging.

Noise covariance incorporated MEG-MUSIC algorithm: a method for multiple-dipole estimation tolerant of the influence of background brain activity.

PubMed

Sekihara, K; Poeppel, D; Marantz, A; Koizumi, H; Miyashita, Y

1997-09-01

This paper proposes a method of localizing multiple current dipoles from spatio-temporal biomagnetic data. The method is based on the multiple signal classification (MUSIC) algorithm and is tolerant of the influence of background brain activity. In this method, the noise covariance matrix is estimated using a portion of the data that contains noise, but does not contain any signal information. Then, a modified noise subspace projector is formed using the generalized eigenvectors of the noise and measured-data covariance matrices. The MUSIC localizer is calculated using this noise subspace projector and the noise covariance matrix. The results from a computer simulation have verified the effectiveness of the method. The method was then applied to source estimation for auditory-evoked fields elicited by syllable speech sounds. The results strongly suggest the method's effectiveness in removing the influence of background activity.

Representation learning: a unified deep learning framework for automatic prostate MR segmentation.

PubMed

Liao, Shu; Gao, Yaozong; Oto, Aytekin; Shen, Dinggang

2013-01-01

Image representation plays an important role in medical image analysis. The key to the success of different medical image analysis algorithms is heavily dependent on how we represent the input data, namely features used to characterize the input image. In the literature, feature engineering remains as an active research topic, and many novel hand-crafted features are designed such as Haar wavelet, histogram of oriented gradient, and local binary patterns. However, such features are not designed with the guidance of the underlying dataset at hand. To this end, we argue that the most effective features should be designed in a learning based manner, namely representation learning, which can be adapted to different patient datasets at hand. In this paper, we introduce a deep learning framework to achieve this goal. Specifically, a stacked independent subspace analysis (ISA) network is adopted to learn the most effective features in a hierarchical and unsupervised manner. The learnt features are adapted to the dataset at hand and encode high level semantic anatomical information. The proposed method is evaluated on the application of automatic prostate MR segmentation. Experimental results show that significant segmentation accuracy improvement can be achieved by the proposed deep learning method compared to other state-of-the-art segmentation approaches.

Ensemble based adaptive over-sampling method for imbalanced data learning in computer aided detection of microaneurysm.

PubMed

Ren, Fulong; Cao, Peng; Li, Wei; Zhao, Dazhe; Zaiane, Osmar

2017-01-01

Diabetic retinopathy (DR) is a progressive disease, and its detection at an early stage is crucial for saving a patient's vision. An automated screening system for DR can help in reduce the chances of complete blindness due to DR along with lowering the work load on ophthalmologists. Among the earliest signs of DR are microaneurysms (MAs). However, current schemes for MA detection appear to report many false positives because detection algorithms have high sensitivity. Inevitably some non-MAs structures are labeled as MAs in the initial MAs identification step. This is a typical "class imbalance problem". Class imbalanced data has detrimental effects on the performance of conventional classifiers. In this work, we propose an ensemble based adaptive over-sampling algorithm for overcoming the class imbalance problem in the false positive reduction, and we use Boosting, Bagging, Random subspace as the ensemble framework to improve microaneurysm detection. The ensemble based over-sampling methods we proposed combine the strength of adaptive over-sampling and ensemble. The objective of the amalgamation of ensemble and adaptive over-sampling is to reduce the induction biases introduced from imbalanced data and to enhance the generalization classification performance of extreme learning machines (ELM). Experimental results show that our ASOBoost method has higher area under the ROC curve (AUC) and G-mean values than many existing class imbalance learning methods. Copyright © 2016 Elsevier Ltd. All rights reserved.

An Intelligent Architecture Based on Field Programmable Gate Arrays Designed to Detect Moving Objects by Using Principal Component Analysis

PubMed Central

Bravo, Ignacio; Mazo, Manuel; Lázaro, José L.; Gardel, Alfredo; Jiménez, Pedro; Pizarro, Daniel

2010-01-01

This paper presents a complete implementation of the Principal Component Analysis (PCA) algorithm in Field Programmable Gate Array (FPGA) devices applied to high rate background segmentation of images. The classical sequential execution of different parts of the PCA algorithm has been parallelized. This parallelization has led to the specific development and implementation in hardware of the different stages of PCA, such as computation of the correlation matrix, matrix diagonalization using the Jacobi method and subspace projections of images. On the application side, the paper presents a motion detection algorithm, also entirely implemented on the FPGA, and based on the developed PCA core. This consists of dynamically thresholding the differences between the input image and the one obtained by expressing the input image using the PCA linear subspace previously obtained as a background model. The proposal achieves a high ratio of processed images (up to 120 frames per second) and high quality segmentation results, with a completely embedded and reliable hardware architecture based on commercial CMOS sensors and FPGA devices. PMID:22163406

An intelligent architecture based on Field Programmable Gate Arrays designed to detect moving objects by using Principal Component Analysis.

PubMed

Bravo, Ignacio; Mazo, Manuel; Lázaro, José L; Gardel, Alfredo; Jiménez, Pedro; Pizarro, Daniel

2010-01-01

This paper presents a complete implementation of the Principal Component Analysis (PCA) algorithm in Field Programmable Gate Array (FPGA) devices applied to high rate background segmentation of images. The classical sequential execution of different parts of the PCA algorithm has been parallelized. This parallelization has led to the specific development and implementation in hardware of the different stages of PCA, such as computation of the correlation matrix, matrix diagonalization using the Jacobi method and subspace projections of images. On the application side, the paper presents a motion detection algorithm, also entirely implemented on the FPGA, and based on the developed PCA core. This consists of dynamically thresholding the differences between the input image and the one obtained by expressing the input image using the PCA linear subspace previously obtained as a background model. The proposal achieves a high ratio of processed images (up to 120 frames per second) and high quality segmentation results, with a completely embedded and reliable hardware architecture based on commercial CMOS sensors and FPGA devices.

Imaging of downward-looking linear array SAR using three-dimensional spatial smoothing MUSIC algorithm

NASA Astrophysics Data System (ADS)

Zhang, Siqian; Kuang, Gangyao

2014-10-01

In this paper, a novel three-dimensional imaging algorithm of downward-looking linear array SAR is presented. To improve the resolution, multiple signal classification (MUSIC) algorithm has been used. However, since the scattering centers are always correlated in real SAR system, the estimated covariance matrix becomes singular. To address the problem, a three-dimensional spatial smoothing method is proposed in this paper to restore the singular covariance matrix to a full-rank one. The three-dimensional signal matrix can be divided into a set of orthogonal three-dimensional subspaces. The main idea of the method is based on extracting the array correlation matrix as the average of all correlation matrices from the subspaces. In addition, the spectral height of the peaks contains no information with regard to the scattering intensity of the different scattering centers, thus it is difficulty to reconstruct the backscattering information. The least square strategy is used to estimate the amplitude of the scattering center in this paper. The above results of the theoretical analysis are verified by 3-D scene simulations and experiments on real data.

Scalable Methods for Uncertainty Quantification, Data Assimilation and Target Accuracy Assessment for Multi-Physics Advanced Simulation of Light Water Reactors

NASA Astrophysics Data System (ADS)

Khuwaileh, Bassam

High fidelity simulation of nuclear reactors entails large scale applications characterized with high dimensionality and tremendous complexity where various physics models are integrated in the form of coupled models (e.g. neutronic with thermal-hydraulic feedback). Each of the coupled modules represents a high fidelity formulation of the first principles governing the physics of interest. Therefore, new developments in high fidelity multi-physics simulation and the corresponding sensitivity/uncertainty quantification analysis are paramount to the development and competitiveness of reactors achieved through enhanced understanding of the design and safety margins. Accordingly, this dissertation introduces efficient and scalable algorithms for performing efficient Uncertainty Quantification (UQ), Data Assimilation (DA) and Target Accuracy Assessment (TAA) for large scale, multi-physics reactor design and safety problems. This dissertation builds upon previous efforts for adaptive core simulation and reduced order modeling algorithms and extends these efforts towards coupled multi-physics models with feedback. The core idea is to recast the reactor physics analysis in terms of reduced order models. This can be achieved via identifying the important/influential degrees of freedom (DoF) via the subspace analysis, such that the required analysis can be recast by considering the important DoF only. In this dissertation, efficient algorithms for lower dimensional subspace construction have been developed for single physics and multi-physics applications with feedback. Then the reduced subspace is used to solve realistic, large scale forward (UQ) and inverse problems (DA and TAA). Once the elite set of DoF is determined, the uncertainty/sensitivity/target accuracy assessment and data assimilation analysis can be performed accurately and efficiently for large scale, high dimensional multi-physics nuclear engineering applications. Hence, in this work a Karhunen-Loeve (KL) based algorithm previously developed to quantify the uncertainty for single physics models is extended for large scale multi-physics coupled problems with feedback effect. Moreover, a non-linear surrogate based UQ approach is developed, used and compared to performance of the KL approach and brute force Monte Carlo (MC) approach. On the other hand, an efficient Data Assimilation (DA) algorithm is developed to assess information about model's parameters: nuclear data cross-sections and thermal-hydraulics parameters. Two improvements are introduced in order to perform DA on the high dimensional problems. First, a goal-oriented surrogate model can be used to replace the original models in the depletion sequence (MPACT -- COBRA-TF - ORIGEN). Second, approximating the complex and high dimensional solution space with a lower dimensional subspace makes the sampling process necessary for DA possible for high dimensional problems. Moreover, safety analysis and design optimization depend on the accurate prediction of various reactor attributes. Predictions can be enhanced by reducing the uncertainty associated with the attributes of interest. Accordingly, an inverse problem can be defined and solved to assess the contributions from sources of uncertainty; and experimental effort can be subsequently directed to further improve the uncertainty associated with these sources. In this dissertation a subspace-based gradient-free and nonlinear algorithm for inverse uncertainty quantification namely the Target Accuracy Assessment (TAA) has been developed and tested. The ideas proposed in this dissertation were first validated using lattice physics applications simulated using SCALE6.1 package (Pressurized Water Reactor (PWR) and Boiling Water Reactor (BWR) lattice models). Ultimately, the algorithms proposed her were applied to perform UQ and DA for assembly level (CASL progression problem number 6) and core wide problems representing Watts Bar Nuclear 1 (WBN1) for cycle 1 of depletion (CASL Progression Problem Number 9) modeled via simulated using VERA-CS which consists of several multi-physics coupled models. The analysis and algorithms developed in this dissertation were encoded and implemented in a newly developed tool kit algorithms for Reduced Order Modeling based Uncertainty/Sensitivity Estimator (ROMUSE).

Hyperspectral Vehicle BRDF Learning: An Exploration of Vehicle Reflectance Variation and Optimal Measures of Spectral Similarity for Vehicle Reacquisition and Tracking Algorithms

NASA Astrophysics Data System (ADS)

Svejkosky, Joseph

The spectral signatures of vehicles in hyperspectral imagery exhibit temporal variations due to the preponderance of surfaces with material properties that display non-Lambertian bi-directional reflectance distribution functions (BRDFs). These temporal variations are caused by changing illumination conditions, changing sun-target-sensor geometry, changing road surface properties, and changing vehicle orientations. To quantify these variations and determine their relative importance in a sub-pixel vehicle reacquisition and tracking scenario, a hyperspectral vehicle BRDF sampling experiment was conducted in which four vehicles were rotated at different orientations and imaged over a six-hour period. The hyperspectral imagery was calibrated using novel in-scene methods and converted to reflectance imagery. The resulting BRDF sampled time-series imagery showed a strong vehicle level BRDF dependence on vehicle shape in off-nadir imaging scenarios and a strong dependence on vehicle color in simulated nadir imaging scenarios. The imagery also exhibited spectral features characteristic of sampling the BRDF of non-Lambertian targets, which were subsequently verified with simulations. In addition, the imagery demonstrated that the illumination contribution from vehicle adjacent horizontal surfaces significantly altered the shape and magnitude of the vehicle reflectance spectrum. The results of the BRDF sampling experiment illustrate the need for a target vehicle BRDF model and detection scheme that incorporates non-Lambertian BRDFs. A new detection algorithm called Eigenvector Loading Regression (ELR) is proposed that learns a hyperspectral vehicle BRDF from a series of BRDF measurements using regression in a lower dimensional space and then applies the learned BRDF to make test spectrum predictions. In cases of non-Lambertian vehicle BRDF, this detection methodology performs favorably when compared to subspace detections algorithms and graph-based detection algorithms that do not account for the target BRDF. The algorithms are compared using a test environment in which observed spectral reflectance signatures from the BRDF sampling experiment are implanted into aerial hyperspectral imagery that contain large quantities of vehicles.

Density-matrix-based algorithm for solving eigenvalue problems

NASA Astrophysics Data System (ADS)

Polizzi, Eric

2009-03-01

A fast and stable numerical algorithm for solving the symmetric eigenvalue problem is presented. The technique deviates fundamentally from the traditional Krylov subspace iteration based techniques (Arnoldi and Lanczos algorithms) or other Davidson-Jacobi techniques and takes its inspiration from the contour integration and density-matrix representation in quantum mechanics. It will be shown that this algorithm—named FEAST—exhibits high efficiency, robustness, accuracy, and scalability on parallel architectures. Examples from electronic structure calculations of carbon nanotubes are presented, and numerical performances and capabilities are discussed.

Robust Joint Graph Sparse Coding for Unsupervised Spectral Feature Selection.

PubMed

Zhu, Xiaofeng; Li, Xuelong; Zhang, Shichao; Ju, Chunhua; Wu, Xindong

2017-06-01

In this paper, we propose a new unsupervised spectral feature selection model by embedding a graph regularizer into the framework of joint sparse regression for preserving the local structures of data. To do this, we first extract the bases of training data by previous dictionary learning methods and, then, map original data into the basis space to generate their new representations, by proposing a novel joint graph sparse coding (JGSC) model. In JGSC, we first formulate its objective function by simultaneously taking subspace learning and joint sparse regression into account, then, design a new optimization solution to solve the resulting objective function, and further prove the convergence of the proposed solution. Furthermore, we extend JGSC to a robust JGSC (RJGSC) via replacing the least square loss function with a robust loss function, for achieving the same goals and also avoiding the impact of outliers. Finally, experimental results on real data sets showed that both JGSC and RJGSC outperformed the state-of-the-art algorithms in terms of k -nearest neighbor classification performance.

Tracking and recognition face in videos with incremental local sparse representation model

NASA Astrophysics Data System (ADS)

Wang, Chao; Wang, Yunhong; Zhang, Zhaoxiang

2013-10-01

This paper addresses the problem of tracking and recognizing faces via incremental local sparse representation. First a robust face tracking algorithm is proposed via employing local sparse appearance and covariance pooling method. In the following face recognition stage, with the employment of a novel template update strategy, which combines incremental subspace learning, our recognition algorithm adapts the template to appearance changes and reduces the influence of occlusion and illumination variation. This leads to a robust video-based face tracking and recognition with desirable performance. In the experiments, we test the quality of face recognition in real-world noisy videos on YouTube database, which includes 47 celebrities. Our proposed method produces a high face recognition rate at 95% of all videos. The proposed face tracking and recognition algorithms are also tested on a set of noisy videos under heavy occlusion and illumination variation. The tracking results on challenging benchmark videos demonstrate that the proposed tracking algorithm performs favorably against several state-of-the-art methods. In the case of the challenging dataset in which faces undergo occlusion and illumination variation, and tracking and recognition experiments under significant pose variation on the University of California, San Diego (Honda/UCSD) database, our proposed method also consistently demonstrates a high recognition rate.

An adaptive optimal control for smart structures based on the subspace tracking identification technique

NASA Astrophysics Data System (ADS)

Ripamonti, Francesco; Resta, Ferruccio; Borroni, Massimo; Cazzulani, Gabriele

2014-04-01

A new method for the real-time identification of mechanical system modal parameters is used in order to design different adaptive control logics aiming to reduce the vibrations in a carbon fiber plate smart structure. It is instrumented with three piezoelectric actuators, three accelerometers and three strain gauges. The real-time identification is based on a recursive subspace tracking algorithm whose outputs are elaborated by an ARMA model. A statistical approach is finally applied to choose the modal parameter correct values. These are given in input to model-based control logics such as a gain scheduling and an adaptive LQR control.

Scalable posterior approximations for large-scale Bayesian inverse problems via likelihood-informed parameter and state reduction

NASA Astrophysics Data System (ADS)

Cui, Tiangang; Marzouk, Youssef; Willcox, Karen

2016-06-01

Two major bottlenecks to the solution of large-scale Bayesian inverse problems are the scaling of posterior sampling algorithms to high-dimensional parameter spaces and the computational cost of forward model evaluations. Yet incomplete or noisy data, the state variation and parameter dependence of the forward model, and correlations in the prior collectively provide useful structure that can be exploited for dimension reduction in this setting-both in the parameter space of the inverse problem and in the state space of the forward model. To this end, we show how to jointly construct low-dimensional subspaces of the parameter space and the state space in order to accelerate the Bayesian solution of the inverse problem. As a byproduct of state dimension reduction, we also show how to identify low-dimensional subspaces of the data in problems with high-dimensional observations. These subspaces enable approximation of the posterior as a product of two factors: (i) a projection of the posterior onto a low-dimensional parameter subspace, wherein the original likelihood is replaced by an approximation involving a reduced model; and (ii) the marginal prior distribution on the high-dimensional complement of the parameter subspace. We present and compare several strategies for constructing these subspaces using only a limited number of forward and adjoint model simulations. The resulting posterior approximations can rapidly be characterized using standard sampling techniques, e.g., Markov chain Monte Carlo. Two numerical examples demonstrate the accuracy and efficiency of our approach: inversion of an integral equation in atmospheric remote sensing, where the data dimension is very high; and the inference of a heterogeneous transmissivity field in a groundwater system, which involves a partial differential equation forward model with high dimensional state and parameters.

Parallel eigenanalysis of finite element models in a completely connected architecture

NASA Technical Reports Server (NTRS)

Akl, F. A.; Morel, M. R.

1989-01-01

A parallel algorithm is presented for the solution of the generalized eigenproblem in linear elastic finite element analysis, (K)(phi) = (M)(phi)(omega), where (K) and (M) are of order N, and (omega) is order of q. The concurrent solution of the eigenproblem is based on the multifrontal/modified subspace method and is achieved in a completely connected parallel architecture in which each processor is allowed to communicate with all other processors. The algorithm was successfully implemented on a tightly coupled multiple-instruction multiple-data parallel processing machine, Cray X-MP. A finite element model is divided into m domains each of which is assumed to process n elements. Each domain is then assigned to a processor or to a logical processor (task) if the number of domains exceeds the number of physical processors. The macrotasking library routines are used in mapping each domain to a user task. Computational speed-up and efficiency are used to determine the effectiveness of the algorithm. The effect of the number of domains, the number of degrees-of-freedom located along the global fronts and the dimension of the subspace on the performance of the algorithm are investigated. A parallel finite element dynamic analysis program, p-feda, is documented and the performance of its subroutines in parallel environment is analyzed.

Efficient Statistically Accurate Algorithms for the Fokker-Planck Equation in Large Dimensions

NASA Astrophysics Data System (ADS)

Chen, N.; Majda, A.

2017-12-01

Solving the Fokker-Planck equation for high-dimensional complex turbulent dynamical systems is an important and practical issue. However, most traditional methods suffer from the curse of dimensionality and have difficulties in capturing the fat tailed highly intermittent probability density functions (PDFs) of complex systems in turbulence, neuroscience and excitable media. In this article, efficient statistically accurate algorithms are developed for solving both the transient and the equilibrium solutions of Fokker-Planck equations associated with high-dimensional nonlinear turbulent dynamical systems with conditional Gaussian structures. The algorithms involve a hybrid strategy that requires only a small number of ensembles. Here, a conditional Gaussian mixture in a high-dimensional subspace via an extremely efficient parametric method is combined with a judicious non-parametric Gaussian kernel density estimation in the remaining low-dimensional subspace. Particularly, the parametric method, which is based on an effective data assimilation framework, provides closed analytical formulae for determining the conditional Gaussian distributions in the high-dimensional subspace. Therefore, it is computationally efficient and accurate. The full non-Gaussian PDF of the system is then given by a Gaussian mixture. Different from the traditional particle methods, each conditional Gaussian distribution here covers a significant portion of the high-dimensional PDF. Therefore a small number of ensembles is sufficient to recover the full PDF, which overcomes the curse of dimensionality. Notably, the mixture distribution has a significant skill in capturing the transient behavior with fat tails of the high-dimensional non-Gaussian PDFs, and this facilitates the algorithms in accurately describing the intermittency and extreme events in complex turbulent systems. It is shown in a stringent set of test problems that the method only requires an order of O(100) ensembles to successfully recover the highly non-Gaussian transient PDFs in up to 6 dimensions with only small errors.

Model, analysis, and evaluation of the effects of analog VLSI arithmetic on linear subspace-based image recognition.

PubMed

Carvajal, Gonzalo; Figueroa, Miguel

2014-07-01

Typical image recognition systems operate in two stages: feature extraction to reduce the dimensionality of the input space, and classification based on the extracted features. Analog Very Large Scale Integration (VLSI) is an attractive technology to achieve compact and low-power implementations of these computationally intensive tasks for portable embedded devices. However, device mismatch limits the resolution of the circuits fabricated with this technology. Traditional layout techniques to reduce the mismatch aim to increase the resolution at the transistor level, without considering the intended application. Relating mismatch parameters to specific effects in the application level would allow designers to apply focalized mismatch compensation techniques according to predefined performance/cost tradeoffs. This paper models, analyzes, and evaluates the effects of mismatched analog arithmetic in both feature extraction and classification circuits. For the feature extraction, we propose analog adaptive linear combiners with on-chip learning for both Least Mean Square (LMS) and Generalized Hebbian Algorithm (GHA). Using mathematical abstractions of analog circuits, we identify mismatch parameters that are naturally compensated during the learning process, and propose cost-effective guidelines to reduce the effect of the rest. For the classification, we derive analog models for the circuits necessary to implement Nearest Neighbor (NN) approach and Radial Basis Function (RBF) networks, and use them to emulate analog classifiers with standard databases of face and hand-writing digits. Formal analysis and experiments show how we can exploit adaptive structures and properties of the input space to compensate the effects of device mismatch at the application level, thus reducing the design overhead of traditional layout techniques. Results are also directly extensible to multiple application domains using linear subspace methods. Copyright © 2014 Elsevier Ltd. All rights reserved.

Reverse time migration by Krylov subspace reduced order modeling

NASA Astrophysics Data System (ADS)

Basir, Hadi Mahdavi; Javaherian, Abdolrahim; Shomali, Zaher Hossein; Firouz-Abadi, Roohollah Dehghani; Gholamy, Shaban Ali

2018-04-01

Imaging is a key step in seismic data processing. To date, a myriad of advanced pre-stack depth migration approaches have been developed; however, reverse time migration (RTM) is still considered as the high-end imaging algorithm. The main limitations associated with the performance cost of reverse time migration are the intensive computation of the forward and backward simulations, time consumption, and memory allocation related to imaging condition. Based on the reduced order modeling, we proposed an algorithm, which can be adapted to all the aforementioned factors. Our proposed method benefit from Krylov subspaces method to compute certain mode shapes of the velocity model computed by as an orthogonal base of reduced order modeling. Reverse time migration by reduced order modeling is helpful concerning the highly parallel computation and strongly reduces the memory requirement of reverse time migration. The synthetic model results showed that suggested method can decrease the computational costs of reverse time migration by several orders of magnitudes, compared with reverse time migration by finite element method.

A novel framework for feature extraction in multi-sensor action potential sorting.

PubMed

Wu, Shun-Chi; Swindlehurst, A Lee; Nenadic, Zoran

2015-09-30

Extracellular recordings of multi-unit neural activity have become indispensable in neuroscience research. The analysis of the recordings begins with the detection of the action potentials (APs), followed by a classification step where each AP is associated with a given neural source. A feature extraction step is required prior to classification in order to reduce the dimensionality of the data and the impact of noise, allowing source clustering algorithms to work more efficiently. In this paper, we propose a novel framework for multi-sensor AP feature extraction based on the so-called Matched Subspace Detector (MSD), which is shown to be a natural generalization of standard single-sensor algorithms. Clustering using both simulated data and real AP recordings taken in the locust antennal lobe demonstrates that the proposed approach yields features that are discriminatory and lead to promising results. Unlike existing methods, the proposed algorithm finds joint spatio-temporal feature vectors that match the dominant subspace observed in the two-dimensional data without needs for a forward propagation model and AP templates. The proposed MSD approach provides more discriminatory features for unsupervised AP sorting applications. Copyright © 2015 Elsevier B.V. All rights reserved.

Systematic Dimensionality Reduction for Quantum Walks: Optimal Spatial Search and Transport on Non-Regular Graphs

PubMed Central

Novo, Leonardo; Chakraborty, Shantanav; Mohseni, Masoud; Neven, Hartmut; Omar, Yasser

2015-01-01

Continuous time quantum walks provide an important framework for designing new algorithms and modelling quantum transport and state transfer problems. Often, the graph representing the structure of a problem contains certain symmetries that confine the dynamics to a smaller subspace of the full Hilbert space. In this work, we use invariant subspace methods, that can be computed systematically using the Lanczos algorithm, to obtain the reduced set of states that encompass the dynamics of the problem at hand without the specific knowledge of underlying symmetries. First, we apply this method to obtain new instances of graphs where the spatial quantum search algorithm is optimal: complete graphs with broken links and complete bipartite graphs, in particular, the star graph. These examples show that regularity and high-connectivity are not needed to achieve optimal spatial search. We also show that this method considerably simplifies the calculation of quantum transport efficiencies. Furthermore, we observe improved efficiencies by removing a few links from highly symmetric graphs. Finally, we show that this reduction method also allows us to obtain an upper bound for the fidelity of a single qubit transfer on an XY spin network. PMID:26330082

Beta Hebbian Learning as a New Method for Exploratory Projection Pursuit.

PubMed

Quintián, Héctor; Corchado, Emilio

2017-09-01

In this research, a novel family of learning rules called Beta Hebbian Learning (BHL) is thoroughly investigated to extract information from high-dimensional datasets by projecting the data onto low-dimensional (typically two dimensional) subspaces, improving the existing exploratory methods by providing a clear representation of data's internal structure. BHL applies a family of learning rules derived from the Probability Density Function (PDF) of the residual based on the beta distribution. This family of rules may be called Hebbian in that all use a simple multiplication of the output of the neural network with some function of the residuals after feedback. The derived learning rules can be linked to an adaptive form of Exploratory Projection Pursuit and with artificial distributions, the networks perform as the theory suggests they should: the use of different learning rules derived from different PDFs allows the identification of "interesting" dimensions (as far from the Gaussian distribution as possible) in high-dimensional datasets. This novel algorithm, BHL, has been tested over seven artificial datasets to study the behavior of BHL parameters, and was later applied successfully over four real datasets, comparing its results, in terms of performance, with other well-known Exploratory and projection models such as Maximum Likelihood Hebbian Learning (MLHL), Locally-Linear Embedding (LLE), Curvilinear Component Analysis (CCA), Isomap and Neural Principal Component Analysis (Neural PCA).

Matrix Product Operator Simulations of Quantum Algorithms

DTIC Science & Technology

2015-02-01

parallel to the Grover subspace parametrically: (Zi|φ〉)‖ = s cos γ|α〉+ s sin γ|β〉, s = √ a(k)2 (N − 1)2 + b(k)2, γ = tan −1 ( b(k)(N − 1) a(k) ) (6.32) Each...of this vector parallel to the Grover subspace in parametric form: (XiZi|φ〉)‖ = s cos(γ)|α〉+ s sin(γ)|β〉, s = 1√ N − 1 , γ = tan −1 ( cot (( k + 1 2 ) θ...quant- ph/0001106, 2000. Bibliography 146 [30] Jérémie Roland and Nicolas J Cerf. Quantum search by local adiabatic evolution. Physical Review A, 65(4

Generalizing the self-healing diffusion Monte Carlo approach to finite temperature: a path for the optimization of low-energy many-body basis expansions

DOE Office of Scientific and Technical Information (OSTI.GOV)

Kim, Jeongnim; Reboredo, Fernando A.

The self-healing diffusion Monte Carlo method for complex functions [F. A. Reboredo J. Chem. Phys. {\\bf 136}, 204101 (2012)] and some ideas of the correlation function Monte Carlo approach [D. M. Ceperley and B. Bernu, J. Chem. Phys. {\\bf 89}, 6316 (1988)] are blended to obtain a method for the calculation of thermodynamic properties of many-body systems at low temperatures. In order to allow the evolution in imaginary time to describe the density matrix, we remove the fixed-node restriction using complex antisymmetric trial wave functions. A statistical method is derived for the calculation of finite temperature properties of many-body systemsmore » near the ground state. In the process we also obtain a parallel algorithm that optimizes the many-body basis of a small subspace of the many-body Hilbert space. This small subspace is optimized to have maximum overlap with the one expanded by the lower energy eigenstates of a many-body Hamiltonian. We show in a model system that the Helmholtz free energy is minimized within this subspace as the iteration number increases. We show that the subspace expanded by the small basis systematically converges towards the subspace expanded by the lowest energy eigenstates. Possible applications of this method to calculate the thermodynamic properties of many-body systems near the ground state are discussed. The resulting basis can be also used to accelerate the calculation of the ground or excited states with Quantum Monte Carlo.« less

Fast Query-Optimized Kernel-Machine Classification

NASA Technical Reports Server (NTRS)

Mazzoni, Dominic; DeCoste, Dennis

2004-01-01

A recently developed algorithm performs kernel-machine classification via incremental approximate nearest support vectors. The algorithm implements support-vector machines (SVMs) at speeds 10 to 100 times those attainable by use of conventional SVM algorithms. The algorithm offers potential benefits for classification of images, recognition of speech, recognition of handwriting, and diverse other applications in which there are requirements to discern patterns in large sets of data. SVMs constitute a subset of kernel machines (KMs), which have become popular as models for machine learning and, more specifically, for automated classification of input data on the basis of labeled training data. While similar in many ways to k-nearest-neighbors (k-NN) models and artificial neural networks (ANNs), SVMs tend to be more accurate. Using representations that scale only linearly in the numbers of training examples, while exploring nonlinear (kernelized) feature spaces that are exponentially larger than the original input dimensionality, KMs elegantly and practically overcome the classic curse of dimensionality. However, the price that one must pay for the power of KMs is that query-time complexity scales linearly with the number of training examples, making KMs often orders of magnitude more computationally expensive than are ANNs, decision trees, and other popular machine learning alternatives. The present algorithm treats an SVM classifier as a special form of a k-NN. The algorithm is based partly on an empirical observation that one can often achieve the same classification as that of an exact KM by using only small fraction of the nearest support vectors (SVs) of a query. The exact KM output is a weighted sum over the kernel values between the query and the SVs. In this algorithm, the KM output is approximated with a k-NN classifier, the output of which is a weighted sum only over the kernel values involving k selected SVs. Before query time, there are gathered statistics about how misleading the output of the k-NN model can be, relative to the outputs of the exact KM for a representative set of examples, for each possible k from 1 to the total number of SVs. From these statistics, there are derived upper and lower thresholds for each step k. These thresholds identify output levels for which the particular variant of the k-NN model already leans so strongly positively or negatively that a reversal in sign is unlikely, given the weaker SV neighbors still remaining. At query time, the partial output of each query is incrementally updated, stopping as soon as it exceeds the predetermined statistical thresholds of the current step. For an easy query, stopping can occur as early as step k = 1. For more difficult queries, stopping might not occur until nearly all SVs are touched. A key empirical observation is that this approach can tolerate very approximate nearest-neighbor orderings. In experiments, SVs and queries were projected to a subspace comprising the top few principal- component dimensions and neighbor orderings were computed in that subspace. This approach ensured that the overhead of the nearest-neighbor computations was insignificant, relative to that of the exact KM computation.

CaSPIAN: A Causal Compressive Sensing Algorithm for Discovering Directed Interactions in Gene Networks

PubMed Central

Emad, Amin; Milenkovic, Olgica

2014-01-01

We introduce a novel algorithm for inference of causal gene interactions, termed CaSPIAN (Causal Subspace Pursuit for Inference and Analysis of Networks), which is based on coupling compressive sensing and Granger causality techniques. The core of the approach is to discover sparse linear dependencies between shifted time series of gene expressions using a sequential list-version of the subspace pursuit reconstruction algorithm and to estimate the direction of gene interactions via Granger-type elimination. The method is conceptually simple and computationally efficient, and it allows for dealing with noisy measurements. Its performance as a stand-alone platform without biological side-information was tested on simulated networks, on the synthetic IRMA network in Saccharomyces cerevisiae, and on data pertaining to the human HeLa cell network and the SOS network in E. coli. The results produced by CaSPIAN are compared to the results of several related algorithms, demonstrating significant improvements in inference accuracy of documented interactions. These findings highlight the importance of Granger causality techniques for reducing the number of false-positives, as well as the influence of noise and sampling period on the accuracy of the estimates. In addition, the performance of the method was tested in conjunction with biological side information of the form of sparse “scaffold networks”, to which new edges were added using available RNA-seq or microarray data. These biological priors aid in increasing the sensitivity and precision of the algorithm in the small sample regime. PMID:24622336

A unified classifier for robust face recognition based on combining multiple subspace algorithms

NASA Astrophysics Data System (ADS)

Ijaz Bajwa, Usama; Ahmad Taj, Imtiaz; Waqas Anwar, Muhammad

2012-10-01

Face recognition being the fastest growing biometric technology has expanded manifold in the last few years. Various new algorithms and commercial systems have been proposed and developed. However, none of the proposed or developed algorithm is a complete solution because it may work very well on one set of images with say illumination changes but may not work properly on another set of image variations like expression variations. This study is motivated by the fact that any single classifier cannot claim to show generally better performance against all facial image variations. To overcome this shortcoming and achieve generality, combining several classifiers using various strategies has been studied extensively also incorporating the question of suitability of any classifier for this task. The study is based on the outcome of a comprehensive comparative analysis conducted on a combination of six subspace extraction algorithms and four distance metrics on three facial databases. The analysis leads to the selection of the most suitable classifiers which performs better on one task or the other. These classifiers are then combined together onto an ensemble classifier by two different strategies of weighted sum and re-ranking. The results of the ensemble classifier show that these strategies can be effectively used to construct a single classifier that can successfully handle varying facial image conditions of illumination, aging and facial expressions.

The use of Lanczos's method to solve the large generalized symmetric definite eigenvalue problem

NASA Technical Reports Server (NTRS)

Jones, Mark T.; Patrick, Merrell L.

1989-01-01

The generalized eigenvalue problem, Kx = Lambda Mx, is of significant practical importance, especially in structural enginering where it arises as the vibration and buckling problem. A new algorithm, LANZ, based on Lanczos's method is developed. LANZ uses a technique called dynamic shifting to improve the efficiency and reliability of the Lanczos algorithm. A new algorithm for solving the tridiagonal matrices that arise when using Lanczos's method is described. A modification of Parlett and Scott's selective orthogonalization algorithm is proposed. Results from an implementation of LANZ on a Convex C-220 show it to be superior to a subspace iteration code.

Evolutionary Ensemble for In Silico Prediction of Ames Test Mutagenicity

NASA Astrophysics Data System (ADS)

Chen, Huanhuan; Yao, Xin

Driven by new regulations and animal welfare, the need to develop in silico models has increased recently as alternative approaches to safety assessment of chemicals without animal testing. This paper describes a novel machine learning ensemble approach to building an in silico model for the prediction of the Ames test mutagenicity, one of a battery of the most commonly used experimental in vitro and in vivo genotoxicity tests for safety evaluation of chemicals. Evolutionary random neural ensemble with negative correlation learning (ERNE) [1] was developed based on neural networks and evolutionary algorithms. ERNE combines the method of bootstrap sampling on training data with the method of random subspace feature selection to ensure diversity in creating individuals within an initial ensemble. Furthermore, while evolving individuals within the ensemble, it makes use of the negative correlation learning, enabling individual NNs to be trained as accurate as possible while still manage to maintain them as diverse as possible. Therefore, the resulting individuals in the final ensemble are capable of cooperating collectively to achieve better generalization of prediction. The empirical experiment suggest that ERNE is an effective ensemble approach for predicting the Ames test mutagenicity of chemicals.

Faster than classical quantum algorithm for dense formulas of exact satisfiability and occupation problems

NASA Astrophysics Data System (ADS)

Mandrà, Salvatore; Giacomo Guerreschi, Gian; Aspuru-Guzik, Alán

2016-07-01

We present an exact quantum algorithm for solving the Exact Satisfiability problem, which belongs to the important NP-complete complexity class. The algorithm is based on an intuitive approach that can be divided into two parts: the first step consists in the identification and efficient characterization of a restricted subspace that contains all the valid assignments of the Exact Satisfiability; while the second part performs a quantum search in such restricted subspace. The quantum algorithm can be used either to find a valid assignment (or to certify that no solution exists) or to count the total number of valid assignments. The query complexities for the worst-case are respectively bounded by O(\\sqrt{{2}n-{M\\prime }}) and O({2}n-{M\\prime }), where n is the number of variables and {M}\\prime the number of linearly independent clauses. Remarkably, the proposed quantum algorithm results to be faster than any known exact classical algorithm to solve dense formulas of Exact Satisfiability. As a concrete application, we provide the worst-case complexity for the Hamiltonian cycle problem obtained after mapping it to a suitable Occupation problem. Specifically, we show that the time complexity for the proposed quantum algorithm is bounded by O({2}n/4) for 3-regular undirected graphs, where n is the number of nodes. The same worst-case complexity holds for (3,3)-regular bipartite graphs. As a reference, the current best classical algorithm has a (worst-case) running time bounded by O({2}31n/96). Finally, when compared to heuristic techniques for Exact Satisfiability problems, the proposed quantum algorithm is faster than the classical WalkSAT and Adiabatic Quantum Optimization for random instances with a density of constraints close to the satisfiability threshold, the regime in which instances are typically the hardest to solve. The proposed quantum algorithm can be straightforwardly extended to the generalized version of the Exact Satisfiability known as Occupation problem. The general version of the algorithm is presented and analyzed.

A Subspace Pursuit–based Iterative Greedy Hierarchical Solution to the Neuromagnetic Inverse Problem

PubMed Central

Babadi, Behtash; Obregon-Henao, Gabriel; Lamus, Camilo; Hämäläinen, Matti S.; Brown, Emery N.; Purdon, Patrick L.

2013-01-01

Magnetoencephalography (MEG) is an important non-invasive method for studying activity within the human brain. Source localization methods can be used to estimate spatiotemporal activity from MEG measurements with high temporal resolution, but the spatial resolution of these estimates is poor due to the ill-posed nature of the MEG inverse problem. Recent developments in source localization methodology have emphasized temporal as well as spatial constraints to improve source localization accuracy, but these methods can be computationally intense. Solutions emphasizing spatial sparsity hold tremendous promise, since the underlying neurophysiological processes generating MEG signals are often sparse in nature, whether in the form of focal sources, or distributed sources representing large-scale functional networks. Recent developments in the theory of compressed sensing (CS) provide a rigorous framework to estimate signals with sparse structure. In particular, a class of CS algorithms referred to as greedy pursuit algorithms can provide both high recovery accuracy and low computational complexity. Greedy pursuit algorithms are difficult to apply directly to the MEG inverse problem because of the high-dimensional structure of the MEG source space and the high spatial correlation in MEG measurements. In this paper, we develop a novel greedy pursuit algorithm for sparse MEG source localization that overcomes these fundamental problems. This algorithm, which we refer to as the Subspace Pursuit-based Iterative Greedy Hierarchical (SPIGH) inverse solution, exhibits very low computational complexity while achieving very high localization accuracy. We evaluate the performance of the proposed algorithm using comprehensive simulations, as well as the analysis of human MEG data during spontaneous brain activity and somatosensory stimuli. These studies reveal substantial performance gains provided by the SPIGH algorithm in terms of computational complexity, localization accuracy, and robustness. PMID:24055554

Decomposition of Near-Infrared Spectroscopy Signals Using Oblique Subspace Projections: Applications in Brain Hemodynamic Monitoring.

PubMed

Caicedo, Alexander; Varon, Carolina; Hunyadi, Borbala; Papademetriou, Maria; Tachtsidis, Ilias; Van Huffel, Sabine

2016-01-01

Clinical data is comprised by a large number of synchronously collected biomedical signals that are measured at different locations. Deciphering the interrelationships of these signals can yield important information about their dependence providing some useful clinical diagnostic data. For instance, by computing the coupling between Near-Infrared Spectroscopy signals (NIRS) and systemic variables the status of the hemodynamic regulation mechanisms can be assessed. In this paper we introduce an algorithm for the decomposition of NIRS signals into additive components. The algorithm, SIgnal DEcomposition base on Obliques Subspace Projections (SIDE-ObSP), assumes that the measured NIRS signal is a linear combination of the systemic measurements, following the linear regression model y = Ax + ϵ . SIDE-ObSP decomposes the output such that, each component in the decomposition represents the sole linear influence of one corresponding regressor variable. This decomposition scheme aims at providing a better understanding of the relation between NIRS and systemic variables, and to provide a framework for the clinical interpretation of regression algorithms, thereby, facilitating their introduction into clinical practice. SIDE-ObSP combines oblique subspace projections (ObSP) with the structure of a mean average system in order to define adequate signal subspaces. To guarantee smoothness in the estimated regression parameters, as observed in normal physiological processes, we impose a Tikhonov regularization using a matrix differential operator. We evaluate the performance of SIDE-ObSP by using a synthetic dataset, and present two case studies in the field of cerebral hemodynamics monitoring using NIRS. In addition, we compare the performance of this method with other system identification techniques. In the first case study data from 20 neonates during the first 3 days of life was used, here SIDE-ObSP decoupled the influence of changes in arterial oxygen saturation from the NIRS measurements, facilitating the use of NIRS as a surrogate measure for cerebral blood flow (CBF). The second case study used data from a 3-years old infant under Extra Corporeal Membrane Oxygenation (ECMO), here SIDE-ObSP decomposed cerebral/peripheral tissue oxygenation, as a sum of the partial contributions from different systemic variables, facilitating the comparison between the effects of each systemic variable on the cerebral/peripheral hemodynamics.

Horror Image Recognition Based on Context-Aware Multi-Instance Learning.

PubMed

Li, Bing; Xiong, Weihua; Wu, Ou; Hu, Weiming; Maybank, Stephen; Yan, Shuicheng

2015-12-01

Horror content sharing on the Web is a growing phenomenon that can interfere with our daily life and affect the mental health of those involved. As an important form of expression, horror images have their own characteristics that can evoke extreme emotions. In this paper, we present a novel context-aware multi-instance learning (CMIL) algorithm for horror image recognition. The CMIL algorithm identifies horror images and picks out the regions that cause the sensation of horror in these horror images. It obtains contextual cues among adjacent regions in an image using a random walk on a contextual graph. Borrowing the strength of the fuzzy support vector machine (FSVM), we define a heuristic optimization procedure based on the FSVM to search for the optimal classifier for the CMIL. To improve the initialization of the CMIL, we propose a novel visual saliency model based on the tensor analysis. The average saliency value of each segmented region is set as its initial fuzzy membership in the CMIL. The advantage of the tensor-based visual saliency model is that it not only adaptively selects features, but also dynamically determines fusion weights for saliency value combination from different feature subspaces. The effectiveness of the proposed CMIL model is demonstrated by its use in horror image recognition on two large-scale image sets collected from the Internet.

Generalizing the self-healing diffusion Monte Carlo approach to finite temperature: A path for the optimization of low-energy many-body bases

DOE Office of Scientific and Technical Information (OSTI.GOV)

Reboredo, Fernando A.; Kim, Jeongnim

A statistical method is derived for the calculation of thermodynamic properties of many-body systems at low temperatures. This method is based on the self-healing diffusion Monte Carlo method for complex functions [F. A. Reboredo, J. Chem. Phys. 136, 204101 (2012)] and some ideas of the correlation function Monte Carlo approach [D. M. Ceperley and B. Bernu, J. Chem. Phys. 89, 6316 (1988)]. In order to allow the evolution in imaginary time to describe the density matrix, we remove the fixed-node restriction using complex antisymmetric guiding wave functions. In the process we obtain a parallel algorithm that optimizes a small subspacemore » of the many-body Hilbert space to provide maximum overlap with the subspace spanned by the lowest-energy eigenstates of a many-body Hamiltonian. We show in a model system that the partition function is progressively maximized within this subspace. We show that the subspace spanned by the small basis systematically converges towards the subspace spanned by the lowest energy eigenstates. Possible applications of this method for calculating the thermodynamic properties of many-body systems near the ground state are discussed. The resulting basis can also be used to accelerate the calculation of the ground or excited states with quantum Monte Carlo.« less

Extraction of process zones and low-dimensional attractive subspaces in stochastic fracture mechanics

PubMed Central

Kerfriden, P.; Schmidt, K.M.; Rabczuk, T.; Bordas, S.P.A.

2013-01-01

We propose to identify process zones in heterogeneous materials by tailored statistical tools. The process zone is redefined as the part of the structure where the random process cannot be correctly approximated in a low-dimensional deterministic space. Such a low-dimensional space is obtained by a spectral analysis performed on pre-computed solution samples. A greedy algorithm is proposed to identify both process zone and low-dimensional representative subspace for the solution in the complementary region. In addition to the novelty of the tools proposed in this paper for the analysis of localised phenomena, we show that the reduced space generated by the method is a valid basis for the construction of a reduced order model. PMID:27069423

Performance comparison of machine learning algorithms and number of independent components used in fMRI decoding of belief vs. disbelief.

PubMed

Douglas, P K; Harris, Sam; Yuille, Alan; Cohen, Mark S

2011-05-15

Machine learning (ML) has become a popular tool for mining functional neuroimaging data, and there are now hopes of performing such analyses efficiently in real-time. Towards this goal, we compared accuracy of six different ML algorithms applied to neuroimaging data of persons engaged in a bivariate task, asserting their belief or disbelief of a variety of propositional statements. We performed unsupervised dimension reduction and automated feature extraction using independent component (IC) analysis and extracted IC time courses. Optimization of classification hyperparameters across each classifier occurred prior to assessment. Maximum accuracy was achieved at 92% for Random Forest, followed by 91% for AdaBoost, 89% for Naïve Bayes, 87% for a J48 decision tree, 86% for K*, and 84% for support vector machine. For real-time decoding applications, finding a parsimonious subset of diagnostic ICs might be useful. We used a forward search technique to sequentially add ranked ICs to the feature subspace. For the current data set, we determined that approximately six ICs represented a meaningful basis set for classification. We then projected these six IC spatial maps forward onto a later scanning session within subject. We then applied the optimized ML algorithms to these new data instances, and found that classification accuracy results were reproducible. Additionally, we compared our classification method to our previously published general linear model results on this same data set. The highest ranked IC spatial maps show similarity to brain regions associated with contrasts for belief > disbelief, and disbelief < belief. Copyright © 2010 Elsevier Inc. All rights reserved.

Extended dynamic mode decomposition with dictionary learning: A data-driven adaptive spectral decomposition of the Koopman operator

NASA Astrophysics Data System (ADS)

Li, Qianxiao; Dietrich, Felix; Bollt, Erik M.; Kevrekidis, Ioannis G.

2017-10-01

Numerical approximation methods for the Koopman operator have advanced considerably in the last few years. In particular, data-driven approaches such as dynamic mode decomposition (DMD)51 and its generalization, the extended-DMD (EDMD), are becoming increasingly popular in practical applications. The EDMD improves upon the classical DMD by the inclusion of a flexible choice of dictionary of observables which spans a finite dimensional subspace on which the Koopman operator can be approximated. This enhances the accuracy of the solution reconstruction and broadens the applicability of the Koopman formalism. Although the convergence of the EDMD has been established, applying the method in practice requires a careful choice of the observables to improve convergence with just a finite number of terms. This is especially difficult for high dimensional and highly nonlinear systems. In this paper, we employ ideas from machine learning to improve upon the EDMD method. We develop an iterative approximation algorithm which couples the EDMD with a trainable dictionary represented by an artificial neural network. Using the Duffing oscillator and the Kuramoto Sivashinsky partical differential equation as examples, we show that our algorithm can effectively and efficiently adapt the trainable dictionary to the problem at hand to achieve good reconstruction accuracy without the need to choose a fixed dictionary a priori. Furthermore, to obtain a given accuracy, we require fewer dictionary terms than EDMD with fixed dictionaries. This alleviates an important shortcoming of the EDMD algorithm and enhances the applicability of the Koopman framework to practical problems.

Extended dynamic mode decomposition with dictionary learning: A data-driven adaptive spectral decomposition of the Koopman operator.

PubMed

Li, Qianxiao; Dietrich, Felix; Bollt, Erik M; Kevrekidis, Ioannis G

2017-10-01

Numerical approximation methods for the Koopman operator have advanced considerably in the last few years. In particular, data-driven approaches such as dynamic mode decomposition (DMD) 51 and its generalization, the extended-DMD (EDMD), are becoming increasingly popular in practical applications. The EDMD improves upon the classical DMD by the inclusion of a flexible choice of dictionary of observables which spans a finite dimensional subspace on which the Koopman operator can be approximated. This enhances the accuracy of the solution reconstruction and broadens the applicability of the Koopman formalism. Although the convergence of the EDMD has been established, applying the method in practice requires a careful choice of the observables to improve convergence with just a finite number of terms. This is especially difficult for high dimensional and highly nonlinear systems. In this paper, we employ ideas from machine learning to improve upon the EDMD method. We develop an iterative approximation algorithm which couples the EDMD with a trainable dictionary represented by an artificial neural network. Using the Duffing oscillator and the Kuramoto Sivashinsky partical differential equation as examples, we show that our algorithm can effectively and efficiently adapt the trainable dictionary to the problem at hand to achieve good reconstruction accuracy without the need to choose a fixed dictionary a priori. Furthermore, to obtain a given accuracy, we require fewer dictionary terms than EDMD with fixed dictionaries. This alleviates an important shortcoming of the EDMD algorithm and enhances the applicability of the Koopman framework to practical problems.

SIMULTANEOUS MULTISLICE MAGNETIC RESONANCE FINGERPRINTING WITH LOW-RANK AND SUBSPACE MODELING

PubMed Central

Zhao, Bo; Bilgic, Berkin; Adalsteinsson, Elfar; Griswold, Mark A.; Wald, Lawrence L.; Setsompop, Kawin

2018-01-01

Magnetic resonance fingerprinting (MRF) is a new quantitative imaging paradigm that enables simultaneous acquisition of multiple magnetic resonance tissue parameters (e.g., T1, T2, and spin density). Recently, MRF has been integrated with simultaneous multislice (SMS) acquisitions to enable volumetric imaging with faster scan time. In this paper, we present a new image reconstruction method based on low-rank and subspace modeling for improved SMS-MRF. Here the low-rank model exploits strong spatiotemporal correlation among contrast-weighted images, while the subspace model captures the temporal evolution of magnetization dynamics. With the proposed model, the image reconstruction problem is formulated as a convex optimization problem, for which we develop an algorithm based on variable splitting and the alternating direction method of multipliers. The performance of the proposed method has been evaluated by numerical experiments, and the results demonstrate that the proposed method leads to improved accuracy over the conventional approach. Practically, the proposed method has a potential to allow for a 3x speedup with minimal reconstruction error, resulting in less than 5 sec imaging time per slice. PMID:29060594

Stochastic subspace identification for operational modal analysis of an arch bridge

NASA Astrophysics Data System (ADS)

Loh, Chin-Hsiung; Chen, Ming-Che; Chao, Shu-Hsien

2012-04-01

In this paer the application of output-only system identification technique, known as Stochastic Subspace Identification (SSI) algorithms, for civil infrastructures is carried out. The ability of covariance driven stochastic subspace identification (SSI-COV) was proved through the analysis of the ambient data of an arch bridge under operational condition. A newly developed signal processing technique, Singular Spectrum analysis (SSA), capable to smooth noisy signals, is adopted for pre-processing the recorded data before the SSI. The conjunction of SSA and SSICOV provides a useful criterion for the system order determination. With the aim of estimating accurate modal parameters of the structure in off-line analysis, a stabilization diagram is constructed by plotting the identified poles of the system with increasing the size of data Hankel matrix. Identification task of a real structure, Guandu Bridge, is carried out to identify the system natural frequencies and mode shapes. The uncertainty of the identified model parameters from output-only measurement of the bridge under operation condition, such as temperature and traffic loading conditions, is discussed.

Development of Subspace-based Hybrid Monte Carlo-Deterministric Algorithms for Reactor Physics Calculations

DOE Office of Scientific and Technical Information (OSTI.GOV)

Abdel-Khalik, Hany S.; Zhang, Qiong

2014-05-20

The development of hybrid Monte-Carlo-Deterministic (MC-DT) approaches, taking place over the past few decades, have primarily focused on shielding and detection applications where the analysis requires a small number of responses, i.e. at the detector locations(s). This work further develops a recently introduced global variance reduction approach, denoted by the SUBSPACE approach is designed to allow the use of MC simulation, currently limited to benchmarking calculations, for routine engineering calculations. By way of demonstration, the SUBSPACE approach is applied to assembly level calculations used to generate the few-group homogenized cross-sections. These models are typically expensive and need to be executedmore » in the order of 10 3 - 10 5 times to properly characterize the few-group cross-sections for downstream core-wide calculations. Applicability to k-eigenvalue core-wide models is also demonstrated in this work. Given the favorable results obtained in this work, we believe the applicability of the MC method for reactor analysis calculations could be realized in the near future.« less

Simultaneous multislice magnetic resonance fingerprinting with low-rank and subspace modeling.

PubMed

Bo Zhao; Bilgic, Berkin; Adalsteinsson, Elfar; Griswold, Mark A; Wald, Lawrence L; Setsompop, Kawin

2017-07-01

Magnetic resonance fingerprinting (MRF) is a new quantitative imaging paradigm that enables simultaneous acquisition of multiple magnetic resonance tissue parameters (e.g., T 1 , T 2 , and spin density). Recently, MRF has been integrated with simultaneous multislice (SMS) acquisitions to enable volumetric imaging with faster scan time. In this paper, we present a new image reconstruction method based on low-rank and subspace modeling for improved SMS-MRF. Here the low-rank model exploits strong spatiotemporal correlation among contrast-weighted images, while the subspace model captures the temporal evolution of magnetization dynamics. With the proposed model, the image reconstruction problem is formulated as a convex optimization problem, for which we develop an algorithm based on variable splitting and the alternating direction method of multipliers. The performance of the proposed method has been evaluated by numerical experiments, and the results demonstrate that the proposed method leads to improved accuracy over the conventional approach. Practically, the proposed method has a potential to allow for a 3× speedup with minimal reconstruction error, resulting in less than 5 sec imaging time per slice.

Learning Compositional Shape Models of Multiple Distance Metrics by Information Projection.

PubMed

Luo, Ping; Lin, Liang; Liu, Xiaobai

2016-07-01

This paper presents a novel compositional contour-based shape model by incorporating multiple distance metrics to account for varying shape distortions or deformations. Our approach contains two key steps: 1) contour feature generation and 2) generative model pursuit. For each category, we first densely sample an ensemble of local prototype contour segments from a few positive shape examples and describe each segment using three different types of distance metrics. These metrics are diverse and complementary with each other to capture various shape deformations. We regard the parameterized contour segment plus an additive residual ϵ as a basic subspace, namely, ϵ -ball, in the sense that it represents local shape variance under the certain distance metric. Using these ϵ -balls as features, we then propose a generative learning algorithm to pursue the compositional shape model, which greedily selects the most representative features under the information projection principle. In experiments, we evaluate our model on several public challenging data sets, and demonstrate that the integration of multiple shape distance metrics is capable of dealing various shape deformations, articulations, and background clutter, hence boosting system performance.

Implementation details of the coupled QMR algorithm

NASA Technical Reports Server (NTRS)

Freund, Roland W.; Nachtigal, Noel M.

1992-01-01

The original quasi-minimal residual method (QMR) relies on the three-term look-ahead Lanczos process, to generate basis vectors for the underlying Krylov subspaces. However, empirical observations indicate that, in finite precision arithmetic, three-term vector recurrences are less robust than mathematically equivalent coupled two-term recurrences. Therefore, we recently proposed a new implementation of the QMR method based on a coupled two-term look-ahead Lanczos procedure. In this paper, we describe implementation details of this coupled QMR algorithm, and we present results of numerical experiments.

A combined joint diagonalization-MUSIC algorithm for subsurface targets localization

NASA Astrophysics Data System (ADS)

Wang, Yinlin; Sigman, John B.; Barrowes, Benjamin E.; O'Neill, Kevin; Shubitidze, Fridon

2014-06-01

This paper presents a combined joint diagonalization (JD) and multiple signal classification (MUSIC) algorithm for estimating subsurface objects locations from electromagnetic induction (EMI) sensor data, without solving ill-posed inverse-scattering problems. JD is a numerical technique that finds the common eigenvectors that diagonalize a set of multistatic response (MSR) matrices measured by a time-domain EMI sensor. Eigenvalues from targets of interest (TOI) can be then distinguished automatically from noise-related eigenvalues. Filtering is also carried out in JD to improve the signal-to-noise ratio (SNR) of the data. The MUSIC algorithm utilizes the orthogonality between the signal and noise subspaces in the MSR matrix, which can be separated with information provided by JD. An array of theoreticallycalculated Green's functions are then projected onto the noise subspace, and the location of the target is estimated by the minimum of the projection owing to the orthogonality. This combined method is applied to data from the Time-Domain Electromagnetic Multisensor Towed Array Detection System (TEMTADS). Examples of TEMTADS test stand data and field data collected at Spencer Range, Tennessee are analyzed and presented. Results indicate that due to its noniterative mechanism, the method can be executed fast enough to provide real-time estimation of objects' locations in the field.

Inverse regression-based uncertainty quantification algorithms for high-dimensional models: Theory and practice

DOE Office of Scientific and Technical Information (OSTI.GOV)

Li, Weixuan; Lin, Guang; Li, Bing

2016-09-01

A well-known challenge in uncertainty quantification (UQ) is the "curse of dimensionality". However, many high-dimensional UQ problems are essentially low-dimensional, because the randomness of the quantity of interest (QoI) is caused only by uncertain parameters varying within a low-dimensional subspace, known as the sufficient dimension reduction (SDR) subspace. Motivated by this observation, we propose and demonstrate in this paper an inverse regression-based UQ approach (IRUQ) for high-dimensional problems. Specifically, we use an inverse regression procedure to estimate the SDR subspace and then convert the original problem to a low-dimensional one, which can be efficiently solved by building a response surface model such as a polynomial chaos expansion. The novelty and advantages of the proposed approach is seen in its computational efficiency and practicality. Comparing with Monte Carlo, the traditionally preferred approach for high-dimensional UQ, IRUQ with a comparable cost generally gives much more accurate solutions even for high-dimensional problems, and even when the dimension reduction is not exactly sufficient. Theoretically, IRUQ is proved to converge twice as fast as the approach it uses seeking the SDR subspace. For example, while a sliced inverse regression method converges to the SDR subspace at the rate ofmore » $$O(n^{-1/2})$$, the corresponding IRUQ converges at $$O(n^{-1})$$. IRUQ also provides several desired conveniences in practice. It is non-intrusive, requiring only a simulator to generate realizations of the QoI, and there is no need to compute the high-dimensional gradient of the QoI. Finally, error bars can be derived for the estimation results reported by IRUQ.« less

Removal of EOG Artifacts from EEG Recordings Using Stationary Subspace Analysis

PubMed Central

Zeng, Hong; Song, Aiguo

2014-01-01

An effective approach is proposed in this paper to remove ocular artifacts from the raw EEG recording. The proposed approach first conducts the blind source separation on the raw EEG recording by the stationary subspace analysis (SSA) algorithm. Unlike the classic blind source separation algorithms, SSA is explicitly tailored to the understanding of distribution changes, where both the mean and the covariance matrix are taken into account. In addition, neither independency nor uncorrelation is required among the sources by SSA. Thereby, it can concentrate artifacts in fewer components than the representative blind source separation methods. Next, the components that are determined to be related to the ocular artifacts are projected back to be subtracted from EEG signals, producing the clean EEG data eventually. The experimental results on both the artificially contaminated EEG data and real EEG data have demonstrated the effectiveness of the proposed method, in particular for the cases where limited number of electrodes are used for the recording, as well as when the artifact contaminated signal is highly nonstationary and the underlying sources cannot be assumed to be independent or uncorrelated. PMID:24550696

Active shape models unleashed

NASA Astrophysics Data System (ADS)

Kirschner, Matthias; Wesarg, Stefan

2011-03-01

Active Shape Models (ASMs) are a popular family of segmentation algorithms which combine local appearance models for boundary detection with a statistical shape model (SSM). They are especially popular in medical imaging due to their ability for fast and accurate segmentation of anatomical structures even in large and noisy 3D images. A well-known limitation of ASMs is that the shape constraints are over-restrictive, because the segmentations are bounded by the Principal Component Analysis (PCA) subspace learned from the training data. To overcome this limitation, we propose a new energy minimization approach which combines an external image energy with an internal shape model energy. Our shape energy uses the Distance From Feature Space (DFFS) concept to allow deviations from the PCA subspace in a theoretically sound and computationally fast way. In contrast to previous approaches, our model does not rely on post-processing with constrained free-form deformation or additional complex local energy models. In addition to the energy minimization approach, we propose a new method for liver detection, a new method for initializing an SSM and an improved k-Nearest Neighbour (kNN)-classifier for boundary detection. Our ASM is evaluated with leave-one-out tests on a data set with 34 tomographic CT scans of the liver and is compared to an ASM with standard shape constraints. The quantitative results of our experiments show that we achieve higher segmentation accuracy with our energy minimization approach than with standard shape constraints.nym

An Efficient Distributed Compressed Sensing Algorithm for Decentralized Sensor Network.

PubMed

Liu, Jing; Huang, Kaiyu; Zhang, Guoxian

2017-04-20

We consider the joint sparsity Model 1 (JSM-1) in a decentralized scenario, where a number of sensors are connected through a network and there is no fusion center. A novel algorithm, named distributed compact sensing matrix pursuit (DCSMP), is proposed to exploit the computational and communication capabilities of the sensor nodes. In contrast to the conventional distributed compressed sensing algorithms adopting a random sensing matrix, the proposed algorithm focuses on the deterministic sensing matrices built directly on the real acquisition systems. The proposed DCSMP algorithm can be divided into two independent parts, the common and innovation support set estimation processes. The goal of the common support set estimation process is to obtain an estimated common support set by fusing the candidate support set information from an individual node and its neighboring nodes. In the following innovation support set estimation process, the measurement vector is projected into a subspace that is perpendicular to the subspace spanned by the columns indexed by the estimated common support set, to remove the impact of the estimated common support set. We can then search the innovation support set using an orthogonal matching pursuit (OMP) algorithm based on the projected measurement vector and projected sensing matrix. In the proposed DCSMP algorithm, the process of estimating the common component/support set is decoupled with that of estimating the innovation component/support set. Thus, the inaccurately estimated common support set will have no impact on estimating the innovation support set. It is proven that under the condition the estimated common support set contains the true common support set, the proposed algorithm can find the true innovation set correctly. Moreover, since the innovation support set estimation process is independent of the common support set estimation process, there is no requirement for the cardinality of both sets; thus, the proposed DCSMP algorithm is capable of tackling the unknown sparsity problem successfully.

Beating the curse of dimension with accurate statistics for the Fokker-Planck equation in complex turbulent systems.

PubMed

Chen, Nan; Majda, Andrew J

2017-12-05

Solving the Fokker-Planck equation for high-dimensional complex dynamical systems is an important issue. Recently, the authors developed efficient statistically accurate algorithms for solving the Fokker-Planck equations associated with high-dimensional nonlinear turbulent dynamical systems with conditional Gaussian structures, which contain many strong non-Gaussian features such as intermittency and fat-tailed probability density functions (PDFs). The algorithms involve a hybrid strategy with a small number of samples [Formula: see text], where a conditional Gaussian mixture in a high-dimensional subspace via an extremely efficient parametric method is combined with a judicious Gaussian kernel density estimation in the remaining low-dimensional subspace. In this article, two effective strategies are developed and incorporated into these algorithms. The first strategy involves a judicious block decomposition of the conditional covariance matrix such that the evolutions of different blocks have no interactions, which allows an extremely efficient parallel computation due to the small size of each individual block. The second strategy exploits statistical symmetry for a further reduction of [Formula: see text] The resulting algorithms can efficiently solve the Fokker-Planck equation with strongly non-Gaussian PDFs in much higher dimensions even with orders in the millions and thus beat the curse of dimension. The algorithms are applied to a [Formula: see text]-dimensional stochastic coupled FitzHugh-Nagumo model for excitable media. An accurate recovery of both the transient and equilibrium non-Gaussian PDFs requires only [Formula: see text] samples! In addition, the block decomposition facilitates the algorithms to efficiently capture the distinct non-Gaussian features at different locations in a [Formula: see text]-dimensional two-layer inhomogeneous Lorenz 96 model, using only [Formula: see text] samples. Copyright © 2017 the Author(s). Published by PNAS.

ANNIT - An Efficient Inversion Algorithm based on Prediction Principles

NASA Astrophysics Data System (ADS)

Růžek, B.; Kolář, P.

2009-04-01

Solution of inverse problems represents meaningful job in geophysics. The amount of data is continuously increasing, methods of modeling are being improved and the computer facilities are also advancing great technical progress. Therefore the development of new and efficient algorithms and computer codes for both forward and inverse modeling is still up to date. ANNIT is contributing to this stream since it is a tool for efficient solution of a set of non-linear equations. Typical geophysical problems are based on parametric approach. The system is characterized by a vector of parameters p, the response of the system is characterized by a vector of data d. The forward problem is usually represented by unique mapping F(p)=d. The inverse problem is much more complex and the inverse mapping p=G(d) is available in an analytical or closed form only exceptionally and generally it may not exist at all. Technically, both forward and inverse mapping F and G are sets of non-linear equations. ANNIT solves such situation as follows: (i) joint subspaces {pD, pM} of original data and model spaces D, M, resp. are searched for, within which the forward mapping F is sufficiently smooth that the inverse mapping G does exist, (ii) numerical approximation of G in subspaces {pD, pM} is found, (iii) candidate solution is predicted by using this numerical approximation. ANNIT is working in an iterative way in cycles. The subspaces {pD, pM} are searched for by generating suitable populations of individuals (models) covering data and model spaces. The approximation of the inverse mapping is made by using three methods: (a) linear regression, (b) Radial Basis Function Network technique, (c) linear prediction (also known as "Kriging"). The ANNIT algorithm has built in also an archive of already evaluated models. Archive models are re-used in a suitable way and thus the number of forward evaluations is minimized. ANNIT is now implemented both in MATLAB and SCILAB. Numerical tests show good performance of the algorithm. Both versions and documentation are available on Internet and anybody can download them. The goal of this presentation is to offer the algorithm and computer codes for anybody interested in the solution to inverse problems.

An efficient shooting algorithm for Evans function calculations in large systems

NASA Astrophysics Data System (ADS)

Humpherys, Jeffrey; Zumbrun, Kevin

2006-08-01

In Evans function computations of the spectra of asymptotically constant-coefficient linear operators, a basic issue is the efficient and numerically stable computation of subspaces evolving according to the associated eigenvalue ODE. For small systems, a fast, shooting algorithm may be obtained by representing subspaces as single exterior products [J.C. Alexander, R. Sachs, Linear instability of solitary waves of a Boussinesq-type equation: A computer assisted computation, Nonlinear World 2 (4) (1995) 471-507; L.Q. Brin, Numerical testing of the stability of viscous shock waves, Ph.D. Thesis, Indiana University, Bloomington, 1998; L.Q. Brin, Numerical testing of the stability of viscous shock waves, Math. Comp. 70 (235) (2001) 1071-1088; L.Q. Brin, K. Zumbrun, Analytically varying eigenvectors and the stability of viscous shock waves, in: Seventh Workshop on Partial Differential Equations, Part I, 2001, Rio de Janeiro, Mat. Contemp. 22 (2002) 19-32; T.J. Bridges, G. Derks, G. Gottwald, Stability and instability of solitary waves of the fifth-order KdV equation: A numerical framework, Physica D 172 (1-4) (2002) 190-216]. For large systems, however, the dimension of the exterior-product space quickly becomes prohibitive, growing as (n/k), where n is the dimension of the system written as a first-order ODE and k (typically ˜n/2) is the dimension of the subspace. We resolve this difficulty by the introduction of a simple polar coordinate algorithm representing “pure” (monomial) products as scalar multiples of orthonormal bases, for which the angular equation is a numerically optimized version of the continuous orthogonalization method of Drury-Davey [A. Davey, An automatic orthonormalization method for solving stiff boundary value problems, J. Comput. Phys. 51 (2) (1983) 343-356; L.O. Drury, Numerical solution of Orr-Sommerfeld-type equations, J. Comput. Phys. 37 (1) (1980) 133-139] and the radial equation is evaluable by quadrature. Notably, the polar-coordinate method preserves the important property of analyticity with respect to parameters.

Improving deep convolutional neural networks with mixed maxout units.

PubMed

Zhao, Hui-Zhen; Liu, Fu-Xian; Li, Long-Yue

2017-01-01

Motivated by insights from the maxout-units-based deep Convolutional Neural Network (CNN) that "non-maximal features are unable to deliver" and "feature mapping subspace pooling is insufficient," we present a novel mixed variant of the recently introduced maxout unit called a mixout unit. Specifically, we do so by calculating the exponential probabilities of feature mappings gained by applying different convolutional transformations over the same input and then calculating the expected values according to their exponential probabilities. Moreover, we introduce the Bernoulli distribution to balance the maximum values with the expected values of the feature mappings subspace. Finally, we design a simple model to verify the pooling ability of mixout units and a Mixout-units-based Network-in-Network (NiN) model to analyze the feature learning ability of the mixout models. We argue that our proposed units improve the pooling ability and that mixout models can achieve better feature learning and classification performance.

Word Spotting and Recognition with Embedded Attributes.

PubMed

Almazán, Jon; Gordo, Albert; Fornés, Alicia; Valveny, Ernest

2014-12-01

This paper addresses the problems of word spotting and word recognition on images. In word spotting, the goal is to find all instances of a query word in a dataset of images. In recognition, the goal is to recognize the content of the word image, usually aided by a dictionary or lexicon. We describe an approach in which both word images and text strings are embedded in a common vectorial subspace. This is achieved by a combination of label embedding and attributes learning, and a common subspace regression. In this subspace, images and strings that represent the same word are close together, allowing one to cast recognition and retrieval tasks as a nearest neighbor problem. Contrary to most other existing methods, our representation has a fixed length, is low dimensional, and is very fast to compute and, especially, to compare. We test our approach on four public datasets of both handwritten documents and natural images showing results comparable or better than the state-of-the-art on spotting and recognition tasks.

Subjective comparison and evaluation of speech enhancement algorithms

PubMed Central

Hu, Yi; Loizou, Philipos C.

2007-01-01

Making meaningful comparisons between the performance of the various speech enhancement algorithms proposed over the years, has been elusive due to lack of a common speech database, differences in the types of noise used and differences in the testing methodology. To facilitate such comparisons, we report on the development of a noisy speech corpus suitable for evaluation of speech enhancement algorithms. This corpus is subsequently used for the subjective evaluation of 13 speech enhancement methods encompassing four classes of algorithms: spectral subtractive, subspace, statistical-model based and Wiener-type algorithms. The subjective evaluation was performed by Dynastat, Inc. using the ITU-T P.835 methodology designed to evaluate the speech quality along three dimensions: signal distortion, noise distortion and overall quality. This paper reports the results of the subjective tests. PMID:18046463

Learning with incomplete information in the committee machine.

PubMed

Bergmann, Urs M; Kühn, Reimer; Stamatescu, Ion-Olimpiu

2009-12-01

We study the problem of learning with incomplete information in a student-teacher setup for the committee machine. The learning algorithm combines unsupervised Hebbian learning of a series of associations with a delayed reinforcement step, in which the set of previously learnt associations is partly and indiscriminately unlearnt, to an extent that depends on the success rate of the student on these previously learnt associations. The relevant learning parameter lambda represents the strength of Hebbian learning. A coarse-grained analysis of the system yields a set of differential equations for overlaps of student and teacher weight vectors, whose solutions provide a complete description of the learning behavior. It reveals complicated dynamics showing that perfect generalization can be obtained if the learning parameter exceeds a threshold lambda ( c ), and if the initial value of the overlap between student and teacher weights is non-zero. In case of convergence, the generalization error exhibits a power law decay as a function of the number of examples used in training, with an exponent that depends on the parameter lambda. An investigation of the system flow in a subspace with broken permutation symmetry between hidden units reveals a bifurcation point lambda* above which perfect generalization does not depend on initial conditions. Finally, we demonstrate that cases of a complexity mismatch between student and teacher are optimally resolved in the sense that an over-complex student can emulate a less complex teacher rule, while an under-complex student reaches a state which realizes the minimal generalization error compatible with the complexity mismatch.

The Wang Landau parallel algorithm for the simple grids. Optimizing OpenMPI parallel implementation

NASA Astrophysics Data System (ADS)

Kussainov, A. S.

2017-12-01

The Wang Landau Monte Carlo algorithm to calculate density of states for the different simple spin lattices was implemented. The energy space was split between the individual threads and balanced according to the expected runtime for the individual processes. Custom spin clustering mechanism, necessary for overcoming of the critical slowdown in the certain energy subspaces, was devised. Stable reconstruction of the density of states was of primary importance. Some data post-processing techniques were involved to produce the expected smooth density of states.

Controller reduction by preserving impulse response energy

NASA Technical Reports Server (NTRS)

Craig, Roy R., Jr.; Su, Tzu-Jeng

1989-01-01

A model order reduction algorithm based on a Krylov recurrence formulation is developed to reduce order of controllers. The reduced-order controller is obtained by projecting the full-order LQG controller onto a Krylov subspace in which either the controllability or the observability grammian is equal to the identity matrix. The reduced-order controller preserves the impulse response energy of the full-order controller and has a parameter-matching property. Two numerical examples drawn from other controller reduction literature are used to illustrate the efficacy of the proposed reduction algorithm.

Novel angle estimation for bistatic MIMO radar using an improved MUSIC

NASA Astrophysics Data System (ADS)

Li, Jianfeng; Zhang, Xiaofei; Chen, Han

2014-09-01

In this article, we study the problem of angle estimation for bistatic multiple-input multiple-output (MIMO) radar and propose an improved multiple signal classification (MUSIC) algorithm for joint direction of departure (DOD) and direction of arrival (DOA) estimation. The proposed algorithm obtains initial estimations of angles obtained from the signal subspace and uses the local one-dimensional peak searches to achieve the joint estimations of DOD and DOA. The angle estimation performance of the proposed algorithm is better than that of estimation of signal parameters via rotational invariance techniques (ESPRIT) algorithm, and is almost the same as that of two-dimensional MUSIC. Furthermore, the proposed algorithm can be suitable for irregular array geometry, obtain automatically paired DOD and DOA estimations, and avoid two-dimensional peak searching. The simulation results verify the effectiveness and improvement of the algorithm.

Transfer Kernel Common Spatial Patterns for Motor Imagery Brain-Computer Interface Classification.

PubMed

Dai, Mengxi; Zheng, Dezhi; Liu, Shucong; Zhang, Pengju

2018-01-01

Motor-imagery-based brain-computer interfaces (BCIs) commonly use the common spatial pattern (CSP) as preprocessing step before classification. The CSP method is a supervised algorithm. Therefore a lot of time-consuming training data is needed to build the model. To address this issue, one promising approach is transfer learning, which generalizes a learning model can extract discriminative information from other subjects for target classification task. To this end, we propose a transfer kernel CSP (TKCSP) approach to learn a domain-invariant kernel by directly matching distributions of source subjects and target subjects. The dataset IVa of BCI Competition III is used to demonstrate the validity by our proposed methods. In the experiment, we compare the classification performance of the TKCSP against CSP, CSP for subject-to-subject transfer (CSP SJ-to-SJ), regularizing CSP (RCSP), stationary subspace CSP (ssCSP), multitask CSP (mtCSP), and the combined mtCSP and ssCSP (ss + mtCSP) method. The results indicate that the superior mean classification performance of TKCSP can achieve 81.14%, especially in case of source subjects with fewer number of training samples. Comprehensive experimental evidence on the dataset verifies the effectiveness and efficiency of the proposed TKCSP approach over several state-of-the-art methods.

Transfer Kernel Common Spatial Patterns for Motor Imagery Brain-Computer Interface Classification

PubMed Central

Dai, Mengxi; Liu, Shucong; Zhang, Pengju

2018-01-01

Motor-imagery-based brain-computer interfaces (BCIs) commonly use the common spatial pattern (CSP) as preprocessing step before classification. The CSP method is a supervised algorithm. Therefore a lot of time-consuming training data is needed to build the model. To address this issue, one promising approach is transfer learning, which generalizes a learning model can extract discriminative information from other subjects for target classification task. To this end, we propose a transfer kernel CSP (TKCSP) approach to learn a domain-invariant kernel by directly matching distributions of source subjects and target subjects. The dataset IVa of BCI Competition III is used to demonstrate the validity by our proposed methods. In the experiment, we compare the classification performance of the TKCSP against CSP, CSP for subject-to-subject transfer (CSP SJ-to-SJ), regularizing CSP (RCSP), stationary subspace CSP (ssCSP), multitask CSP (mtCSP), and the combined mtCSP and ssCSP (ss + mtCSP) method. The results indicate that the superior mean classification performance of TKCSP can achieve 81.14%, especially in case of source subjects with fewer number of training samples. Comprehensive experimental evidence on the dataset verifies the effectiveness and efficiency of the proposed TKCSP approach over several state-of-the-art methods. PMID:29743934

Solving large-scale dynamic systems using band Lanczos method in Rockwell NASTRAN on CRAY X-MP

NASA Technical Reports Server (NTRS)

Gupta, V. K.; Zillmer, S. D.; Allison, R. E.

1986-01-01

The improved cost effectiveness using better models, more accurate and faster algorithms and large scale computing offers more representative dynamic analyses. The band Lanczos eigen-solution method was implemented in Rockwell's version of 1984 COSMIC-released NASTRAN finite element structural analysis computer program to effectively solve for structural vibration modes including those of large complex systems exceeding 10,000 degrees of freedom. The Lanczos vectors were re-orthogonalized locally using the Lanczos Method and globally using the modified Gram-Schmidt method for sweeping rigid-body modes and previously generated modes and Lanczos vectors. The truncated band matrix was solved for vibration frequencies and mode shapes using Givens rotations. Numerical examples are included to demonstrate the cost effectiveness and accuracy of the method as implemented in ROCKWELL NASTRAN. The CRAY version is based on RPK's COSMIC/NASTRAN. The band Lanczos method was more reliable and accurate and converged faster than the single vector Lanczos Method. The band Lanczos method was comparable to the subspace iteration method which was a block version of the inverse power method. However, the subspace matrix tended to be fully populated in the case of subspace iteration and not as sparse as a band matrix.

ERP denoising in multichannel EEG data using contrasts between signal and noise subspaces.

PubMed

Ivannikov, Andriy; Kalyakin, Igor; Hämäläinen, Jarmo; Leppänen, Paavo H T; Ristaniemi, Tapani; Lyytinen, Heikki; Kärkkäinen, Tommi

2009-06-15

In this paper, a new method intended for ERP denoising in multichannel EEG data is discussed. The denoising is done by separating ERP/noise subspaces in multidimensional EEG data by a linear transformation and the following dimension reduction by ignoring noise components during inverse transformation. The separation matrix is found based on the assumption that ERP sources are deterministic for all repetitions of the same type of stimulus within the experiment, while the other noise sources do not obey the determinancy property. A detailed derivation of the technique is given together with the analysis of the results of its application to a real high-density EEG data set. The interpretation of the results and the performance of the proposed method under conditions, when the basic assumptions are violated - e.g. the problem is underdetermined - are also discussed. Moreover, we study how the factors of the number of channels and trials used by the method influence the effectiveness of ERP/noise subspaces separation. In addition, we explore also the impact of different data resampling strategies on the performance of the considered algorithm. The results can help in determining the optimal parameters of the equipment/methods used to elicit and reliably estimate ERPs.

View subspaces for indexing and retrieval of 3D models

NASA Astrophysics Data System (ADS)

Dutagaci, Helin; Godil, Afzal; Sankur, Bülent; Yemez, Yücel

2010-02-01

View-based indexing schemes for 3D object retrieval are gaining popularity since they provide good retrieval results. These schemes are coherent with the theory that humans recognize objects based on their 2D appearances. The viewbased techniques also allow users to search with various queries such as binary images, range images and even 2D sketches. The previous view-based techniques use classical 2D shape descriptors such as Fourier invariants, Zernike moments, Scale Invariant Feature Transform-based local features and 2D Digital Fourier Transform coefficients. These methods describe each object independent of others. In this work, we explore data driven subspace models, such as Principal Component Analysis, Independent Component Analysis and Nonnegative Matrix Factorization to describe the shape information of the views. We treat the depth images obtained from various points of the view sphere as 2D intensity images and train a subspace to extract the inherent structure of the views within a database. We also show the benefit of categorizing shapes according to their eigenvalue spread. Both the shape categorization and data-driven feature set conjectures are tested on the PSB database and compared with the competitor view-based 3D shape retrieval algorithms.

PRIFIRA: General regularization using prior-conditioning for fast radio interferometric imaging†

NASA Astrophysics Data System (ADS)

Naghibzadeh, Shahrzad; van der Veen, Alle-Jan

2018-06-01

Image formation in radio astronomy is a large-scale inverse problem that is inherently ill-posed. We present a general algorithmic framework based on a Bayesian-inspired regularized maximum likelihood formulation of the radio astronomical imaging problem with a focus on diffuse emission recovery from limited noisy correlation data. The algorithm is dubbed PRIor-conditioned Fast Iterative Radio Astronomy (PRIFIRA) and is based on a direct embodiment of the regularization operator into the system by right preconditioning. The resulting system is then solved using an iterative method based on projections onto Krylov subspaces. We motivate the use of a beamformed image (which includes the classical "dirty image") as an efficient prior-conditioner. Iterative reweighting schemes generalize the algorithmic framework and can account for different regularization operators that encourage sparsity of the solution. The performance of the proposed method is evaluated based on simulated one- and two-dimensional array arrangements as well as actual data from the core stations of the Low Frequency Array radio telescope antenna configuration, and compared to state-of-the-art imaging techniques. We show the generality of the proposed method in terms of regularization schemes while maintaining a competitive reconstruction quality with the current reconstruction techniques. Furthermore, we show that exploiting Krylov subspace methods together with the proper noise-based stopping criteria results in a great improvement in imaging efficiency.

Sliding Window Generalized Kernel Affine Projection Algorithm Using Projection Mappings

NASA Astrophysics Data System (ADS)

Slavakis, Konstantinos; Theodoridis, Sergios

2008-12-01

Very recently, a solution to the kernel-based online classification problem has been given by the adaptive projected subgradient method (APSM). The developed algorithm can be considered as a generalization of a kernel affine projection algorithm (APA) and the kernel normalized least mean squares (NLMS). Furthermore, sparsification of the resulting kernel series expansion was achieved by imposing a closed ball (convex set) constraint on the norm of the classifiers. This paper presents another sparsification method for the APSM approach to the online classification task by generating a sequence of linear subspaces in a reproducing kernel Hilbert space (RKHS). To cope with the inherent memory limitations of online systems and to embed tracking capabilities to the design, an upper bound on the dimension of the linear subspaces is imposed. The underlying principle of the design is the notion of projection mappings. Classification is performed by metric projection mappings, sparsification is achieved by orthogonal projections, while the online system's memory requirements and tracking are attained by oblique projections. The resulting sparsification scheme shows strong similarities with the classical sliding window adaptive schemes. The proposed design is validated by the adaptive equalization problem of a nonlinear communication channel, and is compared with classical and recent stochastic gradient descent techniques, as well as with the APSM's solution where sparsification is performed by a closed ball constraint on the norm of the classifiers.

Target detection using the background model from the topological anomaly detection algorithm

NASA Astrophysics Data System (ADS)

Dorado Munoz, Leidy P.; Messinger, David W.; Ziemann, Amanda K.

2013-05-01

The Topological Anomaly Detection (TAD) algorithm has been used as an anomaly detector in hyperspectral and multispectral images. TAD is an algorithm based on graph theory that constructs a topological model of the background in a scene, and computes an anomalousness ranking for all of the pixels in the image with respect to the background in order to identify pixels with uncommon or strange spectral signatures. The pixels that are modeled as background are clustered into groups or connected components, which could be representative of spectral signatures of materials present in the background. Therefore, the idea of using the background components given by TAD in target detection is explored in this paper. In this way, these connected components are characterized in three different approaches, where the mean signature and endmembers for each component are calculated and used as background basis vectors in Orthogonal Subspace Projection (OSP) and Adaptive Subspace Detector (ASD). Likewise, the covariance matrix of those connected components is estimated and used in detectors: Constrained Energy Minimization (CEM) and Adaptive Coherence Estimator (ACE). The performance of these approaches and the different detectors is compared with a global approach, where the background characterization is derived directly from the image. Experiments and results using self-test data set provided as part of the RIT blind test target detection project are shown.

Comparative analysis of different weight matrices in subspace system identification for structural health monitoring

NASA Astrophysics Data System (ADS)

Shokravi, H.; Bakhary, NH

2017-11-01

Subspace System Identification (SSI) is considered as one of the most reliable tools for identification of system parameters. Performance of a SSI scheme is considerably affected by the structure of the associated identification algorithm. Weight matrix is a variable in SSI that is used to reduce the dimensionality of the state-space equation. Generally one of the weight matrices of Principle Component (PC), Unweighted Principle Component (UPC) and Canonical Variate Analysis (CVA) are used in the structure of a SSI algorithm. An increasing number of studies in the field of structural health monitoring are using SSI for damage identification. However, studies that evaluate the performance of the weight matrices particularly in association with accuracy, noise resistance, and time complexity properties are very limited. In this study, the accuracy, noise-robustness, and time-efficiency of the weight matrices are compared using different qualitative and quantitative metrics. Three evaluation metrics of pole analysis, fit values and elapsed time are used in the assessment process. A numerical model of a mass-spring-dashpot and operational data is used in this research paper. It is observed that the principal components obtained using PC algorithms are more robust against noise uncertainty and give more stable results for the pole distribution. Furthermore, higher estimation accuracy is achieved using UPC algorithm. CVA had the worst performance for pole analysis and time efficiency analysis. The superior performance of the UPC algorithm in the elapsed time is attributed to using unit weight matrices. The obtained results demonstrated that the process of reducing dimensionality in CVA and PC has not enhanced the time efficiency but yield an improved modal identification in PC.

On Convergence of Extended Dynamic Mode Decomposition to the Koopman Operator

NASA Astrophysics Data System (ADS)

Korda, Milan; Mezić, Igor

2018-04-01

Extended dynamic mode decomposition (EDMD) (Williams et al. in J Nonlinear Sci 25(6):1307-1346, 2015) is an algorithm that approximates the action of the Koopman operator on an N-dimensional subspace of the space of observables by sampling at M points in the state space. Assuming that the samples are drawn either independently or ergodically from some measure μ , it was shown in Klus et al. (J Comput Dyn 3(1):51-79, 2016) that, in the limit as M→ ∞, the EDMD operator K_{N,M} converges to K_N, where K_N is the L_2(μ )-orthogonal projection of the action of the Koopman operator on the finite-dimensional subspace of observables. We show that, as N → ∞, the operator K_N converges in the strong operator topology to the Koopman operator. This in particular implies convergence of the predictions of future values of a given observable over any finite time horizon, a fact important for practical applications such as forecasting, estimation and control. In addition, we show that accumulation points of the spectra of K_N correspond to the eigenvalues of the Koopman operator with the associated eigenfunctions converging weakly to an eigenfunction of the Koopman operator, provided that the weak limit of the eigenfunctions is nonzero. As a by-product, we propose an analytic version of the EDMD algorithm which, under some assumptions, allows one to construct K_N directly, without the use of sampling. Finally, under additional assumptions, we analyze convergence of K_{N,N} (i.e., M=N), proving convergence, along a subsequence, to weak eigenfunctions (or eigendistributions) related to the eigenmeasures of the Perron-Frobenius operator. No assumptions on the observables belonging to a finite-dimensional invariant subspace of the Koopman operator are required throughout.

Accelerating the weighted histogram analysis method by direct inversion in the iterative subspace.

PubMed

Zhang, Cheng; Lai, Chun-Liang; Pettitt, B Montgomery

The weighted histogram analysis method (WHAM) for free energy calculations is a valuable tool to produce free energy differences with the minimal errors. Given multiple simulations, WHAM obtains from the distribution overlaps the optimal statistical estimator of the density of states, from which the free energy differences can be computed. The WHAM equations are often solved by an iterative procedure. In this work, we use a well-known linear algebra algorithm which allows for more rapid convergence to the solution. We find that the computational complexity of the iterative solution to WHAM and the closely-related multiple Bennett acceptance ratio (MBAR) method can be improved by using the method of direct inversion in the iterative subspace. We give examples from a lattice model, a simple liquid and an aqueous protein solution.

Using parallel banded linear system solvers in generalized eigenvalue problems

NASA Technical Reports Server (NTRS)

Zhang, Hong; Moss, William F.

1993-01-01

Subspace iteration is a reliable and cost effective method for solving positive definite banded symmetric generalized eigenproblems, especially in the case of large scale problems. This paper discusses an algorithm that makes use of two parallel banded solvers in subspace iteration. A shift is introduced to decompose the banded linear systems into relatively independent subsystems and to accelerate the iterations. With this shift, an eigenproblem is mapped efficiently into the memories of a multiprocessor and a high speed-up is obtained for parallel implementations. An optimal shift is a shift that balances total computation and communication costs. Under certain conditions, we show how to estimate an optimal shift analytically using the decay rate for the inverse of a banded matrix, and how to improve this estimate. Computational results on iPSC/2 and iPSC/860 multiprocessors are presented.

Heuristic approach to image registration

NASA Astrophysics Data System (ADS)

Gertner, Izidor; Maslov, Igor V.

2000-08-01

Image registration, i.e. correct mapping of images obtained from different sensor readings onto common reference frame, is a critical part of multi-sensor ATR/AOR systems based on readings from different types of sensors. In order to fuse two different sensor readings of the same object, the readings have to be put into a common coordinate system. This task can be formulated as optimization problem in a space of all possible affine transformations of an image. In this paper, a combination of heuristic methods is explored to register gray- scale images. The modification of Genetic Algorithm is used as the first step in global search for optimal transformation. It covers the entire search space with (randomly or heuristically) scattered probe points and helps significantly reduce the search space to a subspace of potentially most successful transformations. Due to its discrete character, however, Genetic Algorithm in general can not converge while coming close to the optimum. Its termination point can be specified either as some predefined number of generations or as achievement of a certain acceptable convergence level. To refine the search, potential optimal subspaces are searched using more delicate and efficient for local search Taboo and Simulated Annealing methods.

Fuzzy Subspace Clustering

NASA Astrophysics Data System (ADS)

Borgelt, Christian

In clustering we often face the situation that only a subset of the available attributes is relevant for forming clusters, even though this may not be known beforehand. In such cases it is desirable to have a clustering algorithm that automatically weights attributes or even selects a proper subset. In this paper I study such an approach for fuzzy clustering, which is based on the idea to transfer an alternative to the fuzzifier (Klawonn and Höppner, What is fuzzy about fuzzy clustering? Understanding and improving the concept of the fuzzifier, In: Proc. 5th Int. Symp. on Intelligent Data Analysis, 254-264, Springer, Berlin, 2003) to attribute weighting fuzzy clustering (Keller and Klawonn, Int J Uncertain Fuzziness Knowl Based Syst 8:735-746, 2000). In addition, by reformulating Gustafson-Kessel fuzzy clustering, a scheme for weighting and selecting principal axes can be obtained. While in Borgelt (Feature weighting and feature selection in fuzzy clustering, In: Proc. 17th IEEE Int. Conf. on Fuzzy Systems, IEEE Press, Piscataway, NJ, 2008) I already presented such an approach for a global selection of attributes and principal axes, this paper extends it to a cluster-specific selection, thus arriving at a fuzzy subspace clustering algorithm (Parsons, Haque, and Liu, 2004).

Decentralized system identification using stochastic subspace identification on wireless smart sensor networks

NASA Astrophysics Data System (ADS)

Sim, Sung-Han; Spencer, Billie F., Jr.; Park, Jongwoong; Jung, Hyungjo

2012-04-01

Wireless Smart Sensor Networks (WSSNs) facilitates a new paradigm to structural identification and monitoring for civil infrastructure. Conventional monitoring systems based on wired sensors and centralized data acquisition and processing have been considered to be challenging and costly due to cabling and expensive equipment and maintenance costs. WSSNs have emerged as a technology that can overcome such difficulties, making deployment of a dense array of sensors on large civil structures both feasible and economical. However, as opposed to wired sensor networks in which centralized data acquisition and processing is common practice, WSSNs require decentralized computing algorithms to reduce data transmission due to the limitation associated with wireless communication. Thus, several system identification methods have been implemented to process sensor data and extract essential information, including Natural Excitation Technique with Eigensystem Realization Algorithm, Frequency Domain Decomposition (FDD), and Random Decrement Technique (RDT); however, Stochastic Subspace Identification (SSI) has not been fully utilized in WSSNs, while SSI has the strong potential to enhance the system identification. This study presents a decentralized system identification using SSI in WSSNs. The approach is implemented on MEMSIC's Imote2 sensor platform and experimentally verified using a 5-story shear building model.

Separation and imaging diffractions by a sparsity-promoting model and subspace trust-region algorithm

NASA Astrophysics Data System (ADS)

Yu, Caixia; Zhao, Jingtao; Wang, Yanfei; Wang, Chengxiang; Geng, Weifeng

2017-03-01

The small-scale geologic inhomogeneities or discontinuities, such as tiny faults, cavities or fractures, generally have spatial scales comparable to or even smaller than the seismic wavelength. Therefore, the seismic responses of these objects are coded in diffractions and an attempt to high-resolution imaging can be made if we can appropriately image them. As the amplitudes of reflections can be several orders of magnitude larger than those of diffractions, one of the key problems of diffraction imaging is to suppress reflections and at the same time to preserve diffractions. A sparsity-promoting method for separating diffractions in the common-offset domain is proposed that uses the Kirchhoff integral formula to enforce the sparsity of diffractions and the linear Radon transform to formulate reflections. A subspace trust-region algorithm that can provide globally convergent solutions is employed for solving this large-scale computation problem. The method not only allows for separation of diffractions in the case of interfering events but also ensures a high fidelity of the separated diffractions. Numerical experiment and field application demonstrate the good performance of the proposed method in imaging the small-scale geological features related to the migration channel and storage spaces of carbonate reservoirs.

Development of Xi'an-CI package – applying the hole–particle symmetry in multi-reference electronic correlation calculations

NASA Astrophysics Data System (ADS)

Suo, Bingbing; Lei, Yibo; Han, Huixian; Wang, Yubin

2018-04-01

This mini-review introduces our works on the Xi'an-CI (configuration interaction) package using graphical unitary group approach (GUGA). Taking advantage of the hole-particle symmetry in GUGA, the Galfand states used to span the CI space are classified into CI subspaces according to the number of holes and particles, and the coupling coefficients used to calculate Hamiltonian matrix elements could be factorised into the segment factors in the hole, active and external spaces. An efficient multi-reference CI with single and double excitations (MRCISD) algorithm is thus developed that reduces the storage requirement and increases the number of correlated electrons significantly. The hole-particle symmetry also gives rise to a doubly contracted MRCISD approach. Moreover, the internally contracted Gelfand states are defined within the CI subspace arising from the hole-particle symmetry, which makes the implementation of internally contracted MRCISD in the framework of GUGA possible. In addition to MRCISD, the development of multi-reference second-order perturbation theory (MRPT2) also benefits from the hole-particle symmetry. A configuration-based MRPT2 algorithm is proposed and extended to the multi-state n-electron valence-state second-order perturbation theory.

A comparative study of linear and nonlinear anomaly detectors for hyperspectral imagery

NASA Astrophysics Data System (ADS)

Goldberg, Hirsh; Nasrabadi, Nasser M.

2007-04-01

In this paper we implement various linear and nonlinear subspace-based anomaly detectors for hyperspectral imagery. First, a dual window technique is used to separate the local area around each pixel into two regions - an inner-window region (IWR) and an outer-window region (OWR). Pixel spectra from each region are projected onto a subspace which is defined by projection bases that can be generated in several ways. Here we use three common pattern classification techniques (Principal Component Analysis (PCA), Fisher Linear Discriminant (FLD) Analysis, and the Eigenspace Separation Transform (EST)) to generate projection vectors. In addition to these three algorithms, the well-known Reed-Xiaoli (RX) anomaly detector is also implemented. Each of the four linear methods is then implicitly defined in a high- (possibly infinite-) dimensional feature space by using a nonlinear mapping associated with a kernel function. Using a common machine-learning technique known as the kernel trick all dot products in the feature space are replaced with a Mercer kernel function defined in terms of the original input data space. To determine how anomalous a given pixel is, we then project the current test pixel spectra and the spectral mean vector of the OWR onto the linear and nonlinear projection vectors in order to exploit the statistical differences between the IWR and OWR pixels. Anomalies are detected if the separation of the projection of the current test pixel spectra and the OWR mean spectra are greater than a certain threshold. Comparisons are made using receiver operating characteristics (ROC) curves.

A hybrid approach to generating search subspaces in dynamically constrained 4-dimensional data assimilation

NASA Astrophysics Data System (ADS)

Yaremchuk, Max; Martin, Paul; Beattie, Christopher

2017-09-01

Development and maintenance of the linearized and adjoint code for advanced circulation models is a challenging issue, requiring a significant proportion of total effort in operational data assimilation (DA). The ensemble-based DA techniques provide a derivative-free alternative, which appears to be competitive with variational methods in many practical applications. This article proposes a hybrid scheme for generating the search subspaces in the adjoint-free 4-dimensional DA method (a4dVar) that does not use a predefined ensemble. The method resembles 4dVar in that the optimal solution is strongly constrained by model dynamics and search directions are supplied iteratively using information from the current and previous model trajectories generated in the process of optimization. In contrast to 4dVar, which produces a single search direction from exact gradient information, a4dVar employs an ensemble of directions to form a subspace in order to proceed. In the earlier versions of a4dVar, search subspaces were built using the leading EOFs of either the model trajectory or the projections of the model-data misfits onto the range of the background error covariance (BEC) matrix at the current iteration. In the present study, we blend both approaches and explore a hybrid scheme of ensemble generation in order to improve the performance and flexibility of the algorithm. In addition, we introduce balance constraints into the BEC structure and periodically augment the search ensemble with BEC eigenvectors to avoid repeating minimization over already explored subspaces. Performance of the proposed hybrid a4dVar (ha4dVar) method is compared with that of standard 4dVar in a realistic regional configuration assimilating real data into the Navy Coastal Ocean Model (NCOM). It is shown that the ha4dVar converges faster than a4dVar and can be potentially competitive with 4dvar both in terms of the required computational time and the forecast skill.

Quantum search algorithms on a regular lattice

NASA Astrophysics Data System (ADS)

Hein, Birgit; Tanner, Gregor

2010-07-01

Quantum algorithms for searching for one or more marked items on a d-dimensional lattice provide an extension of Grover’s search algorithm including a spatial component. We demonstrate that these lattice search algorithms can be viewed in terms of the level dynamics near an avoided crossing of a one-parameter family of quantum random walks. We give approximations for both the level splitting at the avoided crossing and the effectively two-dimensional subspace of the full Hilbert space spanning the level crossing. This makes it possible to give the leading order behavior for the search time and the localization probability in the limit of large lattice size including the leading order coefficients. For d=2 and d=3, these coefficients are calculated explicitly. Closed form expressions are given for higher dimensions.

Improving deep convolutional neural networks with mixed maxout units

PubMed Central

Liu, Fu-xian; Li, Long-yue

2017-01-01

Motivated by insights from the maxout-units-based deep Convolutional Neural Network (CNN) that “non-maximal features are unable to deliver” and “feature mapping subspace pooling is insufficient,” we present a novel mixed variant of the recently introduced maxout unit called a mixout unit. Specifically, we do so by calculating the exponential probabilities of feature mappings gained by applying different convolutional transformations over the same input and then calculating the expected values according to their exponential probabilities. Moreover, we introduce the Bernoulli distribution to balance the maximum values with the expected values of the feature mappings subspace. Finally, we design a simple model to verify the pooling ability of mixout units and a Mixout-units-based Network-in-Network (NiN) model to analyze the feature learning ability of the mixout models. We argue that our proposed units improve the pooling ability and that mixout models can achieve better feature learning and classification performance. PMID:28727737

An Improved Ensemble Learning Method for Classifying High-Dimensional and Imbalanced Biomedicine Data.

PubMed

Yu, Hualong; Ni, Jun

2014-01-01

Training classifiers on skewed data can be technically challenging tasks, especially if the data is high-dimensional simultaneously, the tasks can become more difficult. In biomedicine field, skewed data type often appears. In this study, we try to deal with this problem by combining asymmetric bagging ensemble classifier (asBagging) that has been presented in previous work and an improved random subspace (RS) generation strategy that is called feature subspace (FSS). Specifically, FSS is a novel method to promote the balance level between accuracy and diversity of base classifiers in asBagging. In view of the strong generalization capability of support vector machine (SVM), we adopt it to be base classifier. Extensive experiments on four benchmark biomedicine data sets indicate that the proposed ensemble learning method outperforms many baseline approaches in terms of Accuracy, F-measure, G-mean and AUC evaluation criterions, thus it can be regarded as an effective and efficient tool to deal with high-dimensional and imbalanced biomedical data.

Randomized Subspace Learning for Proline Cis-Trans Isomerization Prediction.

PubMed

Al-Jarrah, Omar Y; Yoo, Paul D; Taha, Kamal; Muhaidat, Sami; Shami, Abdallah; Zaki, Nazar

2015-01-01

Proline residues are common source of kinetic complications during folding. The X-Pro peptide bond is the only peptide bond for which the stability of the cis and trans conformations is comparable. The cis-trans isomerization (CTI) of X-Pro peptide bonds is a widely recognized rate-limiting factor, which can not only induces additional slow phases in protein folding but also modifies the millisecond and sub-millisecond dynamics of the protein. An accurate computational prediction of proline CTI is of great importance for the understanding of protein folding, splicing, cell signaling, and transmembrane active transport in both the human body and animals. In our earlier work, we successfully developed a biophysically motivated proline CTI predictor utilizing a novel tree-based consensus model with a powerful metalearning technique and achieved 86.58 percent Q2 accuracy and 0.74 Mcc, which is a better result than the results (70-73 percent Q2 accuracies) reported in the literature on the well-referenced benchmark dataset. In this paper, we describe experiments with novel randomized subspace learning and bootstrap seeding techniques as an extension to our earlier work, the consensus models as well as entropy-based learning methods, to obtain better accuracy through a precise and robust learning scheme for proline CTI prediction.

Using parallel evolutionary development for a biologically-inspired computer vision system for mobile robots.

PubMed

Wright, Cameron H G; Barrett, Steven F; Pack, Daniel J

2005-01-01

We describe a new approach to attacking the problem of robust computer vision for mobile robots. The overall strategy is to mimic the biological evolution of animal vision systems. Our basic imaging sensor is based upon the eye of the common house fly, Musca domestica. The computational algorithms are a mix of traditional image processing, subspace techniques, and multilayer neural networks.

Noise facilitation in associative memories of exponential capacity.

PubMed

Karbasi, Amin; Salavati, Amir Hesam; Shokrollahi, Amin; Varshney, Lav R

2014-11-01

Recent advances in associative memory design through structured pattern sets and graph-based inference algorithms have allowed reliable learning and recall of an exponential number of patterns that satisfy certain subspace constraints. Although these designs correct external errors in recall, they assume neurons that compute noiselessly, in contrast to the highly variable neurons in brain regions thought to operate associatively, such as hippocampus and olfactory cortex. Here we consider associative memories with boundedly noisy internal computations and analytically characterize performance. As long as the internal noise level is below a specified threshold, the error probability in the recall phase can be made exceedingly small. More surprising, we show that internal noise improves the performance of the recall phase while the pattern retrieval capacity remains intact: the number of stored patterns does not reduce with noise (up to a threshold). Computational experiments lend additional support to our theoretical analysis. This work suggests a functional benefit to noisy neurons in biological neuronal networks.

Joint Feature Extraction and Classifier Design for ECG-Based Biometric Recognition.

PubMed

Gutta, Sandeep; Cheng, Qi

2016-03-01

Traditional biometric recognition systems often utilize physiological traits such as fingerprint, face, iris, etc. Recent years have seen a growing interest in electrocardiogram (ECG)-based biometric recognition techniques, especially in the field of clinical medicine. In existing ECG-based biometric recognition methods, feature extraction and classifier design are usually performed separately. In this paper, a multitask learning approach is proposed, in which feature extraction and classifier design are carried out simultaneously. Weights are assigned to the features within the kernel of each task. We decompose the matrix consisting of all the feature weights into sparse and low-rank components. The sparse component determines the features that are relevant to identify each individual, and the low-rank component determines the common feature subspace that is relevant to identify all the subjects. A fast optimization algorithm is developed, which requires only the first-order information. The performance of the proposed approach is demonstrated through experiments using the MIT-BIH Normal Sinus Rhythm database.

Greedy subspace clustering.

DOT National Transportation Integrated Search

2016-09-01

We consider the problem of subspace clustering: given points that lie on or near the union of many low-dimensional linear subspaces, recover the subspaces. To this end, one first identifies sets of points close to the same subspace and uses the sets ...

Solute transport with multiple equilibrium-controlled or kinetically controlled chemical reactions

USGS Publications Warehouse

Friedly, John C.; Rubin, Jacob

1992-01-01

A new approach is applied to the problem of modeling solute transport accompanied by many chemical reactions. The approach, based on concepts of the concentration space and its stoichiometric subspaces, uses elements of the subspaces as primary dependent variables. It is shown that the resulting model equations are compact in form, isolate the chemical reaction expressions from flow expressions, and can be used for either equilibrium or kinetically controlled reactions. The implications of the results on numerical algorithms for solving the equations are discussed. The application of the theory is illustrated throughout with examples involving a simple but broadly representative set of reactions previously considered in the literature. Numerical results are presented for four interconnected reactions: a homogeneous complexation reaction, two sorption reactions, and a dissolution/precipitation reaction. Three cases are considered: (1) four kinetically controlled reactions, (2) four equilibrium-controlled reactions, and (3) a system with two kinetically controlled reactions and two equilibrium-controlled reactions.

Krylov subspace methods on supercomputers

NASA Technical Reports Server (NTRS)

Saad, Youcef

1988-01-01

A short survey of recent research on Krylov subspace methods with emphasis on implementation on vector and parallel computers is presented. Conjugate gradient methods have proven very useful on traditional scalar computers, and their popularity is likely to increase as three-dimensional models gain importance. A conservative approach to derive effective iterative techniques for supercomputers has been to find efficient parallel/vector implementations of the standard algorithms. The main source of difficulty in the incomplete factorization preconditionings is in the solution of the triangular systems at each step. A few approaches consisting of implementing efficient forward and backward triangular solutions are described in detail. Polynomial preconditioning as an alternative to standard incomplete factorization techniques is also discussed. Another efficient approach is to reorder the equations so as to improve the structure of the matrix to achieve better parallelism or vectorization. An overview of these and other ideas and their effectiveness or potential for different types of architectures is given.

Feedforward Inhibition and Synaptic Scaling – Two Sides of the Same Coin?

PubMed Central

Lücke, Jörg

2012-01-01

Feedforward inhibition and synaptic scaling are important adaptive processes that control the total input a neuron can receive from its afferents. While often studied in isolation, the two have been reported to co-occur in various brain regions. The functional implications of their interactions remain unclear, however. Based on a probabilistic modeling approach, we show here that fast feedforward inhibition and synaptic scaling interact synergistically during unsupervised learning. In technical terms, we model the input to a neural circuit using a normalized mixture model with Poisson noise. We demonstrate analytically and numerically that, in the presence of lateral inhibition introducing competition between different neurons, Hebbian plasticity and synaptic scaling approximate the optimal maximum likelihood solutions for this model. Our results suggest that, beyond its conventional use as a mechanism to remove undesired pattern variations, input normalization can make typical neural interaction and learning rules optimal on the stimulus subspace defined through feedforward inhibition. Furthermore, learning within this subspace is more efficient in practice, as it helps avoid locally optimal solutions. Our results suggest a close connection between feedforward inhibition and synaptic scaling which may have important functional implications for general cortical processing. PMID:22457610

Feedforward inhibition and synaptic scaling--two sides of the same coin?

PubMed

Keck, Christian; Savin, Cristina; Lücke, Jörg

2012-01-01

Feedforward inhibition and synaptic scaling are important adaptive processes that control the total input a neuron can receive from its afferents. While often studied in isolation, the two have been reported to co-occur in various brain regions. The functional implications of their interactions remain unclear, however. Based on a probabilistic modeling approach, we show here that fast feedforward inhibition and synaptic scaling interact synergistically during unsupervised learning. In technical terms, we model the input to a neural circuit using a normalized mixture model with Poisson noise. We demonstrate analytically and numerically that, in the presence of lateral inhibition introducing competition between different neurons, Hebbian plasticity and synaptic scaling approximate the optimal maximum likelihood solutions for this model. Our results suggest that, beyond its conventional use as a mechanism to remove undesired pattern variations, input normalization can make typical neural interaction and learning rules optimal on the stimulus subspace defined through feedforward inhibition. Furthermore, learning within this subspace is more efficient in practice, as it helps avoid locally optimal solutions. Our results suggest a close connection between feedforward inhibition and synaptic scaling which may have important functional implications for general cortical processing.

Multigrid techniques for nonlinear eigenvalue probems: Solutions of a nonlinear Schroedinger eigenvalue problem in 2D and 3D

NASA Technical Reports Server (NTRS)

Costiner, Sorin; Taasan, Shlomo

1994-01-01

This paper presents multigrid (MG) techniques for nonlinear eigenvalue problems (EP) and emphasizes an MG algorithm for a nonlinear Schrodinger EP. The algorithm overcomes the mentioned difficulties combining the following techniques: an MG projection coupled with backrotations for separation of solutions and treatment of difficulties related to clusters of close and equal eigenvalues; MG subspace continuation techniques for treatment of the nonlinearity; an MG simultaneous treatment of the eigenvectors at the same time with the nonlinearity and with the global constraints. The simultaneous MG techniques reduce the large number of self consistent iterations to only a few or one MG simultaneous iteration and keep the solutions in a right neighborhood where the algorithm converges fast.

Mining biological information from 3D short time-series gene expression data: the OPTricluster algorithm.

PubMed

Tchagang, Alain B; Phan, Sieu; Famili, Fazel; Shearer, Heather; Fobert, Pierre; Huang, Yi; Zou, Jitao; Huang, Daiqing; Cutler, Adrian; Liu, Ziying; Pan, Youlian

2012-04-04

Nowadays, it is possible to collect expression levels of a set of genes from a set of biological samples during a series of time points. Such data have three dimensions: gene-sample-time (GST). Thus they are called 3D microarray gene expression data. To take advantage of the 3D data collected, and to fully understand the biological knowledge hidden in the GST data, novel subspace clustering algorithms have to be developed to effectively address the biological problem in the corresponding space. We developed a subspace clustering algorithm called Order Preserving Triclustering (OPTricluster), for 3D short time-series data mining. OPTricluster is able to identify 3D clusters with coherent evolution from a given 3D dataset using a combinatorial approach on the sample dimension, and the order preserving (OP) concept on the time dimension. The fusion of the two methodologies allows one to study similarities and differences between samples in terms of their temporal expression profile. OPTricluster has been successfully applied to four case studies: immune response in mice infected by malaria (Plasmodium chabaudi), systemic acquired resistance in Arabidopsis thaliana, similarities and differences between inner and outer cotyledon in Brassica napus during seed development, and to Brassica napus whole seed development. These studies showed that OPTricluster is robust to noise and is able to detect the similarities and differences between biological samples. Our analysis showed that OPTricluster generally outperforms other well known clustering algorithms such as the TRICLUSTER, gTRICLUSTER and K-means; it is robust to noise and can effectively mine the biological knowledge hidden in the 3D short time-series gene expression data.

Intelligent measurement and compensation of linear motor force ripple: a projection-based learning approach in the presence of noise

NASA Astrophysics Data System (ADS)

Liu, Yang; Song, Fazhi; Yang, Xiaofeng; Dong, Yue; Tan, Jiubin

2018-06-01

Due to their structural simplicity, linear motors are increasingly receiving attention for use in high velocity and high precision applications. The force ripple, as a space-periodic disturbance, however, would deteriorate the achievable dynamic performance. Conventional force ripple measurement approaches are time-consuming and have high requirements on the experimental conditions. In this paper, a novel learning identification algorithm is proposed for force ripple intelligent measurement and compensation. Existing identification schemes always use all the error signals to update the parameters in the force ripple. However, the error induced by noise is non-effective for force ripple identification, and even deteriorates the identification process. In this paper only the most pertinent information in the error signal is utilized for force ripple identification. Firstly, the effective error signals caused by the reference trajectory and the force ripple are extracted by projecting the overall error signals onto a subspace spanned by the physical model of the linear motor as well as the sinusoidal model of the force ripple. The time delay in the linear motor is compensated in the basis functions. Then, a data-driven approach is proposed to design the learning gain. It balances the trade-off between convergence speed and robustness against noise. Simulation and experimental results validate the proposed method and confirm its effectiveness and superiority.

Verification Techniques for Parameter Selection and Bayesian Model Calibration Presented for an HIV Model

NASA Astrophysics Data System (ADS)

Wentworth, Mami Tonoe

Uncertainty quantification plays an important role when making predictive estimates of model responses. In this context, uncertainty quantification is defined as quantifying and reducing uncertainties, and the objective is to quantify uncertainties in parameter, model and measurements, and propagate the uncertainties through the model, so that one can make a predictive estimate with quantified uncertainties. Two of the aspects of uncertainty quantification that must be performed prior to propagating uncertainties are model calibration and parameter selection. There are several efficient techniques for these processes; however, the accuracy of these methods are often not verified. This is the motivation for our work, and in this dissertation, we present and illustrate verification frameworks for model calibration and parameter selection in the context of biological and physical models. First, HIV models, developed and improved by [2, 3, 8], describe the viral infection dynamics of an HIV disease. These are also used to make predictive estimates of viral loads and T-cell counts and to construct an optimal control for drug therapy. Estimating input parameters is an essential step prior to uncertainty quantification. However, not all the parameters are identifiable, implying that they cannot be uniquely determined by the observations. These unidentifiable parameters can be partially removed by performing parameter selection, a process in which parameters that have minimal impacts on the model response are determined. We provide verification techniques for Bayesian model calibration and parameter selection for an HIV model. As an example of a physical model, we employ a heat model with experimental measurements presented in [10]. A steady-state heat model represents a prototypical behavior for heat conduction and diffusion process involved in a thermal-hydraulic model, which is a part of nuclear reactor models. We employ this simple heat model to illustrate verification techniques for model calibration. For Bayesian model calibration, we employ adaptive Metropolis algorithms to construct densities for input parameters in the heat model and the HIV model. To quantify the uncertainty in the parameters, we employ two MCMC algorithms: Delayed Rejection Adaptive Metropolis (DRAM) [33] and Differential Evolution Adaptive Metropolis (DREAM) [66, 68]. The densities obtained using these methods are compared to those obtained through the direct numerical evaluation of the Bayes' formula. We also combine uncertainties in input parameters and measurement errors to construct predictive estimates for a model response. A significant emphasis is on the development and illustration of techniques to verify the accuracy of sampling-based Metropolis algorithms. We verify the accuracy of DRAM and DREAM by comparing chains, densities and correlations obtained using DRAM, DREAM and the direct evaluation of Bayes formula. We also perform similar analysis for credible and prediction intervals for responses. Once the parameters are estimated, we employ energy statistics test [63, 64] to compare the densities obtained by different methods for the HIV model. The energy statistics are used to test the equality of distributions. We also consider parameter selection and verification techniques for models having one or more parameters that are noninfluential in the sense that they minimally impact model outputs. We illustrate these techniques for a dynamic HIV model but note that the parameter selection and verification framework is applicable to a wide range of biological and physical models. To accommodate the nonlinear input to output relations, which are typical for such models, we focus on global sensitivity analysis techniques, including those based on partial correlations, Sobol indices based on second-order model representations, and Morris indices, as well as a parameter selection technique based on standard errors. A significant objective is to provide verification strategies to assess the accuracy of those techniques, which we illustrate in the context of the HIV model. Finally, we examine active subspace methods as an alternative to parameter subset selection techniques. The objective of active subspace methods is to determine the subspace of inputs that most strongly affect the model response, and to reduce the dimension of the input space. The major difference between active subspace methods and parameter selection techniques is that parameter selection identifies influential parameters whereas subspace selection identifies a linear combination of parameters that impacts the model responses significantly. We employ active subspace methods discussed in [22] for the HIV model and present a verification that the active subspace successfully reduces the input dimensions.

Recognition of explosives fingerprints on objects for courier services using machine learning methods and laser-induced breakdown spectroscopy.

PubMed

Moros, J; Serrano, J; Gallego, F J; Macías, J; Laserna, J J

2013-06-15

During recent years laser-induced breakdown spectroscopy (LIBS) has been considered one of the techniques with larger ability for trace detection of explosives. However, despite of the high sensitivity exhibited for this application, LIBS suffers from a limited selectivity due to difficulties in assigning the molecular origin of the spectral emissions observed. This circumstance makes the recognition of fingerprints a latent challenging problem. In the present manuscript the sorting of six explosives (chloratite, ammonal, DNT, TNT, RDX and PETN) against a broad list of potential harmless interferents (butter, fuel oil, hand cream, olive oil, …), all of them in the form of fingerprints deposited on the surfaces of objects for courier services, has been carried out. When LIBS information is processed through a multi-stage architecture algorithm built from a suitable combination of 3 learning classifiers, an unknown fingerprint may be labeled into a particular class. Neural network classifiers trained by the Levenberg-Marquardt rule were decided within 3D scatter plots projected onto the subspace of the most useful features extracted from the LIBS spectra. Experimental results demonstrate that the presented algorithm sorts fingerprints according to their hazardous character, although its spectral information is virtually identical in appearance, with rates of false negatives and false positives not beyond of 10%. These reported achievements mean a step forward in the technology readiness level of LIBS for this complex application related to defense, homeland security and force protection. Copyright © 2013 Elsevier B.V. All rights reserved.

Macromolecule mapping of the brain using ultrashort-TE acquisition and reference-based metabolite removal.

PubMed

Lam, Fan; Li, Yudu; Clifford, Bryan; Liang, Zhi-Pei

2018-05-01

To develop a practical method for mapping macromolecule distribution in the brain using ultrashort-TE MRSI data. An FID-based chemical shift imaging acquisition without metabolite-nulling pulses was used to acquire ultrashort-TE MRSI data that capture the macromolecule signals with high signal-to-noise-ratio (SNR) efficiency. To remove the metabolite signals from the ultrashort-TE data, single voxel spectroscopy data were obtained to determine a set of high-quality metabolite reference spectra. These spectra were then incorporated into a generalized series (GS) model to represent general metabolite spatiospectral distributions. A time-segmented algorithm was developed to back-extrapolate the GS model-based metabolite distribution from truncated FIDs and remove it from the MRSI data. Numerical simulations and in vivo experiments have been performed to evaluate the proposed method. Simulation results demonstrate accurate metabolite signal extrapolation by the proposed method given a high-quality reference. For in vivo experiments, the proposed method is able to produce spatiospectral distributions of macromolecules in the brain with high SNR from data acquired in about 10 minutes. We further demonstrate that the high-dimensional macromolecule spatiospectral distribution resides in a low-dimensional subspace. This finding provides a new opportunity to use subspace models for quantification and accelerated macromolecule mapping. Robustness of the proposed method is also demonstrated using multiple data sets from the same and different subjects. The proposed method is able to obtain macromolecule distributions in the brain from ultrashort-TE acquisitions. It can also be used for acquiring training data to determine a low-dimensional subspace to represent the macromolecule signals for subspace-based MRSI. Magn Reson Med 79:2460-2469, 2018. © 2017 International Society for Magnetic Resonance in Medicine. © 2017 International Society for Magnetic Resonance in Medicine.

Semiannual Report for Contract NAS1-19480 (Institute for Computer Applications in Science and Engineering)

DTIC Science & Technology

1994-06-01

algorithms for large, irreducibly coupled systems iteratively solve concurrent problems within different subspaces of a Hilbert space, or within different...effective on problems amenable to SIMD solution. Together with researchers at AT&T Bell Labs (Boris Lubachevsky, Albert Greenberg ) we have developed...reasonable measurement. In the study of different speedups, various causes of superlinear speedup are also presented. Greenberg , Albert G., Boris D

An Extended Spectral-Spatial Classification Approach for Hyperspectral Data

NASA Astrophysics Data System (ADS)

Akbari, D.

2017-11-01

In this paper an extended classification approach for hyperspectral imagery based on both spectral and spatial information is proposed. The spatial information is obtained by an enhanced marker-based minimum spanning forest (MSF) algorithm. Three different methods of dimension reduction are first used to obtain the subspace of hyperspectral data: (1) unsupervised feature extraction methods including principal component analysis (PCA), independent component analysis (ICA), and minimum noise fraction (MNF); (2) supervised feature extraction including decision boundary feature extraction (DBFE), discriminate analysis feature extraction (DAFE), and nonparametric weighted feature extraction (NWFE); (3) genetic algorithm (GA). The spectral features obtained are then fed into the enhanced marker-based MSF classification algorithm. In the enhanced MSF algorithm, the markers are extracted from the classification maps obtained by both SVM and watershed segmentation algorithm. To evaluate the proposed approach, the Pavia University hyperspectral data is tested. Experimental results show that the proposed approach using GA achieves an approximately 8 % overall accuracy higher than the original MSF-based algorithm.

Indoor Subspacing to Implement Indoorgml for Indoor Navigation

NASA Astrophysics Data System (ADS)

Jung, H.; Lee, J.

2015-10-01

According to an increasing demand for indoor navigation, there are great attempts to develop applicable indoor network. Representation for a room as a node is not sufficient to apply complex and large buildings. As OGC established IndoorGML, subspacing to partition the space for constructing logical network is introduced. Concerning subspacing for indoor network, transition space like halls or corridors also have to be considered. This study presents the subspacing process for creating an indoor network in shopping mall. Furthermore, categorization of transition space is performed and subspacing of this space is considered. Hall and squares in mall is especially defined for subspacing. Finally, implementation of subspacing process for indoor network is presented.

Optimizing Cubature for Efficient Integration of Subspace Deformations

PubMed Central

An, Steven S.; Kim, Theodore; James, Doug L.

2009-01-01

We propose an efficient scheme for evaluating nonlinear subspace forces (and Jacobians) associated with subspace deformations. The core problem we address is efficient integration of the subspace force density over the 3D spatial domain. Similar to Gaussian quadrature schemes that efficiently integrate functions that lie in particular polynomial subspaces, we propose cubature schemes (multi-dimensional quadrature) optimized for efficient integration of force densities associated with particular subspace deformations, particular materials, and particular geometric domains. We support generic subspace deformation kinematics, and nonlinear hyperelastic materials. For an r-dimensional deformation subspace with O(r) cubature points, our method is able to evaluate subspace forces at O(r2) cost. We also describe composite cubature rules for runtime error estimation. Results are provided for various subspace deformation models, several hyperelastic materials (St.Venant-Kirchhoff, Mooney-Rivlin, Arruda-Boyce), and multimodal (graphics, haptics, sound) applications. We show dramatically better efficiency than traditional Monte Carlo integration. CR Categories: I.6.8 [Simulation and Modeling]: Types of Simulation—Animation, I.3.5 [Computer Graphics]: Computational Geometry and Object Modeling—Physically based modeling G.1.4 [Mathematics of Computing]: Numerical Analysis—Quadrature and Numerical Differentiation PMID:19956777

Improved neutron-gamma discrimination for a 3He neutron detector using subspace learning methods

DOE PAGES

Wang, C. L.; Funk, L. L.; Riedel, R. A.; ...

2017-02-10

3He gas based neutron linear-position-sensitive detectors (LPSDs) have been applied for many neutron scattering instruments. Traditional Pulse-Height Analysis (PHA) for Neutron-Gamma Discrimination (NGD) resulted in the neutron-gamma efficiency ratio on the orders of 10 5-10 6. The NGD ratios of 3He detectors need to be improved for even better scientific results from neutron scattering. Digital Signal Processing (DSP) analyses of waveforms were proposed for obtaining better NGD ratios, based on features extracted from rise-time, pulse amplitude, charge integration, a simplified Wiener filter, and the cross-correlation between individual and template waveforms of neutron and gamma events. Fisher linear discriminant analysis (FLDA)more » and three multivariate analyses (MVAs) of the features were performed. The NGD ratios are improved by about 10 2-10 3 times compared with the traditional PHA method. Finally, our results indicate the NGD capabilities of 3He tube detectors can be significantly improved with subspace-learning based methods, which may result in a reduced data-collection time and better data quality for further data reduction.« less

Intelligent Control of a Sensor-Actuator System via Kernelized Least-Squares Policy Iteration

PubMed Central

Liu, Bo; Chen, Sanfeng; Li, Shuai; Liang, Yongsheng

2012-01-01

In this paper a new framework, called Compressive Kernelized Reinforcement Learning (CKRL), for computing near-optimal policies in sequential decision making with uncertainty is proposed via incorporating the non-adaptive data-independent Random Projections and nonparametric Kernelized Least-squares Policy Iteration (KLSPI). Random Projections are a fast, non-adaptive dimensionality reduction framework in which high-dimensionality data is projected onto a random lower-dimension subspace via spherically random rotation and coordination sampling. KLSPI introduce kernel trick into the LSPI framework for Reinforcement Learning, often achieving faster convergence and providing automatic feature selection via various kernel sparsification approaches. In this approach, policies are computed in a low-dimensional subspace generated by projecting the high-dimensional features onto a set of random basis. We first show how Random Projections constitute an efficient sparsification technique and how our method often converges faster than regular LSPI, while at lower computational costs. Theoretical foundation underlying this approach is a fast approximation of Singular Value Decomposition (SVD). Finally, simulation results are exhibited on benchmark MDP domains, which confirm gains both in computation time and in performance in large feature spaces. PMID:22736969

Domain adaptation via transfer component analysis.

PubMed

Pan, Sinno Jialin; Tsang, Ivor W; Kwok, James T; Yang, Qiang

2011-02-01

Domain adaptation allows knowledge from a source domain to be transferred to a different but related target domain. Intuitively, discovering a good feature representation across domains is crucial. In this paper, we first propose to find such a representation through a new learning method, transfer component analysis (TCA), for domain adaptation. TCA tries to learn some transfer components across domains in a reproducing kernel Hilbert space using maximum mean miscrepancy. In the subspace spanned by these transfer components, data properties are preserved and data distributions in different domains are close to each other. As a result, with the new representations in this subspace, we can apply standard machine learning methods to train classifiers or regression models in the source domain for use in the target domain. Furthermore, in order to uncover the knowledge hidden in the relations between the data labels from the source and target domains, we extend TCA in a semisupervised learning setting, which encodes label information into transfer components learning. We call this extension semisupervised TCA. The main contribution of our work is that we propose a novel dimensionality reduction framework for reducing the distance between domains in a latent space for domain adaptation. We propose both unsupervised and semisupervised feature extraction approaches, which can dramatically reduce the distance between domain distributions by projecting data onto the learned transfer components. Finally, our approach can handle large datasets and naturally lead to out-of-sample generalization. The effectiveness and efficiency of our approach are verified by experiments on five toy datasets and two real-world applications: cross-domain indoor WiFi localization and cross-domain text classification.

Algorithm 971: An Implementation of a Randomized Algorithm for Principal Component Analysis

PubMed Central

LI, HUAMIN; LINDERMAN, GEORGE C.; SZLAM, ARTHUR; STANTON, KELLY P.; KLUGER, YUVAL; TYGERT, MARK

2017-01-01

Recent years have witnessed intense development of randomized methods for low-rank approximation. These methods target principal component analysis and the calculation of truncated singular value decompositions. The present article presents an essentially black-box, foolproof implementation for Mathworks’ MATLAB, a popular software platform for numerical computation. As illustrated via several tests, the randomized algorithms for low-rank approximation outperform or at least match the classical deterministic techniques (such as Lanczos iterations run to convergence) in basically all respects: accuracy, computational efficiency (both speed and memory usage), ease-of-use, parallelizability, and reliability. However, the classical procedures remain the methods of choice for estimating spectral norms and are far superior for calculating the least singular values and corresponding singular vectors (or singular subspaces). PMID:28983138

begin{center} MUSIC Algorithms for Rebar Detection

NASA Astrophysics Data System (ADS)

Leone, G.; Solimene, R.

2012-04-01

In this contribution we consider the problem of detecting and localizing small cross section, with respect to the wavelength, scatterers from their scattered field once a known incident field interrogated the scene where they reside. A pertinent applicative context is rebar detection within concrete pillar. For such a case, scatterers to be detected are represented by rebars themselves or by voids due to their lacking. In both cases, as scatterers have point-like support, a subspace projection method can be conveniently exploited [1]. However, as the field scattered by rebars is stronger than the one due to voids, it is expected that the latter can be difficult to be detected. In order to circumvent this problem, in this contribution we adopt a two-step MUltiple SIgnal Classification (MUSIC) detection algorithm. In particular, the first stage aims at detecting rebars. Once rebar are detected, their positions are exploited to update the Green's function and then a further detection scheme is run to locate voids. However, in this second case, background medium encompasses also the rabars. The analysis is conducted numerically for a simplified two-dimensional scalar scattering geometry. More in detail, as is usual in MUSIC algorithm, a multi-view/multi-static single-frequency configuration is considered [2]. Baratonia, G. Leone, R. Pierri, R. Solimene, "Fault Detection in Grid Scattering by a Time-Reversal MUSIC Approach," Porc. Of ICEAA 2011, Turin, 2011. E. A. Marengo, F. K. Gruber, "Subspace-Based Localization and Inverse Scattering of Multiply Scattering Point Targets," EURASIP Journal on Advances in Signal Processing, 2007, Article ID 17342, 16 pages (2007).

Eigensolution of finite element problems in a completely connected parallel architecture

NASA Technical Reports Server (NTRS)

Akl, F.; Morel, M.

1989-01-01

A parallel algorithm is presented for the solution of the generalized eigenproblem in linear elastic finite element analysis. The algorithm is based on a completely connected parallel architecture in which each processor is allowed to communicate with all other processors. The algorithm is successfully implemented on a tightly coupled MIMD parallel processor. A finite element model is divided into m domains each of which is assumed to process n elements. Each domain is then assigned to a processor or to a logical processor (task) if the number of domains exceeds the number of physical processors. The effect of the number of domains, the number of degrees-of-freedom located along the global fronts, and the dimension of the subspace on the performance of the algorithm is investigated. For a 64-element rectangular plate, speed-ups of 1.86, 3.13, 3.18, and 3.61 are achieved on two, four, six, and eight processors, respectively.

Joint Facial Action Unit Detection and Feature Fusion: A Multi-conditional Learning Approach.

PubMed

Eleftheriadis, Stefanos; Rudovic, Ognjen; Pantic, Maja

2016-10-05

Automated analysis of facial expressions can benefit many domains, from marketing to clinical diagnosis of neurodevelopmental disorders. Facial expressions are typically encoded as a combination of facial muscle activations, i.e., action units. Depending on context, these action units co-occur in specific patterns, and rarely in isolation. Yet, most existing methods for automatic action unit detection fail to exploit dependencies among them, and the corresponding facial features. To address this, we propose a novel multi-conditional latent variable model for simultaneous fusion of facial features and joint action unit detection. Specifically, the proposed model performs feature fusion in a generative fashion via a low-dimensional shared subspace, while simultaneously performing action unit detection using a discriminative classification approach. We show that by combining the merits of both approaches, the proposed methodology outperforms existing purely discriminative/generative methods for the target task. To reduce the number of parameters, and avoid overfitting, a novel Bayesian learning approach based on Monte Carlo sampling is proposed, to integrate out the shared subspace. We validate the proposed method on posed and spontaneous data from three publicly available datasets (CK+, DISFA and Shoulder-pain), and show that both feature fusion and joint learning of action units leads to improved performance compared to the state-of-the-art methods for the task.

Closed-loop control for cardiopulmonary management and intensive care unit sedation using digital imaging

NASA Astrophysics Data System (ADS)

Gholami, Behnood

This dissertation introduces a new problem in the delivery of healthcare, which could result in lower cost and a higher quality of medical care as compared to the current healthcare practice. In particular, a framework is developed for sedation and cardiopulmonary management for patients in the intensive care unit. A method is introduced to automatically detect pain and agitation in nonverbal patients, specifically in sedated patients in the intensive care unit, using their facial expressions. Furthermore, deterministic as well as probabilistic expert systems are developed to suggest the appropriate drug dose based on patient sedation level. Patients in the intensive care unit who require mechanical ventilation due to acute respiratory failure also frequently require the administration of sedative agents. The need for sedation arises both from patient anxiety due to the loss of personal control and the unfamiliar and intrusive environment of the intensive care unit, and also due to pain or other variants of noxious stimuli. In this dissertation, we develop a rule-based expert system for cardiopulmonary management and intensive care unit sedation. Furthermore, we use probability theory to quantify uncertainty and to extend the proposed rule-based expert system to deal with more realistic situations. Pain assessment in patients who are unable to verbally communicate is a challenging problem. The fundamental limitations in pain assessment stem from subjective assessment criteria, rather than quantifiable, measurable data. The relevance vector machine (RVM) classification technique is a Bayesian extension of the support vector machine (SVM) algorithm which achieves comparable performance to SVM while providing posterior probabilities for class memberships and a sparser model. In this dissertation, we use the RVM classification technique to distinguish pain from non-pain as well as assess pain intensity levels. We also correlate our results with the pain intensity assessed by expert and non-expert human examiners. Next, we consider facial expression recognition using an unsupervised learning framework. We show that different facial expressions reside on distinct subspaces if the manifold is unfolded. In particular, semi-definite embedding is used to reduce the dimensionality and unfold the manifold of facial images. Next, generalized principal component analysis is used to fit a series of subspaces to the data points and associate each data point to a subspace. Data points that belong to the same subspace are shown to belong to the same facial expression. In clinical intensive care unit practice sedative/analgesic agents are titrated to achieve a specific level of sedation. The level of sedation is currently based on clinical scoring systems. Examples include the motor activity assessment scale (MAAS), the Richmond agitation-sedation scale (RASS), and the modified Ramsay sedation scale (MRSS). In general, the goal of the clinician is to find the drug dose that maintains the patient at a sedation score corresponding to a moderately sedated state. In this research, we use pharmacokinetic and pharmacodynamic modeling to find an optimal drug dosing control policy to drive the patient to a desired MRSS score. Atrial fibrillation, a cardiac arrhythmia characterized by unsynchronized electrical activity in the atrial chambers of the heart, is a rapidly growing problem in modern societies. One treatment, referred to as catheter ablation, targets specific parts of the left atrium for radio frequency ablation using an intracardiac catheter. As a first step towards the general solution to the computer-assisted segmentation of the left atrial wall, we use shape learning and shape-based image segmentation to identify the endocardial wall of the left atrium in the delayed-enhancement magnetic resonance images. (Abstract shortened by UMI.)

Decentralized System Identification Using Stochastic Subspace Identification for Wireless Sensor Networks

PubMed Central

Cho, Soojin; Park, Jong-Woong; Sim, Sung-Han

2015-01-01

Wireless sensor networks (WSNs) facilitate a new paradigm to structural identification and monitoring for civil infrastructure. Conventional structural monitoring systems based on wired sensors and centralized data acquisition systems are costly for installation as well as maintenance. WSNs have emerged as a technology that can overcome such difficulties, making deployment of a dense array of sensors on large civil structures both feasible and economical. However, as opposed to wired sensor networks in which centralized data acquisition and processing is common practice, WSNs require decentralized computing algorithms to reduce data transmission due to the limitation associated with wireless communication. In this paper, the stochastic subspace identification (SSI) technique is selected for system identification, and SSI-based decentralized system identification (SDSI) is proposed to be implemented in a WSN composed of Imote2 wireless sensors that measure acceleration. The SDSI is tightly scheduled in the hierarchical WSN, and its performance is experimentally verified in a laboratory test using a 5-story shear building model. PMID:25856325

Robust subspace clustering via joint weighted Schatten-p norm and Lq norm minimization

NASA Astrophysics Data System (ADS)

Zhang, Tao; Tang, Zhenmin; Liu, Qing

2017-05-01

Low-rank representation (LRR) has been successfully applied to subspace clustering. However, the nuclear norm in the standard LRR is not optimal for approximating the rank function in many real-world applications. Meanwhile, the L21 norm in LRR also fails to characterize various noises properly. To address the above issues, we propose an improved LRR method, which achieves low rank property via the new formulation with weighted Schatten-p norm and Lq norm (WSPQ). Specifically, the nuclear norm is generalized to be the Schatten-p norm and different weights are assigned to the singular values, and thus it can approximate the rank function more accurately. In addition, Lq norm is further incorporated into WSPQ to model different noises and improve the robustness. An efficient algorithm based on the inexact augmented Lagrange multiplier method is designed for the formulated problem. Extensive experiments on face clustering and motion segmentation clearly demonstrate the superiority of the proposed WSPQ over several state-of-the-art methods.

Two-Dimensional Hermite Filters Simplify the Description of High-Order Statistics of Natural Images.

PubMed

Hu, Qin; Victor, Jonathan D

2016-09-01

Natural image statistics play a crucial role in shaping biological visual systems, understanding their function and design principles, and designing effective computer-vision algorithms. High-order statistics are critical for conveying local features, but they are challenging to study - largely because their number and variety is large. Here, via the use of two-dimensional Hermite (TDH) functions, we identify a covert symmetry in high-order statistics of natural images that simplifies this task. This emerges from the structure of TDH functions, which are an orthogonal set of functions that are organized into a hierarchy of ranks. Specifically, we find that the shape (skewness and kurtosis) of the distribution of filter coefficients depends only on the projection of the function onto a 1-dimensional subspace specific to each rank. The characterization of natural image statistics provided by TDH filter coefficients reflects both their phase and amplitude structure, and we suggest an intuitive interpretation for the special subspace within each rank.

A weighted information criterion for multiple minor components and its adaptive extraction algorithms.

PubMed

Gao, Yingbin; Kong, Xiangyu; Zhang, Huihui; Hou, Li'an

2017-05-01

Minor component (MC) plays an important role in signal processing and data analysis, so it is a valuable work to develop MC extraction algorithms. Based on the concepts of weighted subspace and optimum theory, a weighted information criterion is proposed for searching the optimum solution of a linear neural network. This information criterion exhibits a unique global minimum attained if and only if the state matrix is composed of the desired MCs of an autocorrelation matrix of an input signal. By using gradient ascent method and recursive least square (RLS) method, two algorithms are developed for multiple MCs extraction. The global convergences of the proposed algorithms are also analyzed by the Lyapunov method. The proposed algorithms can extract the multiple MCs in parallel and has advantage in dealing with high dimension matrices. Since the weighted matrix does not require an accurate value, it facilitates the system design of the proposed algorithms for practical applications. The speed and computation advantages of the proposed algorithms are verified through simulations. Copyright © 2017 Elsevier Ltd. All rights reserved.

Multilevel algorithms for nonlinear optimization

NASA Technical Reports Server (NTRS)

Alexandrov, Natalia; Dennis, J. E., Jr.

1994-01-01

Multidisciplinary design optimization (MDO) gives rise to nonlinear optimization problems characterized by a large number of constraints that naturally occur in blocks. We propose a class of multilevel optimization methods motivated by the structure and number of constraints and by the expense of the derivative computations for MDO. The algorithms are an extension to the nonlinear programming problem of the successful class of local Brown-Brent algorithms for nonlinear equations. Our extensions allow the user to partition constraints into arbitrary blocks to fit the application, and they separately process each block and the objective function, restricted to certain subspaces. The methods use trust regions as a globalization strategy, and they have been shown to be globally convergent under reasonable assumptions. The multilevel algorithms can be applied to all classes of MDO formulations. Multilevel algorithms for solving nonlinear systems of equations are a special case of the multilevel optimization methods. In this case, they can be viewed as a trust-region globalization of the Brown-Brent class.

Student Learning of Basis, Span and Linear Independence in Linear Algebra

ERIC Educational Resources Information Center

Stewart, Sepideh; Thomas, Michael O. J.

2010-01-01

One of the earlier, more challenging concepts in linear algebra at university is that of basis. Students are often taught procedurally how to find a basis for a subspace using matrix manipulation, but may struggle with understanding the construct of basis, making further progress harder. We believe one reason for this is because students have…

Subspace Iteration and Immersed Interface Methods: Theory, Algorithm, and Applications

DTIC Science & Technology

2010-08-20

boundary, Navier-Stokes equations Zhilin Li, Kazufumi Ito North Carolina State University Office of Contract and Grants Leazar Hall Lower Level- MC...ORGANIZATION REPORT NUMBER 19a. NAME OF RESPONSIBLE PERSON 19b. TELEPHONE NUMBER Zhilin Li 919-515-3210 3. DATES COVERED (From - To) 15-Aug-2006...Names of Faculty Supported National Academy MemberPERCENT_SUPPORTEDNAME Zhilin Li 0.40 No Kazufumi Ito 0.40 No 0.80FTE Equivalent: 2Total Number

Preserving Symmetry in Preconditioned Krylov Subspace Methods

NASA Technical Reports Server (NTRS)

Chan, Tony F.; Chow, E.; Saad, Y.; Yeung, M. C.

1996-01-01

We consider the problem of solving a linear system Ax = b when A is nearly symmetric and when the system is preconditioned by a symmetric positive definite matrix M. In the symmetric case, one can recover symmetry by using M-inner products in the conjugate gradient (CG) algorithm. This idea can also be used in the nonsymmetric case, and near symmetry can be preserved similarly. Like CG, the new algorithms are mathematically equivalent to split preconditioning, but do not require M to be factored. Better robustness in a specific sense can also be observed. When combined with truncated versions of iterative methods, tests show that this is more effective than the common practice of forfeiting near-symmetry altogether.

Cyclo-stationary linear parameter time-varying subspace realization method applied for identification of horizontal-axis wind turbines

NASA Astrophysics Data System (ADS)

Velazquez, Antonio; Swartz, R. Andrew

2013-04-01

Wind energy is becoming increasingly important worldwide as an alternative renewable energy source. Economical, maintenance and operation are critical issues for large slender dynamic structures, especially for remote offshore wind farms. Health monitoring systems are very promising instruments to assure reliability and good performance of the structure. These sensing and control technologies are typically informed by models based on mechanics or data-driven identification techniques in the time and/or frequency domain. Frequency response functions are popular but are difficult to realize autonomously for structures of higher order and having overlapping frequency content. Instead, time-domain techniques have shown powerful advantages from a practical point of view (e.g. embedded algorithms in wireless-sensor networks), being more suitable to differentiate closely-related modes. Customarily, time-varying effects are often neglected or dismissed to simplify the analysis, but such is not the case for wind loaded structures with spinning multibodies. A more complex scenario is constituted when dealing with both periodic mechanisms responsible for the vibration shaft of the rotor-blade system, and the wind tower substructure interaction. Transformations of the cyclic effects on the vibration data can be applied to isolate inertia quantities different from rotating-generated forces that are typically non-stationary in nature. After applying these transformations, structural identification can be carried out by stationary techniques via data-correlated Eigensystem realizations. In this paper an exploration of a periodic stationary or cyclo-stationary subspace identification technique is presented here by means of a modified Eigensystem Realization Algorithm (ERA) via Stochastic Subspace Identification (SSI) and Linear Parameter Time-Varying (LPTV) techniques. Structural response is assumed under stationary ambient excitation produced by a Gaussian (white) noise assembled in the operative range bandwidth of horizontal-axis wind turbines. ERA-OKID analysis is driven by correlation-function matrices from the stationary ambient response aiming to reduce noise effects. Singular value decomposition (SVD) and eigenvalue analysis are computed in a last stage to get frequencies and mode shapes. Proposed assumptions are carefully weighted to account for the uncertainty of the environment the wind turbines are subjected to. A numerical example is presented based on data acquisition carried out in a BWC XL.1 low power wind turbine device installed in University of California at Davis. Finally, comments and observations are provided on how this subspace realization technique can be extended for modal-parameter identification using exclusively ambient vibration data.

Output-only cyclo-stationary linear-parameter time-varying stochastic subspace identification method for rotating machinery and spinning structures

NASA Astrophysics Data System (ADS)

Velazquez, Antonio; Swartz, R. Andrew

2015-02-01

Economical maintenance and operation are critical issues for rotating machinery and spinning structures containing blade elements, especially large slender dynamic beams (e.g., wind turbines). Structural health monitoring systems represent promising instruments to assure reliability and good performance from the dynamics of the mechanical systems. However, such devices have not been completely perfected for spinning structures. These sensing technologies are typically informed by both mechanistic models coupled with data-driven identification techniques in the time and/or frequency domain. Frequency response functions are popular but are difficult to realize autonomously for structures of higher order, especially when overlapping frequency content is present. Instead, time-domain techniques have shown to possess powerful advantages from a practical point of view (i.e. low-order computational effort suitable for real-time or embedded algorithms) and also are more suitable to differentiate closely-related modes. Customarily, time-varying effects are often neglected or dismissed to simplify this analysis, but such cannot be the case for sinusoidally loaded structures containing spinning multi-bodies. A more complex scenario is constituted when dealing with both periodic mechanisms responsible for the vibration shaft of the rotor-blade system and the interaction of the supporting substructure. Transformations of the cyclic effects on the vibrational data can be applied to isolate inertial quantities that are different from rotation-generated forces that are typically non-stationary in nature. After applying these transformations, structural identification can be carried out by stationary techniques via data-correlated eigensystem realizations. In this paper, an exploration of a periodic stationary or cyclo-stationary subspace identification technique is presented here for spinning multi-blade systems by means of a modified Eigensystem Realization Algorithm (ERA) via stochastic subspace identification (SSI) and linear parameter time-varying (LPTV) techniques. Structural response is assumed to be stationary ambient excitation produced by a Gaussian (white) noise within the operative range bandwidth of the machinery or structure in study. ERA-OKID analysis is driven by correlation-function matrices from the stationary ambient response aiming to reduce noise effects. Singular value decomposition (SVD) and eigenvalue analysis are computed in a last stage to identify frequencies and complex-valued mode shapes. Proposed assumptions are carefully weighted to account for the uncertainty of the environment. A numerical example is carried out based a spinning finite element (SFE) model, and verified using ANSYS® Ver. 12. Finally, comments and observations are provided on how this subspace realization technique can be extended to the problem of modal-parameter identification using only ambient vibration data.

LRSSLMDA: Laplacian Regularized Sparse Subspace Learning for MiRNA-Disease Association prediction

PubMed Central

Huang, Li

2017-01-01

Predicting novel microRNA (miRNA)-disease associations is clinically significant due to miRNAs’ potential roles of diagnostic biomarkers and therapeutic targets for various human diseases. Previous studies have demonstrated the viability of utilizing different types of biological data to computationally infer new disease-related miRNAs. Yet researchers face the challenge of how to effectively integrate diverse datasets and make reliable predictions. In this study, we presented a computational model named Laplacian Regularized Sparse Subspace Learning for MiRNA-Disease Association prediction (LRSSLMDA), which projected miRNAs/diseases’ statistical feature profile and graph theoretical feature profile to a common subspace. It used Laplacian regularization to preserve the local structures of the training data and a L1-norm constraint to select important miRNA/disease features for prediction. The strength of dimensionality reduction enabled the model to be easily extended to much higher dimensional datasets than those exploited in this study. Experimental results showed that LRSSLMDA outperformed ten previous models: the AUC of 0.9178 in global leave-one-out cross validation (LOOCV) and the AUC of 0.8418 in local LOOCV indicated the model’s superior prediction accuracy; and the average AUC of 0.9181+/-0.0004 in 5-fold cross validation justified its accuracy and stability. In addition, three types of case studies further demonstrated its predictive power. Potential miRNAs related to Colon Neoplasms, Lymphoma, Kidney Neoplasms, Esophageal Neoplasms and Breast Neoplasms were predicted by LRSSLMDA. Respectively, 98%, 88%, 96%, 98% and 98% out of the top 50 predictions were validated by experimental evidences. Therefore, we conclude that LRSSLMDA would be a valuable computational tool for miRNA-disease association prediction. PMID:29253885

Overview of Krylov subspace methods with applications to control problems

NASA Technical Reports Server (NTRS)

Saad, Youcef

1989-01-01

An overview of projection methods based on Krylov subspaces are given with emphasis on their application to solving matrix equations that arise in control problems. The main idea of Krylov subspace methods is to generate a basis of the Krylov subspace Span and seek an approximate solution the the original problem from this subspace. Thus, the original matrix problem of size N is approximated by one of dimension m typically much smaller than N. Krylov subspace methods have been very successful in solving linear systems and eigenvalue problems and are now just becoming popular for solving nonlinear equations. It is shown how they can be used to solve partial pole placement problems, Sylvester's equation, and Lyapunov's equation.

Hyperspectral data acquisition and analysis in imaging and real-time active MIR backscattering spectroscopy

NASA Astrophysics Data System (ADS)

Jarvis, Jan; Haertelt, Marko; Hugger, Stefan; Butschek, Lorenz; Fuchs, Frank; Ostendorf, Ralf; Wagner, Joachim; Beyerer, Juergen

2017-04-01

In this work we present data analysis algorithms for detection of hazardous substances in hyperspectral observations acquired using active mid-infrared (MIR) backscattering spectroscopy. We present a novel background extraction algorithm based on the adaptive target generation process proposed by Ren and Chang called the adaptive background generation process (ABGP) that generates a robust and physically meaningful set of background spectra for operation of the well-known adaptive matched subspace detection (AMSD) algorithm. It is shown that the resulting AMSD-ABGP detection algorithm competes well with other widely used detection algorithms. The method is demonstrated in measurement data obtained by two fundamentally different active MIR hyperspectral data acquisition devices. A hyperspectral image sensor applicable in static scenes takes a wavelength sequential approach to hyperspectral data acquisition, whereas a rapid wavelength-scanning single-element detector variant of the same principle uses spatial scanning to generate the hyperspectral observation. It is shown that the measurement timescale of the latter is sufficient for the application of the data analysis algorithms even in dynamic scenarios.

Individualized statistical learning from medical image databases: application to identification of brain lesions.

PubMed

Erus, Guray; Zacharaki, Evangelia I; Davatzikos, Christos

2014-04-01

This paper presents a method for capturing statistical variation of normal imaging phenotypes, with emphasis on brain structure. The method aims to estimate the statistical variation of a normative set of images from healthy individuals, and identify abnormalities as deviations from normality. A direct estimation of the statistical variation of the entire volumetric image is challenged by the high-dimensionality of images relative to smaller sample sizes. To overcome this limitation, we iteratively sample a large number of lower dimensional subspaces that capture image characteristics ranging from fine and localized to coarser and more global. Within each subspace, a "target-specific" feature selection strategy is applied to further reduce the dimensionality, by considering only imaging characteristics present in a test subject's images. Marginal probability density functions of selected features are estimated through PCA models, in conjunction with an "estimability" criterion that limits the dimensionality of estimated probability densities according to available sample size and underlying anatomy variation. A test sample is iteratively projected to the subspaces of these marginals as determined by PCA models, and its trajectory delineates potential abnormalities. The method is applied to segmentation of various brain lesion types, and to simulated data on which superiority of the iterative method over straight PCA is demonstrated. Copyright © 2014 Elsevier B.V. All rights reserved.

Individualized Statistical Learning from Medical Image Databases: Application to Identification of Brain Lesions

PubMed Central

Erus, Guray; Zacharaki, Evangelia I.; Davatzikos, Christos

2014-01-01

This paper presents a method for capturing statistical variation of normal imaging phenotypes, with emphasis on brain structure. The method aims to estimate the statistical variation of a normative set of images from healthy individuals, and identify abnormalities as deviations from normality. A direct estimation of the statistical variation of the entire volumetric image is challenged by the high-dimensionality of images relative to smaller sample sizes. To overcome this limitation, we iteratively sample a large number of lower dimensional subspaces that capture image characteristics ranging from fine and localized to coarser and more global. Within each subspace, a “target-specific” feature selection strategy is applied to further reduce the dimensionality, by considering only imaging characteristics present in a test subject’s images. Marginal probability density functions of selected features are estimated through PCA models, in conjunction with an “estimability” criterion that limits the dimensionality of estimated probability densities according to available sample size and underlying anatomy variation. A test sample is iteratively projected to the subspaces of these marginals as determined by PCA models, and its trajectory delineates potential abnormalities. The method is applied to segmentation of various brain lesion types, and to simulated data on which superiority of the iterative method over straight PCA is demonstrated. PMID:24607564

A Computationally Efficient Parallel Levenberg-Marquardt Algorithm for Large-Scale Big-Data Inversion

NASA Astrophysics Data System (ADS)

Lin, Y.; O'Malley, D.; Vesselinov, V. V.

2015-12-01

Inverse modeling seeks model parameters given a set of observed state variables. However, for many practical problems due to the facts that the observed data sets are often large and model parameters are often numerous, conventional methods for solving the inverse modeling can be computationally expensive. We have developed a new, computationally-efficient Levenberg-Marquardt method for solving large-scale inverse modeling. Levenberg-Marquardt methods require the solution of a dense linear system of equations which can be prohibitively expensive to compute for large-scale inverse problems. Our novel method projects the original large-scale linear problem down to a Krylov subspace, such that the dimensionality of the measurements can be significantly reduced. Furthermore, instead of solving the linear system for every Levenberg-Marquardt damping parameter, we store the Krylov subspace computed when solving the first damping parameter and recycle it for all the following damping parameters. The efficiency of our new inverse modeling algorithm is significantly improved by using these computational techniques. We apply this new inverse modeling method to invert for a random transitivity field. Our algorithm is fast enough to solve for the distributed model parameters (transitivity) at each computational node in the model domain. The inversion is also aided by the use regularization techniques. The algorithm is coded in Julia and implemented in the MADS computational framework (http://mads.lanl.gov). Julia is an advanced high-level scientific programing language that allows for efficient memory management and utilization of high-performance computational resources. By comparing with a Levenberg-Marquardt method using standard linear inversion techniques, our Levenberg-Marquardt method yields speed-up ratio of 15 in a multi-core computational environment and a speed-up ratio of 45 in a single-core computational environment. Therefore, our new inverse modeling method is a powerful tool for large-scale applications.

Acceleration of GPU-based Krylov solvers via data transfer reduction

DOE PAGES

Anzt, Hartwig; Tomov, Stanimire; Luszczek, Piotr; ...

2015-04-08

Krylov subspace iterative solvers are often the method of choice when solving large sparse linear systems. At the same time, hardware accelerators such as graphics processing units continue to offer significant floating point performance gains for matrix and vector computations through easy-to-use libraries of computational kernels. However, as these libraries are usually composed of a well optimized but limited set of linear algebra operations, applications that use them often fail to reduce certain data communications, and hence fail to leverage the full potential of the accelerator. In this study, we target the acceleration of Krylov subspace iterative methods for graphicsmore » processing units, and in particular the Biconjugate Gradient Stabilized solver that significant improvement can be achieved by reformulating the method to reduce data-communications through application-specific kernels instead of using the generic BLAS kernels, e.g. as provided by NVIDIA’s cuBLAS library, and by designing a graphics processing unit specific sparse matrix-vector product kernel that is able to more efficiently use the graphics processing unit’s computing power. Furthermore, we derive a model estimating the performance improvement, and use experimental data to validate the expected runtime savings. Finally, considering that the derived implementation achieves significantly higher performance, we assert that similar optimizations addressing algorithm structure, as well as sparse matrix-vector, are crucial for the subsequent development of high-performance graphics processing units accelerated Krylov subspace iterative methods.« less

Discriminative Transfer Subspace Learning via Low-Rank and Sparse Representation.

PubMed

Xu, Yong; Fang, Xiaozhao; Wu, Jian; Li, Xuelong; Zhang, David

2016-02-01

In this paper, we address the problem of unsupervised domain transfer learning in which no labels are available in the target domain. We use a transformation matrix to transfer both the source and target data to a common subspace, where each target sample can be represented by a combination of source samples such that the samples from different domains can be well interlaced. In this way, the discrepancy of the source and target domains is reduced. By imposing joint low-rank and sparse constraints on the reconstruction coefficient matrix, the global and local structures of data can be preserved. To enlarge the margins between different classes as much as possible and provide more freedom to diminish the discrepancy, a flexible linear classifier (projection) is obtained by learning a non-negative label relaxation matrix that allows the strict binary label matrix to relax into a slack variable matrix. Our method can avoid a potentially negative transfer by using a sparse matrix to model the noise and, thus, is more robust to different types of noise. We formulate our problem as a constrained low-rankness and sparsity minimization problem and solve it by the inexact augmented Lagrange multiplier method. Extensive experiments on various visual domain adaptation tasks show the superiority of the proposed method over the state-of-the art methods. The MATLAB code of our method will be publicly available at http://www.yongxu.org/lunwen.html.

Variational second order density matrix study of F3-: importance of subspace constraints for size-consistency.

PubMed

van Aggelen, Helen; Verstichel, Brecht; Bultinck, Patrick; Van Neck, Dimitri; Ayers, Paul W; Cooper, David L

2011-02-07

Variational second order density matrix theory under "two-positivity" constraints tends to dissociate molecules into unphysical fractionally charged products with too low energies. We aim to construct a qualitatively correct potential energy surface for F(3)(-) by applying subspace energy constraints on mono- and diatomic subspaces of the molecular basis space. Monoatomic subspace constraints do not guarantee correct dissociation: the constraints are thus geometry dependent. Furthermore, the number of subspace constraints needed for correct dissociation does not grow linearly with the number of atoms. The subspace constraints do impose correct chemical properties in the dissociation limit and size-consistency, but the structure of the resulting second order density matrix method does not exactly correspond to a system of noninteracting units.

A non-contact method based on multiple signal classification algorithm to reduce the measurement time for accurately heart rate detection

NASA Astrophysics Data System (ADS)

Bechet, P.; Mitran, R.; Munteanu, M.

2013-08-01

Non-contact methods for the assessment of vital signs are of great interest for specialists due to the benefits obtained in both medical and special applications, such as those for surveillance, monitoring, and search and rescue. This paper investigates the possibility of implementing a digital processing algorithm based on the MUSIC (Multiple Signal Classification) parametric spectral estimation in order to reduce the observation time needed to accurately measure the heart rate. It demonstrates that, by proper dimensioning the signal subspace, the MUSIC algorithm can be optimized in order to accurately assess the heart rate during an 8-28 s time interval. The validation of the processing algorithm performance was achieved by minimizing the mean error of the heart rate after performing simultaneous comparative measurements on several subjects. In order to calculate the error the reference value of heart rate was measured using a classic measurement system through direct contact.

Perfect blind restoration of images blurred by multiple filters: theory and efficient algorithms.

PubMed

Harikumar, G; Bresler, Y

1999-01-01

We address the problem of restoring an image from its noisy convolutions with two or more unknown finite impulse response (FIR) filters. We develop theoretical results about the existence and uniqueness of solutions, and show that under some generically true assumptions, both the filters and the image can be determined exactly in the absence of noise, and stably estimated in its presence. We present efficient algorithms to estimate the blur functions and their sizes. These algorithms are of two types, subspace-based and likelihood-based, and are extensions of techniques proposed for the solution of the multichannel blind deconvolution problem in one dimension. We present memory and computation-efficient techniques to handle the very large matrices arising in the two-dimensional (2-D) case. Once the blur functions are determined, they are used in a multichannel deconvolution step to reconstruct the unknown image. The theoretical and practical implications of edge effects, and "weakly exciting" images are examined. Finally, the algorithms are demonstrated on synthetic and real data.

Implementation of the block-Krylov boundary flexibility method of component synthesis

NASA Technical Reports Server (NTRS)

Carney, Kelly S.; Abdallah, Ayman A.; Hucklebridge, Arthur A.

1993-01-01

A method of dynamic substructuring is presented which utilizes a set of static Ritz vectors as a replacement for normal eigenvectors in component mode synthesis. This set of Ritz vectors is generated in a recurrence relationship, which has the form of a block-Krylov subspace. The initial seed to the recurrence algorithm is based on the boundary flexibility vectors of the component. This algorithm is not load-dependent, is applicable to both fixed and free-interface boundary components, and results in a general component model appropriate for any type of dynamic analysis. This methodology was implemented in the MSC/NASTRAN normal modes solution sequence using DMAP. The accuracy is found to be comparable to that of component synthesis based upon normal modes. The block-Krylov recurrence algorithm is a series of static solutions and so requires significantly less computation than solving the normal eigenspace problem.

Flag-based detection of weak gas signatures in long-wave infrared hyperspectral image sequences

NASA Astrophysics Data System (ADS)

Marrinan, Timothy; Beveridge, J. Ross; Draper, Bruce; Kirby, Michael; Peterson, Chris

2016-05-01

We present a flag manifold based method for detecting chemical plumes in long-wave infrared hyperspectral movies. The method encodes temporal and spatial information related to a hyperspectral pixel into a flag, or nested sequence of linear subspaces. The technique used to create the flags pushes information about the background clutter, ambient conditions, and potential chemical agents into the leading elements of the flags. Exploiting this temporal information allows for a detection algorithm that is sensitive to the presence of weak signals. This method is compared to existing techniques qualitatively on real data and quantitatively on synthetic data to show that the flag-based algorithm consistently performs better on data when the SINRdB is low, and beats the ACE and MF algorithms in probability of detection for low probabilities of false alarm even when the SINRdB is high.

SIAM Conference on Parallel Processing for Scientific Computing, 4th, Chicago, IL, Dec. 11-13, 1989, Proceedings

NASA Technical Reports Server (NTRS)

Dongarra, Jack (Editor); Messina, Paul (Editor); Sorensen, Danny C. (Editor); Voigt, Robert G. (Editor)

1990-01-01

Attention is given to such topics as an evaluation of block algorithm variants in LAPACK and presents a large-grain parallel sparse system solver, a multiprocessor method for the solution of the generalized Eigenvalue problem on an interval, and a parallel QR algorithm for iterative subspace methods on the CM2. A discussion of numerical methods includes the topics of asynchronous numerical solutions of PDEs on parallel computers, parallel homotopy curve tracking on a hypercube, and solving Navier-Stokes equations on the Cedar Multi-Cluster system. A section on differential equations includes a discussion of a six-color procedure for the parallel solution of elliptic systems using the finite quadtree structure, data parallel algorithms for the finite element method, and domain decomposition methods in aerodynamics. Topics dealing with massively parallel computing include hypercube vs. 2-dimensional meshes and massively parallel computation of conservation laws. Performance and tools are also discussed.

Conformational states and folding pathways of peptides revealed by principal-independent component analyses.

PubMed

Nguyen, Phuong H

2007-05-15

Principal component analysis is a powerful method for projecting multidimensional conformational space of peptides or proteins onto lower dimensional subspaces in which the main conformations are present, making it easier to reveal the structures of molecules from e.g. molecular dynamics simulation trajectories. However, the identification of all conformational states is still difficult if the subspaces consist of more than two dimensions. This is mainly due to the fact that the principal components are not independent with each other, and states in the subspaces cannot be visualized. In this work, we propose a simple and fast scheme that allows one to obtain all conformational states in the subspaces. The basic idea is that instead of directly identifying the states in the subspace spanned by principal components, we first transform this subspace into another subspace formed by components that are independent of one other. These independent components are obtained from the principal components by employing the independent component analysis method. Because of independence between components, all states in this new subspace are defined as all possible combinations of the states obtained from each single independent component. This makes the conformational analysis much simpler. We test the performance of the method by analyzing the conformations of the glycine tripeptide and the alanine hexapeptide. The analyses show that our method is simple and quickly reveal all conformational states in the subspaces. The folding pathways between the identified states of the alanine hexapeptide are analyzed and discussed in some detail. 2007 Wiley-Liss, Inc.

A Variational Approach to Video Registration with Subspace Constraints.

PubMed

Garg, Ravi; Roussos, Anastasios; Agapito, Lourdes

2013-01-01

This paper addresses the problem of non-rigid video registration, or the computation of optical flow from a reference frame to each of the subsequent images in a sequence, when the camera views deformable objects. We exploit the high correlation between 2D trajectories of different points on the same non-rigid surface by assuming that the displacement of any point throughout the sequence can be expressed in a compact way as a linear combination of a low-rank motion basis. This subspace constraint effectively acts as a trajectory regularization term leading to temporally consistent optical flow. We formulate it as a robust soft constraint within a variational framework by penalizing flow fields that lie outside the low-rank manifold. The resulting energy functional can be decoupled into the optimization of the brightness constancy and spatial regularization terms, leading to an efficient optimization scheme. Additionally, we propose a novel optimization scheme for the case of vector valued images, based on the dualization of the data term. This allows us to extend our approach to deal with colour images which results in significant improvements on the registration results. Finally, we provide a new benchmark dataset, based on motion capture data of a flag waving in the wind, with dense ground truth optical flow for evaluation of multi-frame optical flow algorithms for non-rigid surfaces. Our experiments show that our proposed approach outperforms state of the art optical flow and dense non-rigid registration algorithms.

Accelerating molecular property calculations with nonorthonormal Krylov space methods

DOE Office of Scientific and Technical Information (OSTI.GOV)

Furche, Filipp; Krull, Brandon T.; Nguyen, Brian D.

Here, we formulate Krylov space methods for large eigenvalue problems and linear equation systems that take advantage of decreasing residual norms to reduce the cost of matrix-vector multiplication. The residuals are used as subspace basis without prior orthonormalization, which leads to generalized eigenvalue problems or linear equation systems on the Krylov space. These nonorthonormal Krylov space (nKs) algorithms are favorable for large matrices with irregular sparsity patterns whose elements are computed on the fly, because fewer operations are necessary as the residual norm decreases as compared to the conventional method, while errors in the desired eigenpairs and solution vectors remainmore » small. We consider real symmetric and symplectic eigenvalue problems as well as linear equation systems and Sylvester equations as they appear in configuration interaction and response theory. The nKs method can be implemented in existing electronic structure codes with minor modifications and yields speed-ups of 1.2-1.8 in typical time-dependent Hartree-Fock and density functional applications without accuracy loss. The algorithm can compute entire linear subspaces simultaneously which benefits electronic spectra and force constant calculations requiring many eigenpairs or solution vectors. The nKs approach is related to difference density methods in electronic ground state calculations, and particularly efficient for integral direct computations of exchange-type contractions. By combination with resolution-of-the-identity methods for Coulomb contractions, three- to fivefold speed-ups of hybrid time-dependent density functional excited state and response calculations are achieved.« less

Accelerating molecular property calculations with nonorthonormal Krylov space methods

DOE PAGES

Furche, Filipp; Krull, Brandon T.; Nguyen, Brian D.; ...

2016-05-03

Here, we formulate Krylov space methods for large eigenvalue problems and linear equation systems that take advantage of decreasing residual norms to reduce the cost of matrix-vector multiplication. The residuals are used as subspace basis without prior orthonormalization, which leads to generalized eigenvalue problems or linear equation systems on the Krylov space. These nonorthonormal Krylov space (nKs) algorithms are favorable for large matrices with irregular sparsity patterns whose elements are computed on the fly, because fewer operations are necessary as the residual norm decreases as compared to the conventional method, while errors in the desired eigenpairs and solution vectors remainmore » small. We consider real symmetric and symplectic eigenvalue problems as well as linear equation systems and Sylvester equations as they appear in configuration interaction and response theory. The nKs method can be implemented in existing electronic structure codes with minor modifications and yields speed-ups of 1.2-1.8 in typical time-dependent Hartree-Fock and density functional applications without accuracy loss. The algorithm can compute entire linear subspaces simultaneously which benefits electronic spectra and force constant calculations requiring many eigenpairs or solution vectors. The nKs approach is related to difference density methods in electronic ground state calculations, and particularly efficient for integral direct computations of exchange-type contractions. By combination with resolution-of-the-identity methods for Coulomb contractions, three- to fivefold speed-ups of hybrid time-dependent density functional excited state and response calculations are achieved.« less

Nonadiabatic holonomic quantum computation in decoherence-free subspaces.

PubMed

Xu, G F; Zhang, J; Tong, D M; Sjöqvist, Erik; Kwek, L C

2012-10-26

Quantum computation that combines the coherence stabilization virtues of decoherence-free subspaces and the fault tolerance of geometric holonomic control is of great practical importance. Some schemes of adiabatic holonomic quantum computation in decoherence-free subspaces have been proposed in the past few years. However, nonadiabatic holonomic quantum computation in decoherence-free subspaces, which avoids a long run-time requirement but with all the robust advantages, remains an open problem. Here, we demonstrate how to realize nonadiabatic holonomic quantum computation in decoherence-free subspaces. By using only three neighboring physical qubits undergoing collective dephasing to encode one logical qubit, we realize a universal set of quantum gates.

Blended particle filters for large-dimensional chaotic dynamical systems

PubMed Central

Majda, Andrew J.; Qi, Di; Sapsis, Themistoklis P.

2014-01-01

A major challenge in contemporary data science is the development of statistically accurate particle filters to capture non-Gaussian features in large-dimensional chaotic dynamical systems. Blended particle filters that capture non-Gaussian features in an adaptively evolving low-dimensional subspace through particles interacting with evolving Gaussian statistics on the remaining portion of phase space are introduced here. These blended particle filters are constructed in this paper through a mathematical formalism involving conditional Gaussian mixtures combined with statistically nonlinear forecast models compatible with this structure developed recently with high skill for uncertainty quantification. Stringent test cases for filtering involving the 40-dimensional Lorenz 96 model with a 5-dimensional adaptive subspace for nonlinear blended filtering in various turbulent regimes with at least nine positive Lyapunov exponents are used here. These cases demonstrate the high skill of the blended particle filter algorithms in capturing both highly non-Gaussian dynamical features as well as crucial nonlinear statistics for accurate filtering in extreme filtering regimes with sparse infrequent high-quality observations. The formalism developed here is also useful for multiscale filtering of turbulent systems and a simple application is sketched below. PMID:24825886

Discriminant analysis of resting-state functional connectivity patterns on the Grassmann manifold

NASA Astrophysics Data System (ADS)

Fan, Yong; Liu, Yong; Jiang, Tianzi; Liu, Zhening; Hao, Yihui; Liu, Haihong

2010-03-01

The functional networks, extracted from fMRI images using independent component analysis, have been demonstrated informative for distinguishing brain states of cognitive functions and neurological diseases. In this paper, we propose a novel algorithm for discriminant analysis of functional networks encoded by spatial independent components. The functional networks of each individual are used as bases for a linear subspace, referred to as a functional connectivity pattern, which facilitates a comprehensive characterization of temporal signals of fMRI data. The functional connectivity patterns of different individuals are analyzed on the Grassmann manifold by adopting a principal angle based subspace distance. In conjunction with a support vector machine classifier, a forward component selection technique is proposed to select independent components for constructing the most discriminative functional connectivity pattern. The discriminant analysis method has been applied to an fMRI based schizophrenia study with 31 schizophrenia patients and 31 healthy individuals. The experimental results demonstrate that the proposed method not only achieves a promising classification performance for distinguishing schizophrenia patients from healthy controls, but also identifies discriminative functional networks that are informative for schizophrenia diagnosis.

Optimal design of focused experiments and surveys

NASA Astrophysics Data System (ADS)

Curtis, Andrew

1999-10-01

Experiments and surveys are often performed to obtain data that constrain some previously underconstrained model. Often, constraints are most desired in a particular subspace of model space. Experiment design optimization requires that the quality of any particular design can be both quantified and then maximized. This study shows how the quality can be defined such that it depends on the amount of information that is focused in the particular subspace of interest. In addition, algorithms are presented which allow one particular focused quality measure (from the class of focused measures) to be evaluated efficiently. A subclass of focused quality measures is also related to the standard variance and resolution measures from linearized inverse theory. The theory presented here requires that the relationship between model parameters and data can be linearized around a reference model without significant loss of information. Physical and financial constraints define the space of possible experiment designs. Cross-well tomographic examples are presented, plus a strategy for survey design to maximize information about linear combinations of parameters such as bulk modulus, κ =λ+ 2μ/3.

DOA Estimation for Underwater Wideband Weak Targets Based on Coherent Signal Subspace and Compressed Sensing.

PubMed

Li, Jun; Lin, Qiu-Hua; Kang, Chun-Yu; Wang, Kai; Yang, Xiu-Ting

2018-03-18

Direction of arrival (DOA) estimation is the basis for underwater target localization and tracking using towed line array sonar devices. A method of DOA estimation for underwater wideband weak targets based on coherent signal subspace (CSS) processing and compressed sensing (CS) theory is proposed. Under the CSS processing framework, wideband frequency focusing is accompanied by a two-sided correlation transformation, allowing the DOA of underwater wideband targets to be estimated based on the spatial sparsity of the targets and the compressed sensing reconstruction algorithm. Through analysis and processing of simulation data and marine trial data, it is shown that this method can accomplish the DOA estimation of underwater wideband weak targets. Results also show that this method can considerably improve the spatial spectrum of weak target signals, enhancing the ability to detect them. It can solve the problems of low directional resolution and unreliable weak-target detection in traditional beamforming technology. Compared with the conventional minimum variance distortionless response beamformers (MVDR), this method has many advantages, such as higher directional resolution, wider detection range, fewer required snapshots and more accurate detection for weak targets.

Krylov-Subspace Recycling via the POD-Augmented Conjugate-Gradient Method

DOE Office of Scientific and Technical Information (OSTI.GOV)

Carlberg, Kevin; Forstall, Virginia; Tuminaro, Ray

This paper presents a new Krylov-subspace-recycling method for efficiently solving sequences of linear systems of equations characterized by varying right-hand sides and symmetric-positive-definite matrices. As opposed to typical truncation strategies used in recycling such as deflation, we propose a truncation method inspired by goal-oriented proper orthogonal decomposition (POD) from model reduction. This idea is based on the observation that model reduction aims to compute a low-dimensional subspace that contains an accurate solution; as such, we expect the proposed method to generate a low-dimensional subspace that is well suited for computing solutions that can satisfy inexact tolerances. In particular, we proposemore » specific goal-oriented POD `ingredients' that align the optimality properties of POD with the objective of Krylov-subspace recycling. To compute solutions in the resulting 'augmented' POD subspace, we propose a hybrid direct/iterative three-stage method that leverages 1) the optimal ordering of POD basis vectors, and 2) well-conditioned reduced matrices. Numerical experiments performed on solid-mechanics problems highlight the benefits of the proposed method over existing approaches for Krylov-subspace recycling.« less

Krylov-Subspace Recycling via the POD-Augmented Conjugate-Gradient Method

DOE PAGES

Carlberg, Kevin; Forstall, Virginia; Tuminaro, Ray

2016-01-01

This paper presents a new Krylov-subspace-recycling method for efficiently solving sequences of linear systems of equations characterized by varying right-hand sides and symmetric-positive-definite matrices. As opposed to typical truncation strategies used in recycling such as deflation, we propose a truncation method inspired by goal-oriented proper orthogonal decomposition (POD) from model reduction. This idea is based on the observation that model reduction aims to compute a low-dimensional subspace that contains an accurate solution; as such, we expect the proposed method to generate a low-dimensional subspace that is well suited for computing solutions that can satisfy inexact tolerances. In particular, we proposemore » specific goal-oriented POD `ingredients' that align the optimality properties of POD with the objective of Krylov-subspace recycling. To compute solutions in the resulting 'augmented' POD subspace, we propose a hybrid direct/iterative three-stage method that leverages 1) the optimal ordering of POD basis vectors, and 2) well-conditioned reduced matrices. Numerical experiments performed on solid-mechanics problems highlight the benefits of the proposed method over existing approaches for Krylov-subspace recycling.« less

Hypercyclic subspaces for Frechet space operators

NASA Astrophysics Data System (ADS)

Petersson, Henrik

2006-07-01

A continuous linear operator is hypercyclic if there is an such that the orbit {Tnx} is dense, and such a vector x is said to be hypercyclic for T. Recent progress show that it is possible to characterize Banach space operators that have a hypercyclic subspace, i.e., an infinite dimensional closed subspace of, except for zero, hypercyclic vectors. The following is known to hold: A Banach space operator T has a hypercyclic subspace if there is a sequence (ni) and an infinite dimensional closed subspace such that T is hereditarily hypercyclic for (ni) and Tni->0 pointwise on E. In this note we extend this result to the setting of Frechet spaces that admit a continuous norm, and study some applications for important function spaces. As an application we also prove that any infinite dimensional separable Frechet space with a continuous norm admits an operator with a hypercyclic subspace.

An efficient linear-scaling CCSD(T) method based on local natural orbitals.

PubMed

Rolik, Zoltán; Szegedy, Lóránt; Ladjánszki, István; Ladóczki, Bence; Kállay, Mihály

2013-09-07

An improved version of our general-order local coupled-cluster (CC) approach [Z. Rolik and M. Kállay, J. Chem. Phys. 135, 104111 (2011)] and its efficient implementation at the CC singles and doubles with perturbative triples [CCSD(T)] level is presented. The method combines the cluster-in-molecule approach of Li and co-workers [J. Chem. Phys. 131, 114109 (2009)] with frozen natural orbital (NO) techniques. To break down the unfavorable fifth-power scaling of our original approach a two-level domain construction algorithm has been developed. First, an extended domain of localized molecular orbitals (LMOs) is assembled based on the spatial distance of the orbitals. The necessary integrals are evaluated and transformed in these domains invoking the density fitting approximation. In the second step, for each occupied LMO of the extended domain a local subspace of occupied and virtual orbitals is constructed including approximate second-order Mo̸ller-Plesset NOs. The CC equations are solved and the perturbative corrections are calculated in the local subspace for each occupied LMO using a highly-efficient CCSD(T) code, which was optimized for the typical sizes of the local subspaces. The total correlation energy is evaluated as the sum of the individual contributions. The computation time of our approach scales linearly with the system size, while its memory and disk space requirements are independent thereof. Test calculations demonstrate that currently our method is one of the most efficient local CCSD(T) approaches and can be routinely applied to molecules of up to 100 atoms with reasonable basis sets.

Human Guidance Behavior Decomposition and Modeling

NASA Astrophysics Data System (ADS)

Feit, Andrew James

Trained humans are capable of high performance, adaptable, and robust first-person dynamic motion guidance behavior. This behavior is exhibited in a wide variety of activities such as driving, piloting aircraft, skiing, biking, and many others. Human performance in such activities far exceeds the current capability of autonomous systems in terms of adaptability to new tasks, real-time motion planning, robustness, and trading safety for performance. The present work investigates the structure of human dynamic motion guidance that enables these performance qualities. This work uses a first-person experimental framework that presents a driving task to the subject, measuring control inputs, vehicle motion, and operator visual gaze movement. The resulting data is decomposed into subspace segment clusters that form primitive elements of action-perception interactive behavior. Subspace clusters are defined by both agent-environment system dynamic constraints and operator control strategies. A key contribution of this work is to define transitions between subspace cluster segments, or subgoals, as points where the set of active constraints, either system or operator defined, changes. This definition provides necessary conditions to determine transition points for a given task-environment scenario that allow a solution trajectory to be planned from known behavior elements. In addition, human gaze behavior during this task contains predictive behavior elements, indicating that the identified control modes are internally modeled. Based on these ideas, a generative, autonomous guidance framework is introduced that efficiently generates optimal dynamic motion behavior in new tasks. The new subgoal planning algorithm is shown to generate solutions to certain tasks more quickly than existing approaches currently used in robotics.

Assessment of beer quality based on foamability and chemical composition using computer vision algorithms, near infrared spectroscopy and machine learning algorithms.

PubMed

Gonzalez Viejo, Claudia; Fuentes, Sigfredo; Torrico, Damir; Howell, Kate; Dunshea, Frank R

2018-01-01

Beer quality is mainly defined by its colour, foamability and foam stability, which are influenced by the chemical composition of the product such as proteins, carbohydrates, pH and alcohol. Traditional methods to assess specific chemical compounds are usually time-consuming and costly. This study used rapid methods to evaluate 15 foam and colour-related parameters using a robotic pourer (RoboBEER) and chemical fingerprinting using near infrared spectroscopy (NIR) from six replicates of 21 beers from three types of fermentation. Results from NIR were used to create partial least squares regression (PLS) and artificial neural networks (ANN) models to predict four chemometrics such as pH, alcohol, Brix and maximum volume of foam. The ANN method was able to create more accurate models (R 2  = 0.95) compared to PLS. Principal components analysis using RoboBEER parameters and NIR overtones related to protein explained 67% of total data variability. Additionally, a sub-space discriminant model using the absorbance values from NIR wavelengths resulted in the successful classification of 85% of beers according to fermentation type. The method proposed showed to be a rapid system based on NIR spectroscopy and RoboBEER outputs of foamability that can be used to infer the quality, production method and chemical parameters of beer with minimal laboratory equipment. © 2017 Society of Chemical Industry. © 2017 Society of Chemical Industry.

Robust uncertainty evaluation for system identification on distributed wireless platforms

NASA Astrophysics Data System (ADS)

Crinière, Antoine; Döhler, Michael; Le Cam, Vincent; Mevel, Laurent

2016-04-01

Health monitoring of civil structures by system identification procedures from automatic control is now accepted as a valid approach. These methods provide frequencies and modeshapes from the structure over time. For a continuous monitoring the excitation of a structure is usually ambient, thus unknown and assumed to be noise. Hence, all estimates from the vibration measurements are realizations of random variables with inherent uncertainty due to (unknown) process and measurement noise and finite data length. The underlying algorithms are usually running under Matlab under the assumption of large memory pool and considerable computational power. Even under these premises, computational and memory usage are heavy and not realistic for being embedded in on-site sensor platforms such as the PEGASE platform. Moreover, the current push for distributed wireless systems calls for algorithmic adaptation for lowering data exchanges and maximizing local processing. Finally, the recent breakthrough in system identification allows us to process both frequency information and its related uncertainty together from one and only one data sequence, at the expense of computational and memory explosion that require even more careful attention than before. The current approach will focus on presenting a system identification procedure called multi-setup subspace identification that allows to process both frequencies and their related variances from a set of interconnected wireless systems with all computation running locally within the limited memory pool of each system before being merged on a host supervisor. Careful attention will be given to data exchanges and I/O satisfying OGC standards, as well as minimizing memory footprints and maximizing computational efficiency. Those systems are built in a way of autonomous operations on field and could be later included in a wide distributed architecture such as the Cloud2SM project. The usefulness of these strategies is illustrated on data from a progressive damage action on a prestressed concrete bridge. References [1] E. Carden and P. Fanning. Vibration based condition monitoring: a review. Structural Health Monitoring, 3(4):355-377, 2004. [2] M. Döhler and L. Mevel. Efficient multi-order uncertainty computation for stochastic subspace identification. Mechanical Systems and Signal Processing, 38(2):346-366, 2013. [3] M.Döhler, L. Mevel. Modular subspace-based system identification from multi-setup measurements. IEEE Transactions on Automatic Control, 57(11):2951-2956, 2012. [4] M. Döhler, X.-B. Lam, and L. Mevel. Uncertainty quantification for modal parameters from stochastic subspace identification on multi-setup measurements. MechanicalSystems and Signal Processing, 36(2):562-581, 2013. [5] A Crinière, J Dumoulin, L Mevel, G Andrade-Barosso, M Simonin. The Cloud2SM Project.European Geosciences Union General Assembly (EGU2015), Apr 2015, Vienne, Austria. 2015.

Expert-guided evolutionary algorithm for layout design of complex space stations

NASA Astrophysics Data System (ADS)

Qian, Zhiqin; Bi, Zhuming; Cao, Qun; Ju, Weiguo; Teng, Hongfei; Zheng, Yang; Zheng, Siyu

2017-08-01

The layout of a space station should be designed in such a way that different equipment and instruments are placed for the station as a whole to achieve the best overall performance. The station layout design is a typical nondeterministic polynomial problem. In particular, how to manage the design complexity to achieve an acceptable solution within a reasonable timeframe poses a great challenge. In this article, a new evolutionary algorithm has been proposed to meet such a challenge. It is called as the expert-guided evolutionary algorithm with a tree-like structure decomposition (EGEA-TSD). Two innovations in EGEA-TSD are (i) to deal with the design complexity, the entire design space is divided into subspaces with a tree-like structure; it reduces the computation and facilitates experts' involvement in the solving process. (ii) A human-intervention interface is developed to allow experts' involvement in avoiding local optimums and accelerating convergence. To validate the proposed algorithm, the layout design of one-space station is formulated as a multi-disciplinary design problem, the developed algorithm is programmed and executed, and the result is compared with those from other two algorithms; it has illustrated the superior performance of the proposed EGEA-TSD.

Restoration of multichannel microwave radiometric images

NASA Technical Reports Server (NTRS)

Chin, R. T.; Yeh, C. L.; Olson, W. S.

1983-01-01

A constrained iterative image restoration method is applied to multichannel diffraction-limited imagery. This method is based on the Gerchberg-Papoulis algorithm utilizing incomplete information and partial constraints. The procedure is described using the orthogonal projection operators which project onto two prescribed subspaces iteratively. Some of its properties and limitations are also presented. The selection of appropriate constraints was emphasized in a practical application. Multichannel microwave images, each having different spatial resolution, were restored to a common highest resolution to demonstrate the effectiveness of the method. Both noise-free and noisy images were used in this investigation.

Object recognition of real targets using modelled SAR images

NASA Astrophysics Data System (ADS)

Zherdev, D. A.

2017-12-01

In this work the problem of recognition is studied using SAR images. The algorithm of recognition is based on the computation of conjugation indices with vectors of class. The support subspaces for each class are constructed by exception of the most and the less correlated vectors in a class. In the study we examine the ability of a significant feature vector size reduce that leads to recognition time decrease. The images of targets form the feature vectors that are transformed using pre-trained convolutional neural network (CNN).

Spatial prediction of landslides using a hybrid machine learning approach based on Random Subspace and Classification and Regression Trees

NASA Astrophysics Data System (ADS)

Pham, Binh Thai; Prakash, Indra; Tien Bui, Dieu

2018-02-01

A hybrid machine learning approach of Random Subspace (RSS) and Classification And Regression Trees (CART) is proposed to develop a model named RSSCART for spatial prediction of landslides. This model is a combination of the RSS method which is known as an efficient ensemble technique and the CART which is a state of the art classifier. The Luc Yen district of Yen Bai province, a prominent landslide prone area of Viet Nam, was selected for the model development. Performance of the RSSCART model was evaluated through the Receiver Operating Characteristic (ROC) curve, statistical analysis methods, and the Chi Square test. Results were compared with other benchmark landslide models namely Support Vector Machines (SVM), single CART, Naïve Bayes Trees (NBT), and Logistic Regression (LR). In the development of model, ten important landslide affecting factors related with geomorphology, geology and geo-environment were considered namely slope angles, elevation, slope aspect, curvature, lithology, distance to faults, distance to rivers, distance to roads, and rainfall. Performance of the RSSCART model (AUC = 0.841) is the best compared with other popular landslide models namely SVM (0.835), single CART (0.822), NBT (0.821), and LR (0.723). These results indicate that performance of the RSSCART is a promising method for spatial landslide prediction.

Machine-Learning Inspired Seismic Phase Detection for Aftershocks of the 2008 MW7.9 Wenchuan Earthquake

NASA Astrophysics Data System (ADS)

Zhu, L.; Li, Z.; Li, C.; Wang, B.; Chen, Z.; McClellan, J. H.; Peng, Z.

2017-12-01

Spatial-temporal evolution of aftershocks is important for illumination of earthquake physics and for rapid response of devastative earthquakes. To improve aftershock catalogs of the 2008 MW7.9 Wenchuan earthquake in Sichuan, China, Alibaba cloud and China Earthquake Administration jointly launched a seismological contest in May 2017 [Fang et al., 2017]. This abstract describes how we handle this problem in this competition. We first used Short-Term Average/Long-Term Average (STA/LTA) and Kurtosis function to obtain over 55000 candidate phase picks (P or S). Based on Signal to Noise Ratio (SNR), about 40000 phases (P or S) are selected. So far, these 40000 phases have a hit rate of 40% among the manually picks. The causes include that 1) there exist false picks (neither P nor S); 2) some P and S arrivals are mis-labeled. To improve our results, we correlate the 40000 phases over continuous waveforms to obtain the phases missed by during the first pass. This results in 120,000 events. After constructing an affinity matrix based on the cross-correlation for newly detected phases, subspace clustering methods [Vidal 2011] are applied to group those phases into separated subspaces. Initial results show good agreement between empirical and clustered labels of P phases. Half of the empirical S phases are clustered into the P phase cluster. This may be a combined effect of 1) mislabeling isolated P phases to S phases and 2) clustering errors due to a small incomplete sample pool. Phases that were falsely detected in the initial results can be also teased out. To better characterize P and S phases, our next step is to apply subspace clustering methods directly to the waveforms, instead of using the cross-correlation coefficients of detected phases. After that, supervised learning, e.g., a convolutional neural network, can be employed to improve the pick accuracy. Updated results will be presented at the meeting.

An accelerated subspace iteration for eigenvector derivatives

NASA Technical Reports Server (NTRS)

Ting, Tienko

1991-01-01

An accelerated subspace iteration method for calculating eigenvector derivatives has been developed. Factors affecting the effectiveness and the reliability of the subspace iteration are identified, and effective strategies concerning these factors are presented. The method has been implemented, and the results of a demonstration problem are presented.

A Higher-Order Generalized Singular Value Decomposition for Comparison of Global mRNA Expression from Multiple Organisms

PubMed Central

Ponnapalli, Sri Priya; Saunders, Michael A.; Van Loan, Charles F.; Alter, Orly

2011-01-01

The number of high-dimensional datasets recording multiple aspects of a single phenomenon is increasing in many areas of science, accompanied by a need for mathematical frameworks that can compare multiple large-scale matrices with different row dimensions. The only such framework to date, the generalized singular value decomposition (GSVD), is limited to two matrices. We mathematically define a higher-order GSVD (HO GSVD) for N≥2 matrices , each with full column rank. Each matrix is exactly factored as Di = UiΣiVT, where V, identical in all factorizations, is obtained from the eigensystem SV = VΛ of the arithmetic mean S of all pairwise quotients of the matrices , i≠j. We prove that this decomposition extends to higher orders almost all of the mathematical properties of the GSVD. The matrix S is nondefective with V and Λ real. Its eigenvalues satisfy λk≥1. Equality holds if and only if the corresponding eigenvector vk is a right basis vector of equal significance in all matrices Di and Dj, that is σi,k/σj,k = 1 for all i and j, and the corresponding left basis vector ui,k is orthogonal to all other vectors in Ui for all i. The eigenvalues λk = 1, therefore, define the “common HO GSVD subspace.” We illustrate the HO GSVD with a comparison of genome-scale cell-cycle mRNA expression from S. pombe, S. cerevisiae and human. Unlike existing algorithms, a mapping among the genes of these disparate organisms is not required. We find that the approximately common HO GSVD subspace represents the cell-cycle mRNA expression oscillations, which are similar among the datasets. Simultaneous reconstruction in the common subspace, therefore, removes the experimental artifacts, which are dissimilar, from the datasets. In the simultaneous sequence-independent classification of the genes of the three organisms in this common subspace, genes of highly conserved sequences but significantly different cell-cycle peak times are correctly classified. PMID:22216090

Drug-target interaction prediction using ensemble learning and dimensionality reduction.

PubMed

Ezzat, Ali; Wu, Min; Li, Xiao-Li; Kwoh, Chee-Keong

2017-10-01

Experimental prediction of drug-target interactions is expensive, time-consuming and tedious. Fortunately, computational methods help narrow down the search space for interaction candidates to be further examined via wet-lab techniques. Nowadays, the number of attributes/features for drugs and targets, as well as the amount of their interactions, are increasing, making these computational methods inefficient or occasionally prohibitive. This motivates us to derive a reduced feature set for prediction. In addition, since ensemble learning techniques are widely used to improve the classification performance, it is also worthwhile to design an ensemble learning framework to enhance the performance for drug-target interaction prediction. In this paper, we propose a framework for drug-target interaction prediction leveraging both feature dimensionality reduction and ensemble learning. First, we conducted feature subspacing to inject diversity into the classifier ensemble. Second, we applied three different dimensionality reduction methods to the subspaced features. Third, we trained homogeneous base learners with the reduced features and then aggregated their scores to derive the final predictions. For base learners, we selected two classifiers, namely Decision Tree and Kernel Ridge Regression, resulting in two variants of ensemble models, EnsemDT and EnsemKRR, respectively. In our experiments, we utilized AUC (Area under ROC Curve) as an evaluation metric. We compared our proposed methods with various state-of-the-art methods under 5-fold cross validation. Experimental results showed EnsemKRR achieving the highest AUC (94.3%) for predicting drug-target interactions. In addition, dimensionality reduction helped improve the performance of EnsemDT. In conclusion, our proposed methods produced significant improvements for drug-target interaction prediction. Copyright © 2017 Elsevier Inc. All rights reserved.

Random Deep Belief Networks for Recognizing Emotions from Speech Signals.

PubMed

Wen, Guihua; Li, Huihui; Huang, Jubing; Li, Danyang; Xun, Eryang

2017-01-01

Now the human emotions can be recognized from speech signals using machine learning methods; however, they are challenged by the lower recognition accuracies in real applications due to lack of the rich representation ability. Deep belief networks (DBN) can automatically discover the multiple levels of representations in speech signals. To make full of its advantages, this paper presents an ensemble of random deep belief networks (RDBN) method for speech emotion recognition. It firstly extracts the low level features of the input speech signal and then applies them to construct lots of random subspaces. Each random subspace is then provided for DBN to yield the higher level features as the input of the classifier to output an emotion label. All outputted emotion labels are then fused through the majority voting to decide the final emotion label for the input speech signal. The conducted experimental results on benchmark speech emotion databases show that RDBN has better accuracy than the compared methods for speech emotion recognition.

Random Deep Belief Networks for Recognizing Emotions from Speech Signals

PubMed Central

Li, Huihui; Huang, Jubing; Li, Danyang; Xun, Eryang

2017-01-01

Now the human emotions can be recognized from speech signals using machine learning methods; however, they are challenged by the lower recognition accuracies in real applications due to lack of the rich representation ability. Deep belief networks (DBN) can automatically discover the multiple levels of representations in speech signals. To make full of its advantages, this paper presents an ensemble of random deep belief networks (RDBN) method for speech emotion recognition. It firstly extracts the low level features of the input speech signal and then applies them to construct lots of random subspaces. Each random subspace is then provided for DBN to yield the higher level features as the input of the classifier to output an emotion label. All outputted emotion labels are then fused through the majority voting to decide the final emotion label for the input speech signal. The conducted experimental results on benchmark speech emotion databases show that RDBN has better accuracy than the compared methods for speech emotion recognition. PMID:28356908

System identification using Nuclear Norm & Tabu Search optimization

NASA Astrophysics Data System (ADS)

Ahmed, Asif A.; Schoen, Marco P.; Bosworth, Ken W.

2018-01-01

In recent years, subspace System Identification (SI) algorithms have seen increased research, stemming from advanced minimization methods being applied to the Nuclear Norm (NN) approach in system identification. These minimization algorithms are based on hard computing methodologies. To the authors’ knowledge, as of now, there has been no work reported that utilizes soft computing algorithms to address the minimization problem within the nuclear norm SI framework. A linear, time-invariant, discrete time system is used in this work as the basic model for characterizing a dynamical system to be identified. The main objective is to extract a mathematical model from collected experimental input-output data. Hankel matrices are constructed from experimental data, and the extended observability matrix is employed to define an estimated output of the system. This estimated output and the actual - measured - output are utilized to construct a minimization problem. An embedded rank measure assures minimum state realization outcomes. Current NN-SI algorithms employ hard computing algorithms for minimization. In this work, we propose a simple Tabu Search (TS) algorithm for minimization. TS algorithm based SI is compared with the iterative Alternating Direction Method of Multipliers (ADMM) line search optimization based NN-SI. For comparison, several different benchmark system identification problems are solved by both approaches. Results show improved performance of the proposed SI-TS algorithm compared to the NN-SI ADMM algorithm.

Discrete-State Stochastic Models of Calcium-Regulated Calcium Influx and Subspace Dynamics Are Not Well-Approximated by ODEs That Neglect Concentration Fluctuations

PubMed Central

Weinberg, Seth H.; Smith, Gregory D.

2012-01-01

Cardiac myocyte calcium signaling is often modeled using deterministic ordinary differential equations (ODEs) and mass-action kinetics. However, spatially restricted “domains” associated with calcium influx are small enough (e.g., 10−17 liters) that local signaling may involve 1–100 calcium ions. Is it appropriate to model the dynamics of subspace calcium using deterministic ODEs or, alternatively, do we require stochastic descriptions that account for the fundamentally discrete nature of these local calcium signals? To address this question, we constructed a minimal Markov model of a calcium-regulated calcium channel and associated subspace. We compared the expected value of fluctuating subspace calcium concentration (a result that accounts for the small subspace volume) with the corresponding deterministic model (an approximation that assumes large system size). When subspace calcium did not regulate calcium influx, the deterministic and stochastic descriptions agreed. However, when calcium binding altered channel activity in the model, the continuous deterministic description often deviated significantly from the discrete stochastic model, unless the subspace volume is unrealistically large and/or the kinetics of the calcium binding are sufficiently fast. This principle was also demonstrated using a physiologically realistic model of calmodulin regulation of L-type calcium channels introduced by Yue and coworkers. PMID:23509597

Hyperspectral image compressing using wavelet-based method

NASA Astrophysics Data System (ADS)

Yu, Hui; Zhang, Zhi-jie; Lei, Bo; Wang, Chen-sheng

2017-10-01

Hyperspectral imaging sensors can acquire images in hundreds of continuous narrow spectral bands. Therefore each object presented in the image can be identified from their spectral response. However, such kind of imaging brings a huge amount of data, which requires transmission, processing, and storage resources for both airborne and space borne imaging. Due to the high volume of hyperspectral image data, the exploration of compression strategies has received a lot of attention in recent years. Compression of hyperspectral data cubes is an effective solution for these problems. Lossless compression of the hyperspectral data usually results in low compression ratio, which may not meet the available resources; on the other hand, lossy compression may give the desired ratio, but with a significant degradation effect on object identification performance of the hyperspectral data. Moreover, most hyperspectral data compression techniques exploits the similarities in spectral dimensions; which requires bands reordering or regrouping, to make use of the spectral redundancy. In this paper, we explored the spectral cross correlation between different bands, and proposed an adaptive band selection method to obtain the spectral bands which contain most of the information of the acquired hyperspectral data cube. The proposed method mainly consist three steps: First, the algorithm decomposes the original hyperspectral imagery into a series of subspaces based on the hyper correlation matrix of the hyperspectral images between different bands. And then the Wavelet-based algorithm is applied to the each subspaces. At last the PCA method is applied to the wavelet coefficients to produce the chosen number of components. The performance of the proposed method was tested by using ISODATA classification method.

A parameter estimation algorithm for LFM/BPSK hybrid modulated signal intercepted by Nyquist folding receiver

NASA Astrophysics Data System (ADS)

Qiu, Zhaoyang; Wang, Pei; Zhu, Jun; Tang, Bin

2016-12-01

Nyquist folding receiver (NYFR) is a novel ultra-wideband receiver architecture which can realize wideband receiving with a small amount of equipment. Linear frequency modulated/binary phase shift keying (LFM/BPSK) hybrid modulated signal is a novel kind of low probability interception signal with wide bandwidth. The NYFR is an effective architecture to intercept the LFM/BPSK signal and the LFM/BPSK signal intercepted by the NYFR will add the local oscillator modulation. A parameter estimation algorithm for the NYFR output signal is proposed. According to the NYFR prior information, the chirp singular value ratio spectrum is proposed to estimate the chirp rate. Then, based on the output self-characteristic, matching component function is designed to estimate Nyquist zone (NZ) index. Finally, matching code and subspace method are employed to estimate the phase change points and code length. Compared with the existing methods, the proposed algorithm has a better performance. It also has no need to construct a multi-channel structure, which means the computational complexity for the NZ index estimation is small. The simulation results demonstrate the efficacy of the proposed algorithm.

ROMUSE 2.0 User Manual

DOE Office of Scientific and Technical Information (OSTI.GOV)

Khuwaileh, Bassam; Turinsky, Paul; Williams, Brian J.

2016-10-04

ROMUSE (Reduced Order Modeling Based Uncertainty/Sensitivity Estimator) is an effort within the Consortium for Advanced Simulation of Light water reactors (CASL) to provide an analysis tool to be used in conjunction with reactor core simulators, especially the Virtual Environment for Reactor Applications (VERA). ROMUSE is written in C++ and is currently capable of performing various types of parameters perturbations, uncertainty quantification, surrogate models construction and subspace analysis. Version 2.0 has the capability to interface with DAKOTA which gives ROMUSE access to the various algorithms implemented within DAKOTA. ROMUSE is mainly designed to interface with VERA and the Comprehensive Modeling andmore » Simulation Suite for Nuclear Safety Analysis and Design (SCALE) [1,2,3], however, ROMUSE can interface with any general model (e.g. python and matlab) with Input/Output (I/O) format that follows the Hierarchical Data Format 5 (HDF5). In this brief user manual, the use of ROMUSE will be overviewed and example problems will be presented and briefly discussed. The algorithms provided here range from algorithms inspired by those discussed in Ref.[4] to nuclear-specific algorithms discussed in Ref. [3].« less

Time reversal acoustics for small targets using decomposition of the time reversal operator

NASA Astrophysics Data System (ADS)

Simko, Peter C.

The method of time reversal acoustics has been the focus of considerable interest over the last twenty years. Time reversal imaging methods have made consistent progress as effective methods for signal processing since the initial demonstration that physical time reversal methods can be used to form convergent wave fields on a localized target, even under conditions of severe multipathing. Computational time reversal methods rely on the properties of the so-called 'time reversal operator' in order to extract information about the target medium. Applications for which time reversal imaging have previously been explored include medical imaging, non-destructive evaluation, and mine detection. Emphasis in this paper will fall on two topics within the general field of computational time reversal imaging. First, we will examine previous work on developing a time reversal imaging algorithm based on the MUltiple SIgnal Classification (MUSIC) algorithm. MUSIC, though computationally very intensive, has demonstrated early promise in simulations using array-based methods applicable to true volumetric (three-dimensional) imaging. We will provide a simple algorithm through which the rank of the time reversal operator subspaces can be properly quantified so that the rank of the associated null subspace can be accurately estimated near the central pulse wavelength in broadband imaging. Second, we will focus on the scattering from small acoustically rigid two dimensional cylindrical targets of elliptical cross section. Analysis of the time reversal operator eigenmodes has been well-studied for symmetric response matrices associated with symmetric systems of scattering targets. We will expand these previous results to include more general scattering systems leading to asymmetric response matrices, for which the analytical complexity increases but the physical interpretation of the time reversal operator remains unchanged. For asymmetric responses, the qualitative properties of the time reversal operator eigenmodes remain consistent with those obtained from the more tightly constrained systems.

Key Frame Extraction in the Summary Space.

PubMed

Li, Xuelong; Zhao, Bin; Lu, Xiaoqiang; Xuelong Li; Bin Zhao; Xiaoqiang Lu; Lu, Xiaoqiang; Li, Xuelong; Zhao, Bin

2018-06-01

Key frame extraction is an efficient way to create the video summary which helps users obtain a quick comprehension of the video content. Generally, the key frames should be representative of the video content, meanwhile, diverse to reduce the redundancy. Based on the assumption that the video data are near a subspace of a high-dimensional space, a new approach, named as key frame extraction in the summary space, is proposed for key frame extraction in this paper. The proposed approach aims to find the representative frames of the video and filter out similar frames from the representative frame set. First of all, the video data are mapped to a high-dimensional space, named as summary space. Then, a new representation is learned for each frame by analyzing the intrinsic structure of the summary space. Specifically, the learned representation can reflect the representativeness of the frame, and is utilized to select representative frames. Next, the perceptual hash algorithm is employed to measure the similarity of representative frames. As a result, the key frame set is obtained after filtering out similar frames from the representative frame set. Finally, the video summary is constructed by assigning the key frames in temporal order. Additionally, the ground truth, created by filtering out similar frames from human-created summaries, is utilized to evaluate the quality of the video summary. Compared with several traditional approaches, the experimental results on 80 videos from two datasets indicate the superior performance of our approach.

Simultaneous Spectral-Spatial Feature Selection and Extraction for Hyperspectral Images.

PubMed

Zhang, Lefei; Zhang, Qian; Du, Bo; Huang, Xin; Tang, Yuan Yan; Tao, Dacheng

2018-01-01

In hyperspectral remote sensing data mining, it is important to take into account of both spectral and spatial information, such as the spectral signature, texture feature, and morphological property, to improve the performances, e.g., the image classification accuracy. In a feature representation point of view, a nature approach to handle this situation is to concatenate the spectral and spatial features into a single but high dimensional vector and then apply a certain dimension reduction technique directly on that concatenated vector before feed it into the subsequent classifier. However, multiple features from various domains definitely have different physical meanings and statistical properties, and thus such concatenation has not efficiently explore the complementary properties among different features, which should benefit for boost the feature discriminability. Furthermore, it is also difficult to interpret the transformed results of the concatenated vector. Consequently, finding a physically meaningful consensus low dimensional feature representation of original multiple features is still a challenging task. In order to address these issues, we propose a novel feature learning framework, i.e., the simultaneous spectral-spatial feature selection and extraction algorithm, for hyperspectral images spectral-spatial feature representation and classification. Specifically, the proposed method learns a latent low dimensional subspace by projecting the spectral-spatial feature into a common feature space, where the complementary information has been effectively exploited, and simultaneously, only the most significant original features have been transformed. Encouraging experimental results on three public available hyperspectral remote sensing datasets confirm that our proposed method is effective and efficient.

Comparing, optimizing, and benchmarking quantum-control algorithms in a unifying programming framework

DOE Office of Scientific and Technical Information (OSTI.GOV)

Machnes, S.; Institute for Theoretical Physics, University of Ulm, D-89069 Ulm; Sander, U.

2011-08-15

For paving the way to novel applications in quantum simulation, computation, and technology, increasingly large quantum systems have to be steered with high precision. It is a typical task amenable to numerical optimal control to turn the time course of pulses, i.e., piecewise constant control amplitudes, iteratively into an optimized shape. Here, we present a comparative study of optimal-control algorithms for a wide range of finite-dimensional applications. We focus on the most commonly used algorithms: GRAPE methods which update all controls concurrently, and Krotov-type methods which do so sequentially. Guidelines for their use are given and open research questions aremore » pointed out. Moreover, we introduce a unifying algorithmic framework, DYNAMO (dynamic optimization platform), designed to provide the quantum-technology community with a convenient matlab-based tool set for optimal control. In addition, it gives researchers in optimal-control techniques a framework for benchmarking and comparing newly proposed algorithms with the state of the art. It allows a mix-and-match approach with various types of gradients, update and step-size methods as well as subspace choices. Open-source code including examples is made available at http://qlib.info.« less

Progressive Visual Analytics: User-Driven Visual Exploration of In-Progress Analytics.

PubMed

Stolper, Charles D; Perer, Adam; Gotz, David

2014-12-01

As datasets grow and analytic algorithms become more complex, the typical workflow of analysts launching an analytic, waiting for it to complete, inspecting the results, and then re-Iaunching the computation with adjusted parameters is not realistic for many real-world tasks. This paper presents an alternative workflow, progressive visual analytics, which enables an analyst to inspect partial results of an algorithm as they become available and interact with the algorithm to prioritize subspaces of interest. Progressive visual analytics depends on adapting analytical algorithms to produce meaningful partial results and enable analyst intervention without sacrificing computational speed. The paradigm also depends on adapting information visualization techniques to incorporate the constantly refining results without overwhelming analysts and provide interactions to support an analyst directing the analytic. The contributions of this paper include: a description of the progressive visual analytics paradigm; design goals for both the algorithms and visualizations in progressive visual analytics systems; an example progressive visual analytics system (Progressive Insights) for analyzing common patterns in a collection of event sequences; and an evaluation of Progressive Insights and the progressive visual analytics paradigm by clinical researchers analyzing electronic medical records.

Subspace in Linear Algebra: Investigating Students' Concept Images and Interactions with the Formal Definition

ERIC Educational Resources Information Center

Wawro, Megan; Sweeney, George F.; Rabin, Jeffrey M.

2011-01-01

This paper reports on a study investigating students' ways of conceptualizing key ideas in linear algebra, with the particular results presented here focusing on student interactions with the notion of subspace. In interviews conducted with eight undergraduates, we found students' initial descriptions of subspace often varied substantially from…

Normal and abnormal tissue identification system and method for medical images such as digital mammograms

NASA Technical Reports Server (NTRS)

Heine, John J. (Inventor); Clarke, Laurence P. (Inventor); Deans, Stanley R. (Inventor); Stauduhar, Richard Paul (Inventor); Cullers, David Kent (Inventor)

2001-01-01

A system and method for analyzing a medical image to determine whether an abnormality is present, for example, in digital mammograms, includes the application of a wavelet expansion to a raw image to obtain subspace images of varying resolution. At least one subspace image is selected that has a resolution commensurate with a desired predetermined detection resolution range. A functional form of a probability distribution function is determined for each selected subspace image, and an optimal statistical normal image region test is determined for each selected subspace image. A threshold level for the probability distribution function is established from the optimal statistical normal image region test for each selected subspace image. A region size comprising at least one sector is defined, and an output image is created that includes a combination of all regions for each selected subspace image. Each region has a first value when the region intensity level is above the threshold and a second value when the region intensity level is below the threshold. This permits the localization of a potential abnormality within the image.

Active Subspaces for Wind Plant Surrogate Modeling

DOE Office of Scientific and Technical Information (OSTI.GOV)

King, Ryan N; Quick, Julian; Dykes, Katherine L

Understanding the uncertainty in wind plant performance is crucial to their cost-effective design and operation. However, conventional approaches to uncertainty quantification (UQ), such as Monte Carlo techniques or surrogate modeling, are often computationally intractable for utility-scale wind plants because of poor congergence rates or the curse of dimensionality. In this paper we demonstrate that wind plant power uncertainty can be well represented with a low-dimensional active subspace, thereby achieving a significant reduction in the dimension of the surrogate modeling problem. We apply the active sub-spaces technique to UQ of plant power output with respect to uncertainty in turbine axial inductionmore » factors, and find a single active subspace direction dominates the sensitivity in power output. When this single active subspace direction is used to construct a quadratic surrogate model, the number of model unknowns can be reduced by up to 3 orders of magnitude without compromising performance on unseen test data. We conclude that the dimension reduction achieved with active subspaces makes surrogate-based UQ approaches tractable for utility-scale wind plants.« less

Interpretation of the MEG-MUSIC scan in biomagnetic source localization

DOE Office of Scientific and Technical Information (OSTI.GOV)

Mosher, J.C.; Lewis, P.S.; Leahy, R.M.

1993-09-01

MEG-Music is a new approach to MEG source localization. MEG-Music is based on a spatio-temporal source model in which the observed biomagnetic fields are generated by a small number of current dipole sources with fixed positions/orientations and varying strengths. From the spatial covariance matrix of the observed fields, a signal subspace can be identified. The rank of this subspace is equal to the number of elemental sources present. This signal sub-space is used in a projection metric that scans the three dimensional head volume. Given a perfect signal subspace estimate and a perfect forward model, the metric will peak atmore » unity at each dipole location. In practice, the signal subspace estimate is contaminated by noise, which in turn yields MUSIC peaks which are less than unity. Previously we examined the lower bounds on localization error, independent of the choice of localization procedure. In this paper, we analyzed the effects of noise and temporal coherence on the signal subspace estimate and the resulting effects on the MEG-MUSIC peaks.« less

Degree-of-Freedom Strengthened Cascade Array for DOD-DOA Estimation in MIMO Array Systems.

PubMed

Yao, Bobin; Dong, Zhi; Zhang, Weile; Wang, Wei; Wu, Qisheng

2018-05-14

In spatial spectrum estimation, difference co-array can provide extra degrees-of-freedom (DOFs) for promoting parameter identifiability and parameter estimation accuracy. For the sake of acquiring as more DOFs as possible with a given number of physical sensors, we herein design a novel sensor array geometry named cascade array. This structure is generated by systematically connecting a uniform linear array (ULA) and a non-uniform linear array, and can provide more DOFs than some exist array structures but less than the upper-bound indicated by minimum redundant array (MRA). We further apply this cascade array into multiple input multiple output (MIMO) array systems, and propose a novel joint direction of departure (DOD) and direction of arrival (DOA) estimation algorithm, which is based on a reduced-dimensional weighted subspace fitting technique. The algorithm is angle auto-paired and computationally efficient. Theoretical analysis and numerical simulations prove the advantages and effectiveness of the proposed array structure and the related algorithm.

Binary tree eigen solver in finite element analysis

NASA Technical Reports Server (NTRS)

Akl, F. A.; Janetzke, D. C.; Kiraly, L. J.

1993-01-01

This paper presents a transputer-based binary tree eigensolver for the solution of the generalized eigenproblem in linear elastic finite element analysis. The algorithm is based on the method of recursive doubling, which parallel implementation of a number of associative operations on an arbitrary set having N elements is of the order of o(log2N), compared to (N-1) steps if implemented sequentially. The hardware used in the implementation of the binary tree consists of 32 transputers. The algorithm is written in OCCAM which is a high-level language developed with the transputers to address parallel programming constructs and to provide the communications between processors. The algorithm can be replicated to match the size of the binary tree transputer network. Parallel and sequential finite element analysis programs have been developed to solve for the set of the least-order eigenpairs using the modified subspace method. The speed-up obtained for a typical analysis problem indicates close agreement with the theoretical prediction given by the method of recursive doubling.

Quantum supercharger library: hyper-parallelism of the Hartree-Fock method.

PubMed

Fernandes, Kyle D; Renison, C Alicia; Naidoo, Kevin J

2015-07-05

We present here a set of algorithms that completely rewrites the Hartree-Fock (HF) computations common to many legacy electronic structure packages (such as GAMESS-US, GAMESS-UK, and NWChem) into a massively parallel compute scheme that takes advantage of hardware accelerators such as Graphical Processing Units (GPUs). The HF compute algorithm is core to a library of routines that we name the Quantum Supercharger Library (QSL). We briefly evaluate the QSL's performance and report that it accelerates a HF 6-31G Self-Consistent Field (SCF) computation by up to 20 times for medium sized molecules (such as a buckyball) when compared with mature Central Processing Unit algorithms available in the legacy codes in regular use by researchers. It achieves this acceleration by massive parallelization of the one- and two-electron integrals and optimization of the SCF and Direct Inversion in the Iterative Subspace routines through the use of GPU linear algebra libraries. © 2015 Wiley Periodicals, Inc. © 2015 Wiley Periodicals, Inc.

Cluster ensemble based on Random Forests for genetic data.

PubMed

Alhusain, Luluah; Hafez, Alaaeldin M

2017-01-01

Clustering plays a crucial role in several application domains, such as bioinformatics. In bioinformatics, clustering has been extensively used as an approach for detecting interesting patterns in genetic data. One application is population structure analysis, which aims to group individuals into subpopulations based on shared genetic variations, such as single nucleotide polymorphisms. Advances in DNA sequencing technology have facilitated the obtainment of genetic datasets with exceptional sizes. Genetic data usually contain hundreds of thousands of genetic markers genotyped for thousands of individuals, making an efficient means for handling such data desirable. Random Forests (RFs) has emerged as an efficient algorithm capable of handling high-dimensional data. RFs provides a proximity measure that can capture different levels of co-occurring relationships between variables. RFs has been widely considered a supervised learning method, although it can be converted into an unsupervised learning method. Therefore, RF-derived proximity measure combined with a clustering technique may be well suited for determining the underlying structure of unlabeled data. This paper proposes, RFcluE, a cluster ensemble approach for determining the underlying structure of genetic data based on RFs. The approach comprises a cluster ensemble framework to combine multiple runs of RF clustering. Experiments were conducted on high-dimensional, real genetic dataset to evaluate the proposed approach. The experiments included an examination of the impact of parameter changes, comparing RFcluE performance against other clustering methods, and an assessment of the relationship between the diversity and quality of the ensemble and its effect on RFcluE performance. This paper proposes, RFcluE, a cluster ensemble approach based on RF clustering to address the problem of population structure analysis and demonstrate the effectiveness of the approach. The paper also illustrates that applying a cluster ensemble approach, combining multiple RF clusterings, produces more robust and higher-quality results as a consequence of feeding the ensemble with diverse views of high-dimensional genetic data obtained through bagging and random subspace, the two key features of the RF algorithm.

A Rapid Convergent Low Complexity Interference Alignment Algorithm for Wireless Sensor Networks.

PubMed

Jiang, Lihui; Wu, Zhilu; Ren, Guanghui; Wang, Gangyi; Zhao, Nan

2015-07-29

Interference alignment (IA) is a novel technique that can effectively eliminate the interference and approach the sum capacity of wireless sensor networks (WSNs) when the signal-to-noise ratio (SNR) is high, by casting the desired signal and interference into different signal subspaces. The traditional alternating minimization interference leakage (AMIL) algorithm for IA shows good performance in high SNR regimes, however, the complexity of the AMIL algorithm increases dramatically as the number of users and antennas increases, posing limits to its applications in the practical systems. In this paper, a novel IA algorithm, called directional quartic optimal (DQO) algorithm, is proposed to minimize the interference leakage with rapid convergence and low complexity. The properties of the AMIL algorithm are investigated, and it is discovered that the difference between the two consecutive iteration results of the AMIL algorithm will approximately point to the convergence solution when the precoding and decoding matrices obtained from the intermediate iterations are sufficiently close to their convergence values. Based on this important property, the proposed DQO algorithm employs the line search procedure so that it can converge to the destination directly. In addition, the optimal step size can be determined analytically by optimizing a quartic function. Numerical results show that the proposed DQO algorithm can suppress the interference leakage more rapidly than the traditional AMIL algorithm, and can achieve the same level of sum rate as that of AMIL algorithm with far less iterations and execution time.

Biomedical image classification based on a cascade of an SVM with a reject option and subspace analysis.

PubMed

Lin, Dongyun; Sun, Lei; Toh, Kar-Ann; Zhang, Jing Bo; Lin, Zhiping

2018-05-01

Automated biomedical image classification could confront the challenges of high level noise, image blur, illumination variation and complicated geometric correspondence among various categorical biomedical patterns in practice. To handle these challenges, we propose a cascade method consisting of two stages for biomedical image classification. At stage 1, we propose a confidence score based classification rule with a reject option for a preliminary decision using the support vector machine (SVM). The testing images going through stage 1 are separated into two groups based on their confidence scores. Those testing images with sufficiently high confidence scores are classified at stage 1 while the others with low confidence scores are rejected and fed to stage 2. At stage 2, the rejected images from stage 1 are first processed by a subspace analysis technique called eigenfeature regularization and extraction (ERE), and then classified by another SVM trained in the transformed subspace learned by ERE. At both stages, images are represented based on two types of local features, i.e., SIFT and SURF, respectively. They are encoded using various bag-of-words (BoW) models to handle biomedical patterns with and without geometric correspondence, respectively. Extensive experiments are implemented to evaluate the proposed method on three benchmark real-world biomedical image datasets. The proposed method significantly outperforms several competing state-of-the-art methods in terms of classification accuracy. Copyright © 2018 Elsevier Ltd. All rights reserved.

Random ensemble learning for EEG classification.

PubMed

Hosseini, Mohammad-Parsa; Pompili, Dario; Elisevich, Kost; Soltanian-Zadeh, Hamid

2018-01-01

Real-time detection of seizure activity in epilepsy patients is critical in averting seizure activity and improving patients' quality of life. Accurate evaluation, presurgical assessment, seizure prevention, and emergency alerts all depend on the rapid detection of seizure onset. A new method of feature selection and classification for rapid and precise seizure detection is discussed wherein informative components of electroencephalogram (EEG)-derived data are extracted and an automatic method is presented using infinite independent component analysis (I-ICA) to select independent features. The feature space is divided into subspaces via random selection and multichannel support vector machines (SVMs) are used to classify these subspaces. The result of each classifier is then combined by majority voting to establish the final output. In addition, a random subspace ensemble using a combination of SVM, multilayer perceptron (MLP) neural network and an extended k-nearest neighbors (k-NN), called extended nearest neighbor (ENN), is developed for the EEG and electrocorticography (ECoG) big data problem. To evaluate the solution, a benchmark ECoG of eight patients with temporal and extratemporal epilepsy was implemented in a distributed computing framework as a multitier cloud-computing architecture. Using leave-one-out cross-validation, the accuracy, sensitivity, specificity, and both false positive and false negative ratios of the proposed method were found to be 0.97, 0.98, 0.96, 0.04, and 0.02, respectively. Application of the solution to cases under investigation with ECoG has also been effected to demonstrate its utility. Copyright © 2017 Elsevier B.V. All rights reserved.

Local Subspace Classifier with Transform-Invariance for Image Classification

NASA Astrophysics Data System (ADS)

Hotta, Seiji

A family of linear subspace classifiers called local subspace classifier (LSC) outperforms the k-nearest neighbor rule (kNN) and conventional subspace classifiers in handwritten digit classification. However, LSC suffers very high sensitivity to image transformations because it uses projection and the Euclidean distances for classification. In this paper, I present a combination of a local subspace classifier (LSC) and a tangent distance (TD) for improving accuracy of handwritten digit recognition. In this classification rule, we can deal with transform-invariance easily because we are able to use tangent vectors for approximation of transformations. However, we cannot use tangent vectors in other type of images such as color images. Hence, kernel LSC (KLSC) is proposed for incorporating transform-invariance into LSC via kernel mapping. The performance of the proposed methods is verified with the experiments on handwritten digit and color image classification.

The trust-region self-consistent field method in Kohn-Sham density-functional theory.

PubMed

Thøgersen, Lea; Olsen, Jeppe; Köhn, Andreas; Jørgensen, Poul; Sałek, Paweł; Helgaker, Trygve

2005-08-15

The trust-region self-consistent field (TRSCF) method is extended to the optimization of the Kohn-Sham energy. In the TRSCF method, both the Roothaan-Hall step and the density-subspace minimization step are replaced by trust-region optimizations of local approximations to the Kohn-Sham energy, leading to a controlled, monotonic convergence towards the optimized energy. Previously the TRSCF method has been developed for optimization of the Hartree-Fock energy, which is a simple quadratic function in the density matrix. However, since the Kohn-Sham energy is a nonquadratic function of the density matrix, the local energy functions must be generalized for use with the Kohn-Sham model. Such a generalization, which contains the Hartree-Fock model as a special case, is presented here. For comparison, a rederivation of the popular direct inversion in the iterative subspace (DIIS) algorithm is performed, demonstrating that the DIIS method may be viewed as a quasi-Newton method, explaining its fast local convergence. In the global region the convergence behavior of DIIS is less predictable. The related energy DIIS technique is also discussed and shown to be inappropriate for the optimization of the Kohn-Sham energy.

DOA Estimation for Underwater Wideband Weak Targets Based on Coherent Signal Subspace and Compressed Sensing

PubMed Central

2018-01-01

Direction of arrival (DOA) estimation is the basis for underwater target localization and tracking using towed line array sonar devices. A method of DOA estimation for underwater wideband weak targets based on coherent signal subspace (CSS) processing and compressed sensing (CS) theory is proposed. Under the CSS processing framework, wideband frequency focusing is accompanied by a two-sided correlation transformation, allowing the DOA of underwater wideband targets to be estimated based on the spatial sparsity of the targets and the compressed sensing reconstruction algorithm. Through analysis and processing of simulation data and marine trial data, it is shown that this method can accomplish the DOA estimation of underwater wideband weak targets. Results also show that this method can considerably improve the spatial spectrum of weak target signals, enhancing the ability to detect them. It can solve the problems of low directional resolution and unreliable weak-target detection in traditional beamforming technology. Compared with the conventional minimum variance distortionless response beamformers (MVDR), this method has many advantages, such as higher directional resolution, wider detection range, fewer required snapshots and more accurate detection for weak targets. PMID:29562642

Sparse electrocardiogram signals recovery based on solving a row echelon-like form of system.

PubMed

Cai, Pingmei; Wang, Guinan; Yu, Shiwei; Zhang, Hongjuan; Ding, Shuxue; Wu, Zikai

2016-02-01

The study of biology and medicine in a noise environment is an evolving direction in biological data analysis. Among these studies, analysis of electrocardiogram (ECG) signals in a noise environment is a challenging direction in personalized medicine. Due to its periodic characteristic, ECG signal can be roughly regarded as sparse biomedical signals. This study proposes a two-stage recovery algorithm for sparse biomedical signals in time domain. In the first stage, the concentration subspaces are found in advance. Then by exploiting these subspaces, the mixing matrix is estimated accurately. In the second stage, based on the number of active sources at each time point, the time points are divided into different layers. Next, by constructing some transformation matrices, these time points form a row echelon-like system. After that, the sources at each layer can be solved out explicitly by corresponding matrix operations. It is noting that all these operations are conducted under a weak sparse condition that the number of active sources is less than the number of observations. Experimental results show that the proposed method has a better performance for sparse ECG signal recovery problem.

An algorithm for separation of mixed sparse and Gaussian sources

PubMed Central

Akkalkotkar, Ameya

2017-01-01

Independent component analysis (ICA) is a ubiquitous method for decomposing complex signal mixtures into a small set of statistically independent source signals. However, in cases in which the signal mixture consists of both nongaussian and Gaussian sources, the Gaussian sources will not be recoverable by ICA and will pollute estimates of the nongaussian sources. Therefore, it is desirable to have methods for mixed ICA/PCA which can separate mixtures of Gaussian and nongaussian sources. For mixtures of purely Gaussian sources, principal component analysis (PCA) can provide a basis for the Gaussian subspace. We introduce a new method for mixed ICA/PCA which we call Mixed ICA/PCA via Reproducibility Stability (MIPReSt). Our method uses a repeated estimations technique to rank sources by reproducibility, combined with decomposition of multiple subsamplings of the original data matrix. These multiple decompositions allow us to assess component stability as the size of the data matrix changes, which can be used to determinine the dimension of the nongaussian subspace in a mixture. We demonstrate the utility of MIPReSt for signal mixtures consisting of simulated sources and real-word (speech) sources, as well as mixture of unknown composition. PMID:28414814

An algorithm for separation of mixed sparse and Gaussian sources.

PubMed

Akkalkotkar, Ameya; Brown, Kevin Scott

2017-01-01

Independent component analysis (ICA) is a ubiquitous method for decomposing complex signal mixtures into a small set of statistically independent source signals. However, in cases in which the signal mixture consists of both nongaussian and Gaussian sources, the Gaussian sources will not be recoverable by ICA and will pollute estimates of the nongaussian sources. Therefore, it is desirable to have methods for mixed ICA/PCA which can separate mixtures of Gaussian and nongaussian sources. For mixtures of purely Gaussian sources, principal component analysis (PCA) can provide a basis for the Gaussian subspace. We introduce a new method for mixed ICA/PCA which we call Mixed ICA/PCA via Reproducibility Stability (MIPReSt). Our method uses a repeated estimations technique to rank sources by reproducibility, combined with decomposition of multiple subsamplings of the original data matrix. These multiple decompositions allow us to assess component stability as the size of the data matrix changes, which can be used to determinine the dimension of the nongaussian subspace in a mixture. We demonstrate the utility of MIPReSt for signal mixtures consisting of simulated sources and real-word (speech) sources, as well as mixture of unknown composition.

A Case-Based Reasoning Method with Rank Aggregation

NASA Astrophysics Data System (ADS)

Sun, Jinhua; Du, Jiao; Hu, Jian

2018-03-01

In order to improve the accuracy of case-based reasoning (CBR), this paper addresses a new CBR framework with the basic principle of rank aggregation. First, the ranking methods are put forward in each attribute subspace of case. The ordering relation between cases on each attribute is got between cases. Then, a sorting matrix is got. Second, the similar case retrieval process from ranking matrix is transformed into a rank aggregation optimal problem, which uses the Kemeny optimal. On the basis, a rank aggregation case-based reasoning algorithm, named RA-CBR, is designed. The experiment result on UCI data sets shows that case retrieval accuracy of RA-CBR algorithm is higher than euclidean distance CBR and mahalanobis distance CBR testing.So we can get the conclusion that RA-CBR method can increase the performance and efficiency of CBR.

Prediction of high-dimensional states subject to respiratory motion: a manifold learning approach

NASA Astrophysics Data System (ADS)

Liu, Wenyang; Sawant, Amit; Ruan, Dan

2016-07-01

The development of high-dimensional imaging systems in image-guided radiotherapy provides important pathways to the ultimate goal of real-time full volumetric motion monitoring. Effective motion management during radiation treatment usually requires prediction to account for system latency and extra signal/image processing time. It is challenging to predict high-dimensional respiratory motion due to the complexity of the motion pattern combined with the curse of dimensionality. Linear dimension reduction methods such as PCA have been used to construct a linear subspace from the high-dimensional data, followed by efficient predictions on the lower-dimensional subspace. In this study, we extend such rationale to a more general manifold and propose a framework for high-dimensional motion prediction with manifold learning, which allows one to learn more descriptive features compared to linear methods with comparable dimensions. Specifically, a kernel PCA is used to construct a proper low-dimensional feature manifold, where accurate and efficient prediction can be performed. A fixed-point iterative pre-image estimation method is used to recover the predicted value in the original state space. We evaluated and compared the proposed method with a PCA-based approach on level-set surfaces reconstructed from point clouds captured by a 3D photogrammetry system. The prediction accuracy was evaluated in terms of root-mean-squared-error. Our proposed method achieved consistent higher prediction accuracy (sub-millimeter) for both 200 ms and 600 ms lookahead lengths compared to the PCA-based approach, and the performance gain was statistically significant.

Identifying quantum phase transitions with adversarial neural networks

NASA Astrophysics Data System (ADS)

Huembeli, Patrick; Dauphin, Alexandre; Wittek, Peter

2018-04-01

The identification of phases of matter is a challenging task, especially in quantum mechanics, where the complexity of the ground state appears to grow exponentially with the size of the system. Traditionally, physicists have to identify the relevant order parameters for the classification of the different phases. We here follow a radically different approach: we address this problem with a state-of-the-art deep learning technique, adversarial domain adaptation. We derive the phase diagram of the whole parameter space starting from a fixed and known subspace using unsupervised learning. This method has the advantage that the input of the algorithm can be directly the ground state without any ad hoc feature engineering. Furthermore, the dimension of the parameter space is unrestricted. More specifically, the input data set contains both labeled and unlabeled data instances. The first kind is a system that admits an accurate analytical or numerical solution, and one can recover its phase diagram. The second type is the physical system with an unknown phase diagram. Adversarial domain adaptation uses both types of data to create invariant feature extracting layers in a deep learning architecture. Once these layers are trained, we can attach an unsupervised learner to the network to find phase transitions. We show the success of this technique by applying it on several paradigmatic models: the Ising model with different temperatures, the Bose-Hubbard model, and the Su-Schrieffer-Heeger model with disorder. The method finds unknown transitions successfully and predicts transition points in close agreement with standard methods. This study opens the door to the classification of physical systems where the phase boundaries are complex such as the many-body localization problem or the Bose glass phase.

Subspace algorithms for identifying separable-in-denominator 2D systems with deterministic-stochastic inputs

NASA Astrophysics Data System (ADS)

Ramos, José A.; Mercère, Guillaume

2016-12-01

In this paper, we present an algorithm for identifying two-dimensional (2D) causal, recursive and separable-in-denominator (CRSD) state-space models in the Roesser form with deterministic-stochastic inputs. The algorithm implements the N4SID, PO-MOESP and CCA methods, which are well known in the literature on 1D system identification, but here we do so for the 2D CRSD Roesser model. The algorithm solves the 2D system identification problem by maintaining the constraint structure imposed by the problem (i.e. Toeplitz and Hankel) and computes the horizontal and vertical system orders, system parameter matrices and covariance matrices of a 2D CRSD Roesser model. From a computational point of view, the algorithm has been presented in a unified framework, where the user can select which of the three methods to use. Furthermore, the identification task is divided into three main parts: (1) computing the deterministic horizontal model parameters, (2) computing the deterministic vertical model parameters and (3) computing the stochastic components. Specific attention has been paid to the computation of a stabilised Kalman gain matrix and a positive real solution when required. The efficiency and robustness of the unified algorithm have been demonstrated via a thorough simulation example.

A Type-2 Block-Component-Decomposition Based 2D AOA Estimation Algorithm for an Electromagnetic Vector Sensor Array

PubMed Central

Gao, Yu-Fei; Gui, Guan; Xie, Wei; Zou, Yan-Bin; Yang, Yue; Wan, Qun

2017-01-01

This paper investigates a two-dimensional angle of arrival (2D AOA) estimation algorithm for the electromagnetic vector sensor (EMVS) array based on Type-2 block component decomposition (BCD) tensor modeling. Such a tensor decomposition method can take full advantage of the multidimensional structural information of electromagnetic signals to accomplish blind estimation for array parameters with higher resolution. However, existing tensor decomposition methods encounter many restrictions in applications of the EMVS array, such as the strict requirement for uniqueness conditions of decomposition, the inability to handle partially-polarized signals, etc. To solve these problems, this paper investigates tensor modeling for partially-polarized signals of an L-shaped EMVS array. The 2D AOA estimation algorithm based on rank-(L1,L2,·) BCD is developed, and the uniqueness condition of decomposition is analyzed. By means of the estimated steering matrix, the proposed algorithm can automatically achieve angle pair-matching. Numerical experiments demonstrate that the present algorithm has the advantages of both accuracy and robustness of parameter estimation. Even under the conditions of lower SNR, small angular separation and limited snapshots, the proposed algorithm still possesses better performance than subspace methods and the canonical polyadic decomposition (CPD) method. PMID:28448431

A Type-2 Block-Component-Decomposition Based 2D AOA Estimation Algorithm for an Electromagnetic Vector Sensor Array.

PubMed

Gao, Yu-Fei; Gui, Guan; Xie, Wei; Zou, Yan-Bin; Yang, Yue; Wan, Qun

2017-04-27

This paper investigates a two-dimensional angle of arrival (2D AOA) estimation algorithm for the electromagnetic vector sensor (EMVS) array based on Type-2 block component decomposition (BCD) tensor modeling. Such a tensor decomposition method can take full advantage of the multidimensional structural information of electromagnetic signals to accomplish blind estimation for array parameters with higher resolution. However, existing tensor decomposition methods encounter many restrictions in applications of the EMVS array, such as the strict requirement for uniqueness conditions of decomposition, the inability to handle partially-polarized signals, etc. To solve these problems, this paper investigates tensor modeling for partially-polarized signals of an L-shaped EMVS array. The 2D AOA estimation algorithm based on rank- ( L 1 , L 2 , · ) BCD is developed, and the uniqueness condition of decomposition is analyzed. By means of the estimated steering matrix, the proposed algorithm can automatically achieve angle pair-matching. Numerical experiments demonstrate that the present algorithm has the advantages of both accuracy and robustness of parameter estimation. Even under the conditions of lower SNR, small angular separation and limited snapshots, the proposed algorithm still possesses better performance than subspace methods and the canonical polyadic decomposition (CPD) method.

LLNL Location and Detection Research

DOE Office of Scientific and Technical Information (OSTI.GOV)

Myers, S C; Harris, D B; Anderson, M L

2003-07-16

We present two LLNL research projects in the topical areas of location and detection. The first project assesses epicenter accuracy using a multiple-event location algorithm, and the second project employs waveform subspace Correlation to detect and identify events at Fennoscandian mines. Accurately located seismic events are the bases of location calibration. A well-characterized set of calibration events enables new Earth model development, empirical calibration, and validation of models. In a recent study, Bondar et al. (2003) develop network coverage criteria for assessing the accuracy of event locations that are determined using single-event, linearized inversion methods. These criteria are conservative andmore » are meant for application to large bulletins where emphasis is on catalog completeness and any given event location may be improved through detailed analysis or application of advanced algorithms. Relative event location techniques are touted as advancements that may improve absolute location accuracy by (1) ensuring an internally consistent dataset, (2) constraining a subset of events to known locations, and (3) taking advantage of station and event correlation structure. Here we present the preliminary phase of this work in which we use Nevada Test Site (NTS) nuclear explosions, with known locations, to test the effect of travel-time model accuracy on relative location accuracy. Like previous studies, we find that the reference velocity-model and relative-location accuracy are highly correlated. We also find that metrics based on travel-time residual of relocated events are not a reliable for assessing either velocity-model or relative-location accuracy. In the topical area of detection, we develop specialized correlation (subspace) detectors for the principal mines surrounding the ARCES station located in the European Arctic. Our objective is to provide efficient screens for explosions occurring in the mines of the Kola Peninsula (Kovdor, Zapolyarny, Olenogorsk, Khibiny) and the major iron mines of northern Sweden (Malmberget, Kiruna). In excess of 90% of the events detected by the ARCES station are mining explosions, and a significant fraction are from these northern mining groups. The primary challenge in developing waveform correlation detectors is the degree of variation in the source time histories of the shots, which can result in poor correlation among events even in close proximity. Our approach to solving this problem is to use lagged subspace correlation detectors, which offer some prospect of compensating for variation and uncertainty in source time functions.« less

Regularization by Functions of Bounded Variation and Applications to Image Enhancement

DOE Office of Scientific and Technical Information (OSTI.GOV)

Casas, E.; Kunisch, K.; Pola, C.

1999-09-15

Optimization problems regularized by bounded variation seminorms are analyzed. The optimality system is obtained and finite-dimensional approximations of bounded variation function spaces as well as of the optimization problems are studied. It is demonstrated that the choice of the vector norm in the definition of the bounded variation seminorm is of special importance for approximating subspaces consisting of piecewise constant functions. Algorithms based on a primal-dual framework that exploit the structure of these nondifferentiable optimization problems are proposed. Numerical examples are given for denoising of blocky images with very high noise.

A Model Comparison for Characterizing Protein Motions from Structure

NASA Astrophysics Data System (ADS)

David, Charles; Jacobs, Donald

2011-10-01

A comparative study is made using three computational models that characterize native state dynamics starting from known protein structures taken from four distinct SCOP classifications. A geometrical simulation is performed, and the results are compared to the elastic network model and molecular dynamics. The essential dynamics is quantified by a direct analysis of a mode subspace constructed from ANM and a principal component analysis on both the FRODA and MD trajectories using root mean square inner product and principal angles. Relative subspace sizes and overlaps are visualized using the projection of displacement vectors on the model modes. Additionally, a mode subspace is constructed from PCA on an exemplar set of X-ray crystal structures in order to determine similarly with respect to the generated ensembles. Quantitative analysis reveals there is significant overlap across the three model subspaces and the model independent subspace. These results indicate that structure is the key determinant for native state dynamics.

Zeno subspace in quantum-walk dynamics

NASA Astrophysics Data System (ADS)

Chandrashekar, C. M.

2010-11-01

We investigate discrete-time quantum-walk evolution under the influence of periodic measurements in position subspace. The undisturbed survival probability of the particle at the position subspace P(0,t) is compared with the survival probability after frequent (n) measurements at interval τ=t/n, P(0,τ)n. We show that P(0,τ)n>P(0,t) leads to the quantum Zeno effect in position subspace when a parameter θ in the quantum coin operations and frequency of measurements is greater than the critical value, θ>θc and n>nc. This Zeno effect in the subspace preserves the dynamics in coin Hilbert space of the walk dynamics and has the potential to play a significant role in quantum tasks such as preserving the quantum state of the particle at any particular position, and to understand the Zeno dynamics in a multidimensional system that is highly transient in nature.

Active Subspaces of Airfoil Shape Parameterizations

NASA Astrophysics Data System (ADS)

Grey, Zachary J.; Constantine, Paul G.

2018-05-01

Design and optimization benefit from understanding the dependence of a quantity of interest (e.g., a design objective or constraint function) on the design variables. A low-dimensional active subspace, when present, identifies important directions in the space of design variables; perturbing a design along the active subspace associated with a particular quantity of interest changes that quantity more, on average, than perturbing the design orthogonally to the active subspace. This low-dimensional structure provides insights that characterize the dependence of quantities of interest on design variables. Airfoil design in a transonic flow field with a parameterized geometry is a popular test problem for design methodologies. We examine two particular airfoil shape parameterizations, PARSEC and CST, and study the active subspaces present in two common design quantities of interest, transonic lift and drag coefficients, under each shape parameterization. We mathematically relate the two parameterizations with a common polynomial series. The active subspaces enable low-dimensional approximations of lift and drag that relate to physical airfoil properties. In particular, we obtain and interpret a two-dimensional approximation of both transonic lift and drag, and we show how these approximation inform a multi-objective design problem.

Adiabatic evolution of decoherence-free subspaces and its shortcuts

NASA Astrophysics Data System (ADS)

Wu, S. L.; Huang, X. L.; Li, H.; Yi, X. X.

2017-10-01

The adiabatic theorem and shortcuts to adiabaticity for time-dependent open quantum systems are explored in this paper. Starting from the definition of dynamical stable decoherence-free subspace, we show that, under a compact adiabatic condition, the quantum state remains in the time-dependent decoherence-free subspace with an extremely high purity, even though the dynamics of the open quantum system may not be adiabatic. The adiabatic condition mentioned here in the adiabatic theorem for open systems is very similar to that for closed quantum systems, except that the operators required to change slowly are the Lindblad operators. We also show that the adiabatic evolution of decoherence-free subspaces depends on the existence of instantaneous decoherence-free subspaces, which requires that the Hamiltonian of open quantum systems be engineered according to the incoherent control protocol. In addition, shortcuts to adiabaticity for adiabatic decoherence-free subspaces are also presented based on the transitionless quantum driving method. Finally, we provide an example that consists of a two-level system coupled to a broadband squeezed vacuum field to show our theory. Our approach employs Markovian master equations and the theory can apply to finite-dimensional quantum open systems.

Deep and Structured Robust Information Theoretic Learning for Image Analysis.

PubMed

Deng, Yue; Bao, Feng; Deng, Xuesong; Wang, Ruiping; Kong, Youyong; Dai, Qionghai

2016-07-07

This paper presents a robust information theoretic (RIT) model to reduce the uncertainties, i.e. missing and noisy labels, in general discriminative data representation tasks. The fundamental pursuit of our model is to simultaneously learn a transformation function and a discriminative classifier that maximize the mutual information of data and their labels in the latent space. In this general paradigm, we respectively discuss three types of the RIT implementations with linear subspace embedding, deep transformation and structured sparse learning. In practice, the RIT and deep RIT are exploited to solve the image categorization task whose performances will be verified on various benchmark datasets. The structured sparse RIT is further applied to a medical image analysis task for brain MRI segmentation that allows group-level feature selections on the brain tissues.

Imaging of heart acoustic based on the sub-space methods using a microphone array.

PubMed

Moghaddasi, Hanie; Almasganj, Farshad; Zoroufian, Arezoo

2017-07-01

Heart disease is one of the leading causes of death around the world. Phonocardiogram (PCG) is an important bio-signal which represents the acoustic activity of heart, typically without any spatiotemporal information of the involved acoustic sources. The aim of this study is to analyze the PCG by employing a microphone array by which the heart internal sound sources could be localized, too. In this paper, it is intended to propose a modality by which the locations of the active sources in the heart could also be investigated, during a cardiac cycle. In this way, a microphone array with six microphones is employed as the recording set up to be put on the human chest. In the following, the Group Delay MUSIC algorithm which is a sub-space based localization method is used to estimate the location of the heart sources in different phases of the PCG. We achieved to 0.14cm mean error for the sources of first heart sound (S 1 ) simulator and 0.21cm mean error for the sources of second heart sound (S 2 ) simulator with Group Delay MUSIC algorithm. The acoustical diagrams created for human subjects show distinct patterns in various phases of the cardiac cycles such as the first and second heart sounds. Moreover, the evaluated source locations for the heart valves are matched with the ones that are obtained via the 4-dimensional (4D) echocardiography applied, to a real human case. Imaging of heart acoustic map presents a new outlook to indicate the acoustic properties of cardiovascular system and disorders of valves and thereby, in the future, could be used as a new diagnostic tool. Copyright © 2017. Published by Elsevier B.V.

A new method to real-normalize measured complex modes

NASA Technical Reports Server (NTRS)

Wei, Max L.; Allemang, Randall J.; Zhang, Qiang; Brown, David L.

1987-01-01

A time domain subspace iteration technique is presented to compute a set of normal modes from the measured complex modes. By using the proposed method, a large number of physical coordinates are reduced to a smaller number of model or principal coordinates. Subspace free decay time responses are computed using properly scaled complex modal vectors. Companion matrix for the general case of nonproportional damping is then derived in the selected vector subspace. Subspace normal modes are obtained through eigenvalue solution of the (M sub N) sup -1 (K sub N) matrix and transformed back to the physical coordinates to get a set of normal modes. A numerical example is presented to demonstrate the outlined theory.

Visual exploration of high-dimensional data through subspace analysis and dynamic projections

DOE PAGES

Liu, S.; Wang, B.; Thiagarajan, J. J.; ...

2015-06-01

Here, we introduce a novel interactive framework for visualizing and exploring high-dimensional datasets based on subspace analysis and dynamic projections. We assume the high-dimensional dataset can be represented by a mixture of low-dimensional linear subspaces with mixed dimensions, and provide a method to reliably estimate the intrinsic dimension and linear basis of each subspace extracted from the subspace clustering. Subsequently, we use these bases to define unique 2D linear projections as viewpoints from which to visualize the data. To understand the relationships among the different projections and to discover hidden patterns, we connect these projections through dynamic projections that createmore » smooth animated transitions between pairs of projections. We introduce the view transition graph, which provides flexible navigation among these projections to facilitate an intuitive exploration. Finally, we provide detailed comparisons with related systems, and use real-world examples to demonstrate the novelty and usability of our proposed framework.« less

Visual Exploration of High-Dimensional Data through Subspace Analysis and Dynamic Projections

DOE Office of Scientific and Technical Information (OSTI.GOV)

Liu, S.; Wang, B.; Thiagarajan, Jayaraman J.

2015-06-01

We introduce a novel interactive framework for visualizing and exploring high-dimensional datasets based on subspace analysis and dynamic projections. We assume the high-dimensional dataset can be represented by a mixture of low-dimensional linear subspaces with mixed dimensions, and provide a method to reliably estimate the intrinsic dimension and linear basis of each subspace extracted from the subspace clustering. Subsequently, we use these bases to define unique 2D linear projections as viewpoints from which to visualize the data. To understand the relationships among the different projections and to discover hidden patterns, we connect these projections through dynamic projections that create smoothmore » animated transitions between pairs of projections. We introduce the view transition graph, which provides flexible navigation among these projections to facilitate an intuitive exploration. Finally, we provide detailed comparisons with related systems, and use real-world examples to demonstrate the novelty and usability of our proposed framework.« less

Tensor Train Neighborhood Preserving Embedding

NASA Astrophysics Data System (ADS)

Wang, Wenqi; Aggarwal, Vaneet; Aeron, Shuchin

2018-05-01

In this paper, we propose a Tensor Train Neighborhood Preserving Embedding (TTNPE) to embed multi-dimensional tensor data into low dimensional tensor subspace. Novel approaches to solve the optimization problem in TTNPE are proposed. For this embedding, we evaluate novel trade-off gain among classification, computation, and dimensionality reduction (storage) for supervised learning. It is shown that compared to the state-of-the-arts tensor embedding methods, TTNPE achieves superior trade-off in classification, computation, and dimensionality reduction in MNIST handwritten digits and Weizmann face datasets.

Computation of Molecular Spectra on a Quantum Processor with an Error-Resilient Algorithm

DOE PAGES

Colless, J. I.; Ramasesh, V. V.; Dahlen, D.; ...

2018-02-12

Harnessing the full power of nascent quantum processors requires the efficient management of a limited number of quantum bits with finite coherent lifetimes. Hybrid algorithms, such as the variational quantum eigensolver (VQE), leverage classical resources to reduce the required number of quantum gates. Experimental demonstrations of VQE have resulted in calculation of Hamiltonian ground states, and a new theoretical approach based on a quantum subspace expansion (QSE) has outlined a procedure for determining excited states that are central to dynamical processes. Here, we use a superconducting-qubit-based processor to apply the QSE approach to the H 2 molecule, extracting both groundmore » and excited states without the need for auxiliary qubits or additional minimization. Further, we show that this extended protocol can mitigate the effects of incoherent errors, potentially enabling larger-scale quantum simulations without the need for complex error-correction techniques.« less

Computation of Molecular Spectra on a Quantum Processor with an Error-Resilient Algorithm

NASA Astrophysics Data System (ADS)

Colless, J. I.; Ramasesh, V. V.; Dahlen, D.; Blok, M. S.; Kimchi-Schwartz, M. E.; McClean, J. R.; Carter, J.; de Jong, W. A.; Siddiqi, I.

2018-02-01

Harnessing the full power of nascent quantum processors requires the efficient management of a limited number of quantum bits with finite coherent lifetimes. Hybrid algorithms, such as the variational quantum eigensolver (VQE), leverage classical resources to reduce the required number of quantum gates. Experimental demonstrations of VQE have resulted in calculation of Hamiltonian ground states, and a new theoretical approach based on a quantum subspace expansion (QSE) has outlined a procedure for determining excited states that are central to dynamical processes. We use a superconducting-qubit-based processor to apply the QSE approach to the H2 molecule, extracting both ground and excited states without the need for auxiliary qubits or additional minimization. Further, we show that this extended protocol can mitigate the effects of incoherent errors, potentially enabling larger-scale quantum simulations without the need for complex error-correction techniques.

Computation of Molecular Spectra on a Quantum Processor with an Error-Resilient Algorithm

DOE Office of Scientific and Technical Information (OSTI.GOV)

Colless, J. I.; Ramasesh, V. V.; Dahlen, D.

Harnessing the full power of nascent quantum processors requires the efficient management of a limited number of quantum bits with finite coherent lifetimes. Hybrid algorithms, such as the variational quantum eigensolver (VQE), leverage classical resources to reduce the required number of quantum gates. Experimental demonstrations of VQE have resulted in calculation of Hamiltonian ground states, and a new theoretical approach based on a quantum subspace expansion (QSE) has outlined a procedure for determining excited states that are central to dynamical processes. Here, we use a superconducting-qubit-based processor to apply the QSE approach to the H 2 molecule, extracting both groundmore » and excited states without the need for auxiliary qubits or additional minimization. Further, we show that this extended protocol can mitigate the effects of incoherent errors, potentially enabling larger-scale quantum simulations without the need for complex error-correction techniques.« less

Adaptive Sparse Representation for Source Localization with Gain/Phase Errors

PubMed Central

Sun, Ke; Liu, Yimin; Meng, Huadong; Wang, Xiqin

2011-01-01

Sparse representation (SR) algorithms can be implemented for high-resolution direction of arrival (DOA) estimation. Additionally, SR can effectively separate the coherent signal sources because the spectrum estimation is based on the optimization technique, such as the L1 norm minimization, but not on subspace orthogonality. However, in the actual source localization scenario, an unknown gain/phase error between the array sensors is inevitable. Due to this nonideal factor, the predefined overcomplete basis mismatches the actual array manifold so that the estimation performance is degraded in SR. In this paper, an adaptive SR algorithm is proposed to improve the robustness with respect to the gain/phase error, where the overcomplete basis is dynamically adjusted using multiple snapshots and the sparse solution is adaptively acquired to match with the actual scenario. The simulation results demonstrate the estimation robustness to the gain/phase error using the proposed method. PMID:22163875

Incorporating signal-dependent noise for hyperspectral target detection

NASA Astrophysics Data System (ADS)

Morman, Christopher J.; Meola, Joseph

2015-05-01

The majority of hyperspectral target detection algorithms are developed from statistical data models employing stationary background statistics or white Gaussian noise models. Stationary background models are inaccurate as a result of two separate physical processes. First, varying background classes often exist in the imagery that possess different clutter statistics. Many algorithms can account for this variability through the use of subspaces or clustering techniques. The second physical process, which is often ignored, is a signal-dependent sensor noise term. For photon counting sensors that are often used in hyperspectral imaging systems, sensor noise increases as the measured signal level increases as a result of Poisson random processes. This work investigates the impact of this sensor noise on target detection performance. A linear noise model is developed describing sensor noise variance as a linear function of signal level. The linear noise model is then incorporated for detection of targets using data collected at Wright Patterson Air Force Base.

Krylov subspace methods for computing hydrodynamic interactions in Brownian dynamics simulations

PubMed Central

Ando, Tadashi; Chow, Edmond; Saad, Yousef; Skolnick, Jeffrey

2012-01-01

Hydrodynamic interactions play an important role in the dynamics of macromolecules. The most common way to take into account hydrodynamic effects in molecular simulations is in the context of a Brownian dynamics simulation. However, the calculation of correlated Brownian noise vectors in these simulations is computationally very demanding and alternative methods are desirable. This paper studies methods based on Krylov subspaces for computing Brownian noise vectors. These methods are related to Chebyshev polynomial approximations, but do not require eigenvalue estimates. We show that only low accuracy is required in the Brownian noise vectors to accurately compute values of dynamic and static properties of polymer and monodisperse suspension models. With this level of accuracy, the computational time of Krylov subspace methods scales very nearly as O(N2) for the number of particles N up to 10 000, which was the limit tested. The performance of the Krylov subspace methods, especially the “block” version, is slightly better than that of the Chebyshev method, even without taking into account the additional cost of eigenvalue estimates required by the latter. Furthermore, at N = 10 000, the Krylov subspace method is 13 times faster than the exact Cholesky method. Thus, Krylov subspace methods are recommended for performing large-scale Brownian dynamics simulations with hydrodynamic interactions. PMID:22897254

Physical subspace in a model of the quantized electromagnetic field coupled to an external field with an indefinite metric

NASA Astrophysics Data System (ADS)

Suzuki, Akito

2008-04-01

We study a model of the quantized electromagnetic field interacting with an external static source ρ in the Feynman (Lorentz) gauge and construct the quantized radiation field Aμ (μ=0,1,2,3) as an operator-valued distribution acting on the Fock space F with an indefinite metric. By using the Gupta subsidiary condition ∂μAμ(x)(+)Ψ=0, one can select the physical subspace Vphys. According to the Gupta-Bleuler formalism, Vphys is a non-negative subspace so that elements of Vphys, called physical states, can be probabilistically interpretable. Indeed, assuming that the external source ρ is infrared regular, i.e., ρ̂/∣k∣3/2ɛL2(R3), we can characterize the physical subspace Vphys and show that Vphys is non-negative. In addition, we find that the Hamiltonian of the model is reduced to the Hamiltonian of the transverse photons with the Coulomb interaction. We, however, prove that the physical subspace is trivial, i.e., Vphys={0}, if and only if the external source ρ is infrared singular, i.e., ρ̂/∣k∣3/2∉L2(R3). We also discuss a representation different from the above representation such that the physical subspace is not trivial under the infrared singular condition.

Fast image interpolation via random forests.

PubMed

Huang, Jun-Jie; Siu, Wan-Chi; Liu, Tian-Rui

2015-10-01

This paper proposes a two-stage framework for fast image interpolation via random forests (FIRF). The proposed FIRF method gives high accuracy, as well as requires low computation. The underlying idea of this proposed work is to apply random forests to classify the natural image patch space into numerous subspaces and learn a linear regression model for each subspace to map the low-resolution image patch to high-resolution image patch. The FIRF framework consists of two stages. Stage 1 of the framework removes most of the ringing and aliasing artifacts in the initial bicubic interpolated image, while Stage 2 further refines the Stage 1 interpolated image. By varying the number of decision trees in the random forests and the number of stages applied, the proposed FIRF method can realize computationally scalable image interpolation. Extensive experimental results show that the proposed FIRF(3, 2) method achieves more than 0.3 dB improvement in peak signal-to-noise ratio over the state-of-the-art nonlocal autoregressive modeling (NARM) method. Moreover, the proposed FIRF(1, 1) obtains similar or better results as NARM while only takes its 0.3% computational time.

A Dimensionally Reduced Clustering Methodology for Heterogeneous Occupational Medicine Data Mining.

PubMed

Saâdaoui, Foued; Bertrand, Pierre R; Boudet, Gil; Rouffiac, Karine; Dutheil, Frédéric; Chamoux, Alain

2015-10-01

Clustering is a set of techniques of the statistical learning aimed at finding structures of heterogeneous partitions grouping homogenous data called clusters. There are several fields in which clustering was successfully applied, such as medicine, biology, finance, economics, etc. In this paper, we introduce the notion of clustering in multifactorial data analysis problems. A case study is conducted for an occupational medicine problem with the purpose of analyzing patterns in a population of 813 individuals. To reduce the data set dimensionality, we base our approach on the Principal Component Analysis (PCA), which is the statistical tool most commonly used in factorial analysis. However, the problems in nature, especially in medicine, are often based on heterogeneous-type qualitative-quantitative measurements, whereas PCA only processes quantitative ones. Besides, qualitative data are originally unobservable quantitative responses that are usually binary-coded. Hence, we propose a new set of strategies allowing to simultaneously handle quantitative and qualitative data. The principle of this approach is to perform a projection of the qualitative variables on the subspaces spanned by quantitative ones. Subsequently, an optimal model is allocated to the resulting PCA-regressed subspaces.

Eigensolution of finite element problems in a completely connected parallel architecture

NASA Technical Reports Server (NTRS)

Akl, Fred A.; Morel, Michael R.

1989-01-01

A parallel algorithm for the solution of the generalized eigenproblem in linear elastic finite element analysis, (K)(phi)=(M)(phi)(omega), where (K) and (M) are of order N, and (omega) is of order q is presented. The parallel algorithm is based on a completely connected parallel architecture in which each processor is allowed to communicate with all other processors. The algorithm has been successfully implemented on a tightly coupled multiple-instruction-multiple-data (MIMD) parallel processing computer, Cray X-MP. A finite element model is divided into m domains each of which is assumed to process n elements. Each domain is then assigned to a processor, or to a logical processor (task) if the number of domains exceeds the number of physical processors. The macro-tasking library routines are used in mapping each domain to a user task. Computational speed-up and efficiency are used to determine the effectiveness of the algorithm. The effect of the number of domains, the number of degrees-of-freedom located along the global fronts and the dimension of the subspace on the performance of the algorithm are investigated. For a 64-element rectangular plate, speed-ups of 1.86, 3.13, 3.18 and 3.61 are achieved on two, four, six and eight processors, respectively.

Excluding Noise from Short Krylov Subspace Approximations to the Truncated Singular Value Decomposition (SVD)

DTIC Science & Technology

2017-09-27

ARL-TR-8161•SEP 2017 US Army Research Laboratory Excluding Noise from Short Krylov Subspace Approximations to the Truncated Singular Value...originator. ARL-TR-8161•SEP 2017 US Army Research Laboratory Excluding Noise from Short Krylov Subspace Approximations to the Truncated Singular Value...unlimited. October 2015–January 2016 US Army Research Laboratory ATTN: RDRL-CIH-C Aberdeen Proving Ground, MD 21005-5066 primary author’s email

Failure detection and identification

NASA Technical Reports Server (NTRS)

Massoumnia, Mohammad-Ali; Verghese, George C.; Willsky, Alan S.

1989-01-01

Using the geometric concept of an unobservability subspace, a solution is given to the problem of detecting and identifying control system component failures in linear, time-invariant systems. Conditions are developed for the existence of a causal, linear, time-invariant processor that can detect and uniquely identify a component failure, first for the case where components can fail simultaneously, and then for the case where they fail only one at a time. Explicit design algorithms are provided when these conditions are satisfied. In addition to time-domain solvability conditions, frequency-domain interpretations of the results are given, and connections are drawn with results already available in the literature.

DOE Office of Scientific and Technical Information (OSTI.GOV)

Jones, J E; Vassilevski, P S; Woodward, C S

This paper provides extensions of an element agglomeration AMG method to nonlinear elliptic problems discretized by the finite element method on general unstructured meshes. The method constructs coarse discretization spaces and corresponding coarse nonlinear operators as well as their Jacobians. We introduce both standard (fairly quasi-uniformly coarsened) and non-standard (coarsened away) coarse meshes and respective finite element spaces. We use both kind of spaces in FAS type coarse subspace correction (or Schwarz) algorithms. Their performance is illustrated on a number of model problems. The coarsened away spaces seem to perform better than the standard spaces for problems with nonlinearities inmore » the principal part of the elliptic operator.« less

The successive projection algorithm as an initialization method for brain tumor segmentation using non-negative matrix factorization.

PubMed

Sauwen, Nicolas; Acou, Marjan; Bharath, Halandur N; Sima, Diana M; Veraart, Jelle; Maes, Frederik; Himmelreich, Uwe; Achten, Eric; Van Huffel, Sabine

2017-01-01

Non-negative matrix factorization (NMF) has become a widely used tool for additive parts-based analysis in a wide range of applications. As NMF is a non-convex problem, the quality of the solution will depend on the initialization of the factor matrices. In this study, the successive projection algorithm (SPA) is proposed as an initialization method for NMF. SPA builds on convex geometry and allocates endmembers based on successive orthogonal subspace projections of the input data. SPA is a fast and reproducible method, and it aligns well with the assumptions made in near-separable NMF analyses. SPA was applied to multi-parametric magnetic resonance imaging (MRI) datasets for brain tumor segmentation using different NMF algorithms. Comparison with common initialization methods shows that SPA achieves similar segmentation quality and it is competitive in terms of convergence rate. Whereas SPA was previously applied as a direct endmember extraction tool, we have shown improved segmentation results when using SPA as an initialization method, as it allows further enhancement of the sources during the NMF iterative procedure.

XY vs X Mixer in Quantum Alternating Operator Ansatz for Optimization Problems with Constraints

NASA Technical Reports Server (NTRS)

Wang, Zhihui; Rubin, Nicholas; Rieffel, Eleanor G.

2018-01-01

Quantum Approximate Optimization Algorithm, further generalized as Quantum Alternating Operator Ansatz (QAOA), is a family of algorithms for combinatorial optimization problems. It is a leading candidate to run on emerging universal quantum computers to gain insight into quantum heuristics. In constrained optimization, penalties are often introduced so that the ground state of the cost Hamiltonian encodes the solution (a standard practice in quantum annealing). An alternative is to choose a mixing Hamiltonian such that the constraint corresponds to a constant of motion and the quantum evolution stays in the feasible subspace. Better performance of the algorithm is speculated due to a much smaller search space. We consider problems with a constant Hamming weight as the constraint. We also compare different methods of generating the generalized W-state, which serves as a natural initial state for the Hamming-weight constraint. Using graph-coloring as an example, we compare the performance of using XY model as a mixer that preserves the Hamming weight with the performance of adding a penalty term in the cost Hamiltonian.

Motor Task Variation Induces Structural Learning

PubMed Central

Braun, Daniel A.; Aertsen, Ad; Wolpert, Daniel M.; Mehring, Carsten

2009-01-01

Summary When we have learned a motor skill, such as cycling or ice-skating, we can rapidly generalize to novel tasks, such as motorcycling or rollerblading [1–8]. Such facilitation of learning could arise through two distinct mechanisms by which the motor system might adjust its control parameters. First, fast learning could simply be a consequence of the proximity of the original and final settings of the control parameters. Second, by structural learning [9–14], the motor system could constrain the parameter adjustments to conform to the control parameters' covariance structure. Thus, facilitation of learning would rely on the novel task parameters' lying on the structure of a lower-dimensional subspace that can be explored more efficiently. To test between these two hypotheses, we exposed subjects to randomly varying visuomotor tasks of fixed structure. Although such randomly varying tasks are thought to prevent learning, we show that when subsequently presented with novel tasks, subjects exhibit three key features of structural learning: facilitated learning of tasks with the same structure, strong reduction in interference normally observed when switching between tasks that require opposite control strategies, and preferential exploration along the learned structure. These results suggest that skill generalization relies on task variation and structural learning. PMID:19217296

Motor task variation induces structural learning.

PubMed

Braun, Daniel A; Aertsen, Ad; Wolpert, Daniel M; Mehring, Carsten

2009-02-24

When we have learned a motor skill, such as cycling or ice-skating, we can rapidly generalize to novel tasks, such as motorcycling or rollerblading [1-8]. Such facilitation of learning could arise through two distinct mechanisms by which the motor system might adjust its control parameters. First, fast learning could simply be a consequence of the proximity of the original and final settings of the control parameters. Second, by structural learning [9-14], the motor system could constrain the parameter adjustments to conform to the control parameters' covariance structure. Thus, facilitation of learning would rely on the novel task parameters' lying on the structure of a lower-dimensional subspace that can be explored more efficiently. To test between these two hypotheses, we exposed subjects to randomly varying visuomotor tasks of fixed structure. Although such randomly varying tasks are thought to prevent learning, we show that when subsequently presented with novel tasks, subjects exhibit three key features of structural learning: facilitated learning of tasks with the same structure, strong reduction in interference normally observed when switching between tasks that require opposite control strategies, and preferential exploration along the learned structure. These results suggest that skill generalization relies on task variation and structural learning.

Subspace Methods for Massive and Messy Data

DTIC Science & Technology

2017-07-12

Subspace Methods for Massive and Messy Data The views, opinions and/or findings contained in this report are those of the author(s) and should not...AGENCY NAME(S) AND ADDRESS (ES) U.S. Army Research Office P.O. Box 12211 Research Triangle Park, NC 27709-2211 REPORT DOCUMENTATION PAGE 11. SPONSOR...Number: W911NF-14-1-0634 Organization: University of Michigan - Ann Arbor Title: Subspace Methods for Massive and Messy Data Report Term: 0-Other

Geometry aware Stationary Subspace Analysis

DTIC Science & Technology

2016-11-22

approach to handling non-stationarity is to remove or minimize it before attempting to analyze the data. In the context of brain computer interface ( BCI ...context of brain computer interface ( BCI ) data analysis, two such note-worthy methods are stationary subspace analysis (SSA) (von Bünau et al., 2009a... BCI systems, is sCSP. Its goal is to project the data onto a subspace in which the various data classes are more separable. The sCSP method directs

Locally indistinguishable subspaces spanned by three-qubit unextendible product bases

NASA Astrophysics Data System (ADS)

Duan, Runyao; Xin, Yu; Ying, Mingsheng

2010-03-01

We study the local distinguishability of general multiqubit states and show that local projective measurements and classical communication are as powerful as the most general local measurements and classical communication. Remarkably, this indicates that the local distinguishability of multiqubit states can be decided efficiently. Another useful consequence is that a set of orthogonal n-qubit states is locally distinguishable only if the summation of their orthogonal Schmidt numbers is less than the total dimension 2n. Employing these results, we show that any orthonormal basis of a subspace spanned by arbitrary three-qubit orthogonal unextendible product bases (UPB) cannot be exactly distinguishable by local operations and classical communication. This not only reveals another intrinsic property of three-qubit orthogonal UPB but also provides a class of locally indistinguishable subspaces with dimension 4. We also explicitly construct locally indistinguishable subspaces with dimensions 3 and 5, respectively. Similar to the bipartite case, these results on multipartite locally indistinguishable subspaces can be used to estimate the one-shot environment-assisted classical capacity of a class of quantum broadcast channels.

Application of vector-valued rational approximations to the matrix eigenvalue problem and connections with Krylov subspace methods

NASA Technical Reports Server (NTRS)

Sidi, Avram

1992-01-01

Let F(z) be a vectored-valued function F: C approaches C sup N, which is analytic at z=0 and meromorphic in a neighborhood of z=0, and let its Maclaurin series be given. We use vector-valued rational approximation procedures for F(z) that are based on its Maclaurin series in conjunction with power iterations to develop bona fide generalizations of the power method for an arbitrary N X N matrix that may be diagonalizable or not. These generalizations can be used to obtain simultaneously several of the largest distinct eigenvalues and the corresponding invariant subspaces, and present a detailed convergence theory for them. In addition, it is shown that the generalized power methods of this work are equivalent to some Krylov subspace methods, among them the methods of Arnoldi and Lanczos. Thus, the theory provides a set of completely new results and constructions for these Krylov subspace methods. This theory suggests at the same time a new mode of usage for these Krylov subspace methods that were observed to possess computational advantages over their common mode of usage.

Observation of entanglement witnesses for orbital angular momentum states

NASA Astrophysics Data System (ADS)

Agnew, M.; Leach, J.; Boyd, R. W.

2012-06-01

Entanglement witnesses provide an efficient means of determining the level of entanglement of a system using the minimum number of measurements. Here we demonstrate the observation of two-dimensional entanglement witnesses in the high-dimensional basis of orbital angular momentum (OAM). In this case, the number of potentially entangled subspaces scales as d(d - 1)/2, where d is the dimension of the space. The choice of OAM as a basis is relevant as each subspace is not necessarily maximally entangled, thus providing the necessary state for certain tests of nonlocality. The expectation value of the witness gives an estimate of the state of each two-dimensional subspace belonging to the d-dimensional Hilbert space. These measurements demonstrate the degree of entanglement and therefore the suitability of the resulting subspaces for quantum information applications.

Entanglement dynamics of coupled qubits and a semi-decoherence free subspace

NASA Astrophysics Data System (ADS)

Campagnano, Gabriele; Hamma, Alioscia; Weiss, Ulrich

2010-01-01

We study the entanglement dynamics and relaxation properties of a system of two interacting qubits in the cases of (I) two independent bosonic baths and (II) one common bath. We find that in the case (II) the existence of a decoherence-free subspace (DFS) makes entanglement dynamics very rich. We show that when the system is initially in a state with a component in the DFS the relaxation time is surprisingly long, showing the existence of semi-decoherence free subspaces.

A manifold learning approach to target detection in high-resolution hyperspectral imagery

NASA Astrophysics Data System (ADS)

Ziemann, Amanda K.

Imagery collected from airborne platforms and satellites provide an important medium for remotely analyzing the content in a scene. In particular, the ability to detect a specific material within a scene is of high importance to both civilian and defense applications. This may include identifying "targets" such as vehicles, buildings, or boats. Sensors that process hyperspectral images provide the high-dimensional spectral information necessary to perform such analyses. However, for a d-dimensional hyperspectral image, it is typical for the data to inherently occupy an m-dimensional space, with m << d. In the remote sensing community, this has led to a recent increase in the use of manifold learning, which aims to characterize the embedded lower-dimensional, non-linear manifold upon which the hyperspectral data inherently lie. Classic hyperspectral data models include statistical, linear subspace, and linear mixture models, but these can place restrictive assumptions on the distribution of the data; this is particularly true when implementing traditional target detection approaches, and the limitations of these models are well-documented. With manifold learning based approaches, the only assumption is that the data reside on an underlying manifold that can be discretely modeled by a graph. The research presented here focuses on the use of graph theory and manifold learning in hyperspectral imagery. Early work explored various graph-building techniques with application to the background model of the Topological Anomaly Detection (TAD) algorithm, which is a graph theory based approach to anomaly detection. This led towards a focus on target detection, and in the development of a specific graph-based model of the data and subsequent dimensionality reduction using manifold learning. An adaptive graph is built on the data, and then used to implement an adaptive version of locally linear embedding (LLE). We artificially induce a target manifold and incorporate it into the adaptive LLE transformation; the artificial target manifold helps to guide the separation of the target data from the background data in the new, lower-dimensional manifold coordinates. Then, target detection is performed in the manifold space.

Multiclassifier information fusion methods for microarray pattern recognition

NASA Astrophysics Data System (ADS)

Braun, Jerome J.; Glina, Yan; Judson, Nicholas; Herzig-Marx, Rachel

2004-04-01

This paper addresses automatic recognition of microarray patterns, a capability that could have a major significance for medical diagnostics, enabling development of diagnostic tools for automatic discrimination of specific diseases. The paper presents multiclassifier information fusion methods for microarray pattern recognition. The input space partitioning approach based on fitness measures that constitute an a-priori gauging of classification efficacy for each subspace is investigated. Methods for generation of fitness measures, generation of input subspaces and their use in the multiclassifier fusion architecture are presented. In particular, two-level quantification of fitness that accounts for the quality of each subspace as well as the quality of individual neighborhoods within the subspace is described. Individual-subspace classifiers are Support Vector Machine based. The decision fusion stage fuses the information from mulitple SVMs along with the multi-level fitness information. Final decision fusion stage techniques, including weighted fusion as well as Dempster-Shafer theory based fusion are investigated. It should be noted that while the above methods are discussed in the context of microarray pattern recognition, they are applicable to a broader range of discrimination problems, in particular to problems involving a large number of information sources irreducible to a low-dimensional feature space.

Generation of skeletal mechanism by means of projected entropy participation indices

NASA Astrophysics Data System (ADS)

Paolucci, Samuel; Valorani, Mauro; Ciottoli, Pietro Paolo; Galassi, Riccardo Malpica

2017-11-01

When the dynamics of reactive systems develop very-slow and very-fast time scales separated by a range of active time scales, with gaps in the fast/active and slow/active time scales, then it is possible to achieve multi-scale adaptive model reduction along-with the integration of the ODEs using the G-Scheme. The scheme assumes that the dynamics is decomposed into active, slow, fast, and invariant subspaces. We derive expressions that establish a direct link between time scales and entropy production by using estimates provided by the G-Scheme. To calculate the contribution to entropy production, we resort to a standard model of a constant pressure, adiabatic, batch reactor, where the mixture temperature of the reactants is initially set above the auto-ignition temperature. Numerical experiments show that the contribution to entropy production of the fast subspace is of the same magnitude as the error threshold chosen for the identification of the decomposition of the tangent space, and the contribution of the slow subspace is generally much smaller than that of the active subspace. The information on entropy production associated with reactions within each subspace is used to define an entropy participation index that is subsequently utilized for model reduction.

Object-oriented and pixel-based classification approach for land cover using airborne long-wave infrared hyperspectral data

NASA Astrophysics Data System (ADS)

Marwaha, Richa; Kumar, Anil; Kumar, Arumugam Senthil

2015-01-01

Our primary objective was to explore a classification algorithm for thermal hyperspectral data. Minimum noise fraction is applied to thermal hyperspectral data and eight pixel-based classifiers, i.e., constrained energy minimization, matched filter, spectral angle mapper (SAM), adaptive coherence estimator, orthogonal subspace projection, mixture-tuned matched filter, target-constrained interference-minimized filter, and mixture-tuned target-constrained interference minimized filter are tested. The long-wave infrared (LWIR) has not yet been exploited for classification purposes. The LWIR data contain emissivity and temperature information about an object. A highest overall accuracy of 90.99% was obtained using the SAM algorithm for the combination of thermal data with a colored digital photograph. Similarly, an object-oriented approach is applied to thermal data. The image is segmented into meaningful objects based on properties such as geometry, length, etc., which are grouped into pixels using a watershed algorithm and an applied supervised classification algorithm, i.e., support vector machine (SVM). The best algorithm in the pixel-based category is the SAM technique. SVM is useful for thermal data, providing a high accuracy of 80.00% at a scale value of 83 and a merge value of 90, whereas for the combination of thermal data with a colored digital photograph, SVM gives the highest accuracy of 85.71% at a scale value of 82 and a merge value of 90.

Multi-Centrality Graph Spectral Decompositions and Their Application to Cyber Intrusion Detection

DOE Office of Scientific and Technical Information (OSTI.GOV)

Chen, Pin-Yu; Choudhury, Sutanay; Hero, Alfred

Many modern datasets can be represented as graphs and hence spectral decompositions such as graph principal component analysis (PCA) can be useful. Distinct from previous graph decomposition approaches based on subspace projection of a single topological feature, e.g., the centered graph adjacency matrix (graph Laplacian), we propose spectral decomposition approaches to graph PCA and graph dictionary learning that integrate multiple features, including graph walk statistics, centrality measures and graph distances to reference nodes. In this paper we propose a new PCA method for single graph analysis, called multi-centrality graph PCA (MC-GPCA), and a new dictionary learning method for ensembles ofmore » graphs, called multi-centrality graph dictionary learning (MC-GDL), both based on spectral decomposition of multi-centrality matrices. As an application to cyber intrusion detection, MC-GPCA can be an effective indicator of anomalous connectivity pattern and MC-GDL can provide discriminative basis for attack classification.« less

Direct Position Determination of Multiple Non-Circular Sources with a Moving Coprime Array.

PubMed

Zhang, Yankui; Ba, Bin; Wang, Daming; Geng, Wei; Xu, Haiyun

2018-05-08

Direct position determination (DPD) is currently a hot topic in wireless localization research as it is more accurate than traditional two-step positioning. However, current DPD algorithms are all based on uniform arrays, which have an insufficient degree of freedom and limited estimation accuracy. To improve the DPD accuracy, this paper introduces a coprime array to the position model of multiple non-circular sources with a moving array. To maximize the advantages of this coprime array, we reconstruct the covariance matrix by vectorization, apply a spatial smoothing technique, and converge the subspace data from each measuring position to establish the cost function. Finally, we obtain the position coordinates of the multiple non-circular sources. The complexity of the proposed method is computed and compared with that of other methods, and the Cramer⁻Rao lower bound of DPD for multiple sources with a moving coprime array, is derived. Theoretical analysis and simulation results show that the proposed algorithm is not only applicable to circular sources, but can also improve the positioning accuracy of non-circular sources. Compared with existing two-step positioning algorithms and DPD algorithms based on uniform linear arrays, the proposed technique offers a significant improvement in positioning accuracy with a slight increase in complexity.

Extended Krylov subspaces approximations of matrix functions. Application to computational electromagnetics

DOE Office of Scientific and Technical Information (OSTI.GOV)

Druskin, V.; Lee, Ping; Knizhnerman, L.

There is now a growing interest in the area of using Krylov subspace approximations to compute the actions of matrix functions. The main application of this approach is the solution of ODE systems, obtained after discretization of partial differential equations by method of lines. In the event that the cost of computing the matrix inverse is relatively inexpensive, it is sometimes attractive to solve the ODE using the extended Krylov subspaces, originated by actions of both positive and negative matrix powers. Examples of such problems can be found frequently in computational electromagnetics.

Problematic projection to the in-sample subspace for a kernelized anomaly detector

DOE PAGES

Theiler, James; Grosklos, Guen

2016-03-07

We examine the properties and performance of kernelized anomaly detectors, with an emphasis on the Mahalanobis-distance-based kernel RX (KRX) algorithm. Although the detector generally performs well for high-bandwidth Gaussian kernels, it exhibits problematic (in some cases, catastrophic) performance for distances that are large compared to the bandwidth. By comparing KRX to two other anomaly detectors, we can trace the problem to a projection in feature space, which arises when a pseudoinverse is used on the covariance matrix in that feature space. Here, we show that a regularized variant of KRX overcomes this difficulty and achieves superior performance over a widemore » range of bandwidths.« less

Face recognition based on two-dimensional discriminant sparse preserving projection

NASA Astrophysics Data System (ADS)

Zhang, Dawei; Zhu, Shanan

2018-04-01

In this paper, a supervised dimensionality reduction algorithm named two-dimensional discriminant sparse preserving projection (2DDSPP) is proposed for face recognition. In order to accurately model manifold structure of data, 2DDSPP constructs within-class affinity graph and between-class affinity graph by the constrained least squares (LS) and l1 norm minimization problem, respectively. Based on directly operating on image matrix, 2DDSPP integrates graph embedding (GE) with Fisher criterion. The obtained projection subspace preserves within-class neighborhood geometry structure of samples, while keeping away samples from different classes. The experimental results on the PIE and AR face databases show that 2DDSPP can achieve better recognition performance.

Fast divide-and-conquer algorithm for evaluating polarization in classical force fields

NASA Astrophysics Data System (ADS)

Nocito, Dominique; Beran, Gregory J. O.

2017-03-01

Evaluation of the self-consistent polarization energy forms a major computational bottleneck in polarizable force fields. In large systems, the linear polarization equations are typically solved iteratively with techniques based on Jacobi iterations (JI) or preconditioned conjugate gradients (PCG). Two new variants of JI are proposed here that exploit domain decomposition to accelerate the convergence of the induced dipoles. The first, divide-and-conquer JI (DC-JI), is a block Jacobi algorithm which solves the polarization equations within non-overlapping sub-clusters of atoms directly via Cholesky decomposition, and iterates to capture interactions between sub-clusters. The second, fuzzy DC-JI, achieves further acceleration by employing overlapping blocks. Fuzzy DC-JI is analogous to an additive Schwarz method, but with distance-based weighting when averaging the fuzzy dipoles from different blocks. Key to the success of these algorithms is the use of K-means clustering to identify natural atomic sub-clusters automatically for both algorithms and to determine the appropriate weights in fuzzy DC-JI. The algorithm employs knowledge of the 3-D spatial interactions to group important elements in the 2-D polarization matrix. When coupled with direct inversion in the iterative subspace (DIIS) extrapolation, fuzzy DC-JI/DIIS in particular converges in a comparable number of iterations as PCG, but with lower computational cost per iteration. In the end, the new algorithms demonstrated here accelerate the evaluation of the polarization energy by 2-3 fold compared to existing implementations of PCG or JI/DIIS.

Signatures of large-scale and local climates on the demography of white-tailed ptarmigan in Rocky Mountain National Park, Colorado, USA.

PubMed

Wang, Guiming; Hobbs, N Thompson; Galbraith, Hector; Giesen, Kenneth M

2002-09-01

Global climate change may impact wildlife populations by affecting local weather patterns, which, in turn, can impact a variety of ecological processes. However, it is not clear that local variations in ecological processes can be explained by large-scale patterns of climate. The North Atlantic oscillation (NAO) is a large-scale climate phenomenon that has been shown to influence the population dynamics of some animals. Although effects of the NAO on vertebrate population dynamics have been studied, it remains uncertain whether it broadly predicts the impact of weather on species. We examined the ability of local weather data and the NAO to explain the annual variation in population dynamics of white-tailed ptarmigan ( Lagopus leucurus) in Rocky Mountain National Park, USA. We performed canonical correlation analysis on the demographic subspace of ptarmigan and local-climate subspace defined by the empirical orthogonal function (EOF) using data from 1975 to 1999. We found that two subspaces were significantly correlated on the first canonical variable. The Pearson correlation coefficient of the first EOF values of the demographic and local-climate subspaces was significant. The population density and the first EOF of local-climate subspace influenced the ptarmigan population with 1-year lags in the Gompertz model. However, the NAO index was neither related to the first two EOF of local-climate subspace nor to the first EOF of the demographic subspace of ptarmigan. Moreover, the NAO index was not a significant term in the Gompertz model for the ptarmigan population. Therefore, local climate had stronger signature on the demography of ptarmigan than did a large-scale index, i.e., the NAO index. We conclude that local responses of wildlife populations to changing climate may not be adequately explained by models that project large-scale climatic patterns.

Approximate kernel competitive learning.

PubMed

Wu, Jian-Sheng; Zheng, Wei-Shi; Lai, Jian-Huang

2015-03-01

Kernel competitive learning has been successfully used to achieve robust clustering. However, kernel competitive learning (KCL) is not scalable for large scale data processing, because (1) it has to calculate and store the full kernel matrix that is too large to be calculated and kept in the memory and (2) it cannot be computed in parallel. In this paper we develop a framework of approximate kernel competitive learning for processing large scale dataset. The proposed framework consists of two parts. First, it derives an approximate kernel competitive learning (AKCL), which learns kernel competitive learning in a subspace via sampling. We provide solid theoretical analysis on why the proposed approximation modelling would work for kernel competitive learning, and furthermore, we show that the computational complexity of AKCL is largely reduced. Second, we propose a pseudo-parallelled approximate kernel competitive learning (PAKCL) based on a set-based kernel competitive learning strategy, which overcomes the obstacle of using parallel programming in kernel competitive learning and significantly accelerates the approximate kernel competitive learning for large scale clustering. The empirical evaluation on publicly available datasets shows that the proposed AKCL and PAKCL can perform comparably as KCL, with a large reduction on computational cost. Also, the proposed methods achieve more effective clustering performance in terms of clustering precision against related approximate clustering approaches. Copyright © 2014 Elsevier Ltd. All rights reserved.

A learning scheme for reach to grasp movements: on EMG-based interfaces using task specific motion decoding models.

PubMed

Liarokapis, Minas V; Artemiadis, Panagiotis K; Kyriakopoulos, Kostas J; Manolakos, Elias S

2013-09-01

A learning scheme based on random forests is used to discriminate between different reach to grasp movements in 3-D space, based on the myoelectric activity of human muscles of the upper-arm and the forearm. Task specificity for motion decoding is introduced in two different levels: Subspace to move toward and object to be grasped. The discrimination between the different reach to grasp strategies is accomplished with machine learning techniques for classification. The classification decision is then used in order to trigger an EMG-based task-specific motion decoding model. Task specific models manage to outperform "general" models providing better estimation accuracy. Thus, the proposed scheme takes advantage of a framework incorporating both a classifier and a regressor that cooperate advantageously in order to split the task space. The proposed learning scheme can be easily used to a series of EMG-based interfaces that must operate in real time, providing data-driven capabilities for multiclass problems, that occur in everyday life complex environments.

Independence and totalness of subspaces in phase space methods

NASA Astrophysics Data System (ADS)

Vourdas, A.

2018-04-01

The concepts of independence and totalness of subspaces are introduced in the context of quasi-probability distributions in phase space, for quantum systems with finite-dimensional Hilbert space. It is shown that due to the non-distributivity of the lattice of subspaces, there are various levels of independence, from pairwise independence up to (full) independence. Pairwise totalness, totalness and other intermediate concepts are also introduced, which roughly express that the subspaces overlap strongly among themselves, and they cover the full Hilbert space. A duality between independence and totalness, that involves orthocomplementation (logical NOT operation), is discussed. Another approach to independence is also studied, using Rota's formalism on independent partitions of the Hilbert space. This is used to define informational independence, which is proved to be equivalent to independence. As an application, the pentagram (used in discussions on contextuality) is analysed using these concepts.

A fast multigrid-based electromagnetic eigensolver for curved metal boundaries on the Yee mesh

DOE Office of Scientific and Technical Information (OSTI.GOV)

Bauer, Carl A., E-mail: carl.bauer@colorado.edu; Werner, Gregory R.; Cary, John R.

For embedded boundary electromagnetics using the Dey–Mittra (Dey and Mittra, 1997) [1] algorithm, a special grad–div matrix constructed in this work allows use of multigrid methods for efficient inversion of Maxwell’s curl–curl matrix. Efficient curl–curl inversions are demonstrated within a shift-and-invert Krylov-subspace eigensolver (open-sourced at ([ofortt]https://github.com/bauerca/maxwell[cfortt])) on the spherical cavity and the 9-cell TESLA superconducting accelerator cavity. The accuracy of the Dey–Mittra algorithm is also examined: frequencies converge with second-order error, and surface fields are found to converge with nearly second-order error. In agreement with previous work (Nieter et al., 2009) [2], neglecting some boundary-cut cell faces (as is requiredmore » in the time domain for numerical stability) reduces frequency convergence to first-order and surface-field convergence to zeroth-order (i.e. surface fields do not converge). Additionally and importantly, neglecting faces can reduce accuracy by an order of magnitude at low resolutions.« less

Efficient solution of parabolic equations by Krylov approximation methods

NASA Technical Reports Server (NTRS)

Gallopoulos, E.; Saad, Y.

1990-01-01

Numerical techniques for solving parabolic equations by the method of lines is addressed. The main motivation for the proposed approach is the possibility of exploiting a high degree of parallelism in a simple manner. The basic idea of the method is to approximate the action of the evolution operator on a given state vector by means of a projection process onto a Krylov subspace. Thus, the resulting approximation consists of applying an evolution operator of a very small dimension to a known vector which is, in turn, computed accurately by exploiting well-known rational approximations to the exponential. Because the rational approximation is only applied to a small matrix, the only operations required with the original large matrix are matrix-by-vector multiplications, and as a result the algorithm can easily be parallelized and vectorized. Some relevant approximation and stability issues are discussed. We present some numerical experiments with the method and compare its performance with a few explicit and implicit algorithms.

Adaptive classifier for steel strip surface defects

NASA Astrophysics Data System (ADS)

Jiang, Mingming; Li, Guangyao; Xie, Li; Xiao, Mang; Yi, Li

2017-01-01

Surface defects detection system has been receiving increased attention as its precision, speed and less cost. One of the most challenges is reacting to accuracy deterioration with time as aged equipment and changed processes. These variables will make a tiny change to the real world model but a big impact on the classification result. In this paper, we propose a new adaptive classifier with a Bayes kernel (BYEC) which update the model with small sample to it adaptive for accuracy deterioration. Firstly, abundant features were introduced to cover lots of information about the defects. Secondly, we constructed a series of SVMs with the random subspace of the features. Then, a Bayes classifier was trained as an evolutionary kernel to fuse the results from base SVMs. Finally, we proposed the method to update the Bayes evolutionary kernel. The proposed algorithm is experimentally compared with different algorithms, experimental results demonstrate that the proposed method can be updated with small sample and fit the changed model well. Robustness, low requirement for samples and adaptive is presented in the experiment.

Multilevel acceleration of scattering-source iterations with application to electron transport

DOE PAGES

Drumm, Clif; Fan, Wesley

2017-08-18

Acceleration/preconditioning strategies available in the SCEPTRE radiation transport code are described. A flexible transport synthetic acceleration (TSA) algorithm that uses a low-order discrete-ordinates (S N) or spherical-harmonics (P N) solve to accelerate convergence of a high-order S N source-iteration (SI) solve is described. Convergence of the low-order solves can be further accelerated by applying off-the-shelf incomplete-factorization or algebraic-multigrid methods. Also available is an algorithm that uses a generalized minimum residual (GMRES) iterative method rather than SI for convergence, using a parallel sweep-based solver to build up a Krylov subspace. TSA has been applied as a preconditioner to accelerate the convergencemore » of the GMRES iterations. The methods are applied to several problems involving electron transport and problems with artificial cross sections with large scattering ratios. These methods were compared and evaluated by considering material discontinuities and scattering anisotropy. Observed accelerations obtained are highly problem dependent, but speedup factors around 10 have been observed in typical applications.« less

Improved dense trajectories for action recognition based on random projection and Fisher vectors

NASA Astrophysics Data System (ADS)

Ai, Shihui; Lu, Tongwei; Xiong, Yudian

2018-03-01

As an important application of intelligent monitoring system, the action recognition in video has become a very important research area of computer vision. In order to improve the accuracy rate of the action recognition in video with improved dense trajectories, one advanced vector method is introduced. Improved dense trajectories combine Fisher Vector with Random Projection. The method realizes the reduction of the characteristic trajectory though projecting the high-dimensional trajectory descriptor into the low-dimensional subspace based on defining and analyzing Gaussian mixture model by Random Projection. And a GMM-FV hybrid model is introduced to encode the trajectory feature vector and reduce dimension. The computational complexity is reduced by Random Projection which can drop Fisher coding vector. Finally, a Linear SVM is used to classifier to predict labels. We tested the algorithm in UCF101 dataset and KTH dataset. Compared with existed some others algorithm, the result showed that the method not only reduce the computational complexity but also improved the accuracy of action recognition.

Trace gas detection in hyperspectral imagery using the wavelet packet subspace

NASA Astrophysics Data System (ADS)

Salvador, Mark A. Z.

This dissertation describes research into a new remote sensing method to detect trace gases in hyperspectral and ultra-spectral data. This new method is based on the wavelet packet transform. It attempts to improve both the computational tractability and the detection of trace gases in airborne and spaceborne spectral imagery. Atmospheric trace gas research supports various Earth science disciplines to include climatology, vulcanology, pollution monitoring, natural disasters, and intelligence and military applications. Hyperspectral and ultra-spectral data significantly increases the data glut of existing Earth science data sets. Spaceborne spectral data in particular significantly increases spectral resolution while performing daily global collections of the earth. Application of the wavelet packet transform to the spectral space of hyperspectral and ultra-spectral imagery data potentially improves remote sensing detection algorithms. It also facilities the parallelization of these methods for high performance computing. This research seeks two science goals, (1) developing a new spectral imagery detection algorithm, and (2) facilitating the parallelization of trace gas detection in spectral imagery data.

Efficient Grammar Induction Algorithm with Parse Forests from Real Corpora

NASA Astrophysics Data System (ADS)

Kurihara, Kenichi; Kameya, Yoshitaka; Sato, Taisuke

The task of inducing grammar structures has received a great deal of attention. The reasons why researchers have studied are different; to use grammar induction as the first stage in building large treebanks or to make up better language models. However, grammar induction has inherent computational complexity. To overcome it, some grammar induction algorithms add new production rules incrementally. They refine the grammar while keeping their computational complexity low. In this paper, we propose a new efficient grammar induction algorithm. Although our algorithm is similar to algorithms which learn a grammar incrementally, our algorithm uses the graphical EM algorithm instead of the Inside-Outside algorithm. We report results of learning experiments in terms of learning speeds. The results show that our algorithm learns a grammar in constant time regardless of the size of the grammar. Since our algorithm decreases syntactic ambiguities in each step, our algorithm reduces required time for learning. This constant-time learning considerably affects learning time for larger grammars. We also reports results of evaluation of criteria to choose nonterminals. Our algorithm refines a grammar based on a nonterminal in each step. Since there can be several criteria to decide which nonterminal is the best, we evaluate them by learning experiments.

Hybrid Reduced Order Modeling Algorithms for Reactor Physics Calculations

NASA Astrophysics Data System (ADS)

Bang, Youngsuk

Reduced order modeling (ROM) has been recognized as an indispensable approach when the engineering analysis requires many executions of high fidelity simulation codes. Examples of such engineering analyses in nuclear reactor core calculations, representing the focus of this dissertation, include the functionalization of the homogenized few-group cross-sections in terms of the various core conditions, e.g. burn-up, fuel enrichment, temperature, etc. This is done via assembly calculations which are executed many times to generate the required functionalization for use in the downstream core calculations. Other examples are sensitivity analysis used to determine important core attribute variations due to input parameter variations, and uncertainty quantification employed to estimate core attribute uncertainties originating from input parameter uncertainties. ROM constructs a surrogate model with quantifiable accuracy which can replace the original code for subsequent engineering analysis calculations. This is achieved by reducing the effective dimensionality of the input parameter, the state variable, or the output response spaces, by projection onto the so-called active subspaces. Confining the variations to the active subspace allows one to construct an ROM model of reduced complexity which can be solved more efficiently. This dissertation introduces a new algorithm to render reduction with the reduction errors bounded based on a user-defined error tolerance which represents the main challenge of existing ROM techniques. Bounding the error is the key to ensuring that the constructed ROM models are robust for all possible applications. Providing such error bounds represents one of the algorithmic contributions of this dissertation to the ROM state-of-the-art. Recognizing that ROM techniques have been developed to render reduction at different levels, e.g. the input parameter space, the state space, and the response space, this dissertation offers a set of novel hybrid ROM algorithms which can be readily integrated into existing methods and offer higher computational efficiency and defendable accuracy of the reduced models. For example, the snapshots ROM algorithm is hybridized with the range finding algorithm to render reduction in the state space, e.g. the flux in reactor calculations. In another implementation, the perturbation theory used to calculate first order derivatives of responses with respect to parameters is hybridized with a forward sensitivity analysis approach to render reduction in the parameter space. Reduction at the state and parameter spaces can be combined to render further reduction at the interface between different physics codes in a multi-physics model with the accuracy quantified in a similar manner to the single physics case. Although the proposed algorithms are generic in nature, we focus here on radiation transport models used in support of the design and analysis of nuclear reactor cores. In particular, we focus on replacing the traditional assembly calculations by ROM models to facilitate the generation of homogenized cross-sections for downstream core calculations. The implication is that assembly calculations could be done instantaneously therefore precluding the need for the expensive evaluation of the few-group cross-sections for all possible core conditions. Given the generic natures of the algorithms, we make an effort to introduce the material in a general form to allow non-nuclear engineers to benefit from this work.

Efficient model learning methods for actor-critic control.

PubMed

Grondman, Ivo; Vaandrager, Maarten; Buşoniu, Lucian; Babuska, Robert; Schuitema, Erik

2012-06-01

We propose two new actor-critic algorithms for reinforcement learning. Both algorithms use local linear regression (LLR) to learn approximations of the functions involved. A crucial feature of the algorithms is that they also learn a process model, and this, in combination with LLR, provides an efficient policy update for faster learning. The first algorithm uses a novel model-based update rule for the actor parameters. The second algorithm does not use an explicit actor but learns a reference model which represents a desired behavior, from which desired control actions can be calculated using the inverse of the learned process model. The two novel methods and a standard actor-critic algorithm are applied to the pendulum swing-up problem, in which the novel methods achieve faster learning than the standard algorithm.

Constraining a Coastal Ocean Model by Surface Observations Using an Ensemble Kalman Filter

NASA Astrophysics Data System (ADS)

De Mey, P. J.; Ayoub, N. K.

2016-02-01

We explore the impact of assimilating sea surface temperature (SST) and sea surface height (SSH) observations in the Bay of Biscay (North-East Atlantic). The study is conducted in the SYMPHONIE coastal circulation model (Marsaleix et al., 2009) on a 3kmx3km grid, with 43 sigma levels. Ensembles are generated by perturbing the wind forcing to analyze the model error subspace spanned by its response to wind forcing uncertainties. The assimilation method is a 4D Ensemble Kalman Filter algorithm with localization. We use the SDAP code developed in the team (https://sourceforge.net/projects/sequoia-dap/). In a first step before the assimilation of real observations, we set up an Ensemble twin experiment protocol where a nature run as well as noisy pseudo-observations of SST and SSH are generated from an Ensemble member (later discarded from the assimilative Ensemble). Our objectives are to assess (1) the adequacy of the choice of error source and perturbation strategy and (2) how effective the surface observational constraint is at constraining the surface and subsurface fields. We first illustrate characteristics of the error subspace generated by the perturbation strategy. We then show that, while the EnKF solves a single seamless problem regardless of the region within our domain, the nature and effectiveness of the data constraint over the shelf differ from those over the abyssal plain.

Optimal image alignment with random projections of manifolds: algorithm and geometric analysis.

PubMed

Kokiopoulou, Effrosyni; Kressner, Daniel; Frossard, Pascal

2011-06-01

This paper addresses the problem of image alignment based on random measurements. Image alignment consists of estimating the relative transformation between a query image and a reference image. We consider the specific problem where the query image is provided in compressed form in terms of linear measurements captured by a vision sensor. We cast the alignment problem as a manifold distance minimization problem in the linear subspace defined by the measurements. The transformation manifold that represents synthesis of shift, rotation, and isotropic scaling of the reference image can be given in closed form when the reference pattern is sparsely represented over a parametric dictionary. We show that the objective function can then be decomposed as the difference of two convex functions (DC) in the particular case where the dictionary is built on Gaussian functions. Thus, the optimization problem becomes a DC program, which in turn can be solved globally by a cutting plane method. The quality of the solution is typically affected by the number of random measurements and the condition number of the manifold that describes the transformations of the reference image. We show that the curvature, which is closely related to the condition number, remains bounded in our image alignment problem, which means that the relative transformation between two images can be determined optimally in a reduced subspace.

Incremental isometric embedding of high-dimensional data using connected neighborhood graphs.

PubMed

Zhao, Dongfang; Yang, Li

2009-01-01

Most nonlinear data embedding methods use bottom-up approaches for capturing the underlying structure of data distributed on a manifold in high dimensional space. These methods often share the first step which defines neighbor points of every data point by building a connected neighborhood graph so that all data points can be embedded to a single coordinate system. These methods are required to work incrementally for dimensionality reduction in many applications. Because input data stream may be under-sampled or skewed from time to time, building connected neighborhood graph is crucial to the success of incremental data embedding using these methods. This paper presents algorithms for updating $k$-edge-connected and $k$-connected neighborhood graphs after a new data point is added or an old data point is deleted. It further utilizes a simple algorithm for updating all-pair shortest distances on the neighborhood graph. Together with incremental classical multidimensional scaling using iterative subspace approximation, this paper devises an incremental version of Isomap with enhancements to deal with under-sampled or unevenly distributed data. Experiments on both synthetic and real-world data sets show that the algorithm is efficient and maintains low dimensional configurations of high dimensional data under various data distributions.

Non-linear eigensolver-based alternative to traditional SCF methods

NASA Astrophysics Data System (ADS)

Gavin, Brendan; Polizzi, Eric

2013-03-01

The self-consistent iterative procedure in Density Functional Theory calculations is revisited using a new, highly efficient and robust algorithm for solving the non-linear eigenvector problem (i.e. H(X)X = EX;) of the Kohn-Sham equations. This new scheme is derived from a generalization of the FEAST eigenvalue algorithm, and provides a fundamental and practical numerical solution for addressing the non-linearity of the Hamiltonian with the occupied eigenvectors. In contrast to SCF techniques, the traditional outer iterations are replaced by subspace iterations that are intrinsic to the FEAST algorithm, while the non-linearity is handled at the level of a projected reduced system which is orders of magnitude smaller than the original one. Using a series of numerical examples, it will be shown that our approach can outperform the traditional SCF mixing techniques such as Pulay-DIIS by providing a high converge rate and by converging to the correct solution regardless of the choice of the initial guess. We also discuss a practical implementation of the technique that can be achieved effectively using the FEAST solver package. This research is supported by NSF under Grant #ECCS-0846457 and Intel Corporation.

Automatic Estimation of Osteoporotic Fracture Cases by Using Ensemble Learning Approaches.

PubMed

Kilic, Niyazi; Hosgormez, Erkan

2016-03-01

Ensemble learning methods are one of the most powerful tools for the pattern classification problems. In this paper, the effects of ensemble learning methods and some physical bone densitometry parameters on osteoporotic fracture detection were investigated. Six feature set models were constructed including different physical parameters and they fed into the ensemble classifiers as input features. As ensemble learning techniques, bagging, gradient boosting and random subspace (RSM) were used. Instance based learning (IBk) and random forest (RF) classifiers applied to six feature set models. The patients were classified into three groups such as osteoporosis, osteopenia and control (healthy), using ensemble classifiers. Total classification accuracy and f-measure were also used to evaluate diagnostic performance of the proposed ensemble classification system. The classification accuracy has reached to 98.85 % by the combination of model 6 (five BMD + five T-score values) using RSM-RF classifier. The findings of this paper suggest that the patients will be able to be warned before a bone fracture occurred, by just examining some physical parameters that can easily be measured without invasive operations.

Gravitational instantons, self-duality, and geometric flows

DOE Office of Scientific and Technical Information (OSTI.GOV)

Bourliot, F.; Estes, J.; Petropoulos, P. M.

2010-05-15

We discuss four-dimensional 'spatially homogeneous' gravitational instantons. These are self-dual solutions of Euclidean vacuum Einstein equations. They are endowed with a product structure RxM{sub 3} leading to a foliation into three-dimensional subspaces evolving in Euclidean time. For a large class of homogeneous subspaces, the dynamics coincides with a geometric flow on the three-dimensional slice, driven by the Ricci tensor plus an so(3) gauge connection. The flowing metric is related to the vielbein of the subspace, while the gauge field is inherited from the anti-self-dual component of the four-dimensional Levi-Civita connection.

On the Convergence of an Implicitly Restarted Arnoldi Method

DOE Office of Scientific and Technical Information (OSTI.GOV)

Lehoucq, Richard B.

We show that Sorensen's [35] implicitly restarted Arnoldi method (including its block extension) is simultaneous iteration with an implicit projection step to accelerate convergence to the invariant subspace of interest. By using the geometric convergence theory for simultaneous iteration due to Watkins and Elsner [43], we prove that an implicitly restarted Arnoldi method can achieve a super-linear rate of convergence to the dominant invariant subspace of a matrix. Moreover, we show how an IRAM computes a nested sequence of approximations for the partial Schur decomposition associated with the dominant invariant subspace of a matrix.

Portfolios and the market geometry

NASA Astrophysics Data System (ADS)

Eleutério, Samuel; Araújo, Tanya; Vilela Mendes, R.

2014-09-01

A geometric analysis of return time series, performed in the past, implied that most of the systematic information in the market is contained in a space of small dimension. Here we have explored subspaces of this space to find out the relative performance of portfolios formed from companies that have the largest projections in each one of the subspaces. As expected, it was found that the best performance portfolios are associated with some of the small eigenvalue subspaces and not to the dominant dimensions. This is found to occur in a systematic fashion over an extended period (1990-2008).

Integrated head package cable carrier for a nuclear power plant

DOEpatents

Meuschke, Robert E.; Trombola, Daniel M.

1995-01-01

A cabling arrangement is provided for a nuclear reactor located within a containment. Structure inside the containment is characterized by a wall having a near side surrounding the reactor vessel defining a cavity, an operating deck outside the cavity, a sub-space below the deck and on a far side of the wall spaced from the near side, and an operating area above the deck. The arrangement includes a movable frame supporting a plurality of cables extending through the frame, each connectable at a first end to a head package on the reactor vessel and each having a second end located in the sub-space. The frame is movable, with the cables, between a first position during normal operation of the reactor when the cables are connected to the head package, located outside the sub-space proximate the head package, and a second position during refueling when the cables are disconnected from the head package, located in the sub-space. In a preferred embodiment, the frame straddles the top of the wall in a substantially horizontal orientation in the first position, pivots about an end distal from the head package to a substantially vertically oriented intermediate position, and is guided, while remaining about vertically oriented, along a track in the sub-space to the second position.

A robust variant of block Jacobi-Davidson for extracting a large number of eigenpairs: Application to grid-based real-space density functional theory

NASA Astrophysics Data System (ADS)

Lee, M.; Leiter, K.; Eisner, C.; Breuer, A.; Wang, X.

2017-09-01

In this work, we investigate a block Jacobi-Davidson (J-D) variant suitable for sparse symmetric eigenproblems where a substantial number of extremal eigenvalues are desired (e.g., ground-state real-space quantum chemistry). Most J-D algorithm variations tend to slow down as the number of desired eigenpairs increases due to frequent orthogonalization against a growing list of solved eigenvectors. In our specification of block J-D, all of the steps of the algorithm are performed in clusters, including the linear solves, which allows us to greatly reduce computational effort with blocked matrix-vector multiplies. In addition, we move orthogonalization against locked eigenvectors and working eigenvectors outside of the inner loop but retain the single Ritz vector projection corresponding to the index of the correction vector. Furthermore, we minimize the computational effort by constraining the working subspace to the current vectors being updated and the latest set of corresponding correction vectors. Finally, we incorporate accuracy thresholds based on the precision required by the Fermi-Dirac distribution. The net result is a significant reduction in the computational effort against most previous block J-D implementations, especially as the number of wanted eigenpairs grows. We compare our approach with another robust implementation of block J-D (JDQMR) and the state-of-the-art Chebyshev filter subspace (CheFSI) method for various real-space density functional theory systems. Versus CheFSI, for first-row elements, our method yields competitive timings for valence-only systems and 4-6× speedups for all-electron systems with up to 10× reduced matrix-vector multiplies. For all-electron calculations on larger elements (e.g., gold) where the wanted spectrum is quite narrow compared to the full spectrum, we observe 60× speedup with 200× fewer matrix-vector multiples vs. CheFSI.

A robust variant of block Jacobi-Davidson for extracting a large number of eigenpairs: Application to grid-based real-space density functional theory.

PubMed

Lee, M; Leiter, K; Eisner, C; Breuer, A; Wang, X

2017-09-21

In this work, we investigate a block Jacobi-Davidson (J-D) variant suitable for sparse symmetric eigenproblems where a substantial number of extremal eigenvalues are desired (e.g., ground-state real-space quantum chemistry). Most J-D algorithm variations tend to slow down as the number of desired eigenpairs increases due to frequent orthogonalization against a growing list of solved eigenvectors. In our specification of block J-D, all of the steps of the algorithm are performed in clusters, including the linear solves, which allows us to greatly reduce computational effort with blocked matrix-vector multiplies. In addition, we move orthogonalization against locked eigenvectors and working eigenvectors outside of the inner loop but retain the single Ritz vector projection corresponding to the index of the correction vector. Furthermore, we minimize the computational effort by constraining the working subspace to the current vectors being updated and the latest set of corresponding correction vectors. Finally, we incorporate accuracy thresholds based on the precision required by the Fermi-Dirac distribution. The net result is a significant reduction in the computational effort against most previous block J-D implementations, especially as the number of wanted eigenpairs grows. We compare our approach with another robust implementation of block J-D (JDQMR) and the state-of-the-art Chebyshev filter subspace (CheFSI) method for various real-space density functional theory systems. Versus CheFSI, for first-row elements, our method yields competitive timings for valence-only systems and 4-6× speedups for all-electron systems with up to 10× reduced matrix-vector multiplies. For all-electron calculations on larger elements (e.g., gold) where the wanted spectrum is quite narrow compared to the full spectrum, we observe 60× speedup with 200× fewer matrix-vector multiples vs. CheFSI.

SemiBoost: boosting for semi-supervised learning.

PubMed

Mallapragada, Pavan Kumar; Jin, Rong; Jain, Anil K; Liu, Yi

2009-11-01

Semi-supervised learning has attracted a significant amount of attention in pattern recognition and machine learning. Most previous studies have focused on designing special algorithms to effectively exploit the unlabeled data in conjunction with labeled data. Our goal is to improve the classification accuracy of any given supervised learning algorithm by using the available unlabeled examples. We call this as the Semi-supervised improvement problem, to distinguish the proposed approach from the existing approaches. We design a metasemi-supervised learning algorithm that wraps around the underlying supervised algorithm and improves its performance using unlabeled data. This problem is particularly important when we need to train a supervised learning algorithm with a limited number of labeled examples and a multitude of unlabeled examples. We present a boosting framework for semi-supervised learning, termed as SemiBoost. The key advantages of the proposed semi-supervised learning approach are: 1) performance improvement of any supervised learning algorithm with a multitude of unlabeled data, 2) efficient computation by the iterative boosting algorithm, and 3) exploiting both manifold and cluster assumption in training classification models. An empirical study on 16 different data sets and text categorization demonstrates that the proposed framework improves the performance of several commonly used supervised learning algorithms, given a large number of unlabeled examples. We also show that the performance of the proposed algorithm, SemiBoost, is comparable to the state-of-the-art semi-supervised learning algorithms.

Time delay estimation using new spectral and adaptive filtering methods with applications to underwater target detection

NASA Astrophysics Data System (ADS)

Hasan, Mohammed A.

1997-11-01

In this dissertation, we present several novel approaches for detection and identification of targets of arbitrary shapes from the acoustic backscattered data and using the incident waveform. This problem is formulated as time- delay estimation and sinusoidal frequency estimation problems which both have applications in many other important areas in signal processing. Solving time-delay estimation problem allows the identification of the specular components in the backscattered signal from elastic and non-elastic targets. Thus, accurate estimation of these time delays would help in determining the existence of certain clues for detecting targets. Several new methods for solving these two problems in the time, frequency and wavelet domains are developed. In the time domain, a new block fast transversal filter (BFTF) is proposed for a fast implementation of the least squares (LS) method. This BFTF algorithm is derived by using data-related constrained block-LS cost function to guarantee global optimality. The new soft-constrained algorithm provides an efficient way of transferring weight information between blocks of data and thus it is computationally very efficient compared with other LS- based schemes. Additionally, the tracking ability of the algorithm can be controlled by varying the block length and/or a soft constrained parameter. The effectiveness of this algorithm is tested on several underwater acoustic backscattered data for elastic targets and non-elastic (cement chunk) objects. In the frequency domain, the time-delay estimation problem is converted to a sinusoidal frequency estimation problem by using the discrete Fourier transform. Then, the lagged sample covariance matrices of the resulting signal are computed and studied in terms of their eigen- structure. These matrices are shown to be robust and effective in extracting bases for the signal and noise subspaces. New MUSIC and matrix pencil-based methods are derived these subspaces. The effectiveness of the method is demonstrated on the problem of detection of multiple specular components in the acoustic backscattered data. Finally, a method for the estimation of time delays using wavelet decomposition is derived. The sub-band adaptive filtering uses discrete wavelet transform for multi- resolution or sub-band decomposition. Joint time delay estimation for identifying multi-specular components and subsequent adaptive filtering processes are performed on the signal in each sub-band. This would provide multiple 'look' of the signal at different resolution scale which results in more accurate estimates for delays associated with the specular components. Simulation results on the simulated and real shallow water data are provided which show the promise of this new scheme for target detection in a heavy cluttered environment.

SNP selection and classification of genome-wide SNP data using stratified sampling random forests.

PubMed

Wu, Qingyao; Ye, Yunming; Liu, Yang; Ng, Michael K

2012-09-01

For high dimensional genome-wide association (GWA) case-control data of complex disease, there are usually a large portion of single-nucleotide polymorphisms (SNPs) that are irrelevant with the disease. A simple random sampling method in random forest using default mtry parameter to choose feature subspace, will select too many subspaces without informative SNPs. Exhaustive searching an optimal mtry is often required in order to include useful and relevant SNPs and get rid of vast of non-informative SNPs. However, it is too time-consuming and not favorable in GWA for high-dimensional data. The main aim of this paper is to propose a stratified sampling method for feature subspace selection to generate decision trees in a random forest for GWA high-dimensional data. Our idea is to design an equal-width discretization scheme for informativeness to divide SNPs into multiple groups. In feature subspace selection, we randomly select the same number of SNPs from each group and combine them to form a subspace to generate a decision tree. The advantage of this stratified sampling procedure can make sure each subspace contains enough useful SNPs, but can avoid a very high computational cost of exhaustive search of an optimal mtry, and maintain the randomness of a random forest. We employ two genome-wide SNP data sets (Parkinson case-control data comprised of 408 803 SNPs and Alzheimer case-control data comprised of 380 157 SNPs) to demonstrate that the proposed stratified sampling method is effective, and it can generate better random forest with higher accuracy and lower error bound than those by Breiman's random forest generation method. For Parkinson data, we also show some interesting genes identified by the method, which may be associated with neurological disorders for further biological investigations.

Noise-enhanced clustering and competitive learning algorithms.

PubMed

Osoba, Osonde; Kosko, Bart

2013-01-01

Noise can provably speed up convergence in many centroid-based clustering algorithms. This includes the popular k-means clustering algorithm. The clustering noise benefit follows from the general noise benefit for the expectation-maximization algorithm because many clustering algorithms are special cases of the expectation-maximization algorithm. Simulations show that noise also speeds up convergence in stochastic unsupervised competitive learning, supervised competitive learning, and differential competitive learning. Copyright © 2012 Elsevier Ltd. All rights reserved.

Q-Learning-Based Adjustable Fixed-Phase Quantum Grover Search Algorithm

NASA Astrophysics Data System (ADS)

Guo, Ying; Shi, Wensha; Wang, Yijun; Hu, Jiankun

2017-02-01

We demonstrate that the rotation phase can be suitably chosen to increase the efficiency of the phase-based quantum search algorithm, leading to a dynamic balance between iterations and success probabilities of the fixed-phase quantum Grover search algorithm with Q-learning for a given number of solutions. In this search algorithm, the proposed Q-learning algorithm, which is a model-free reinforcement learning strategy in essence, is used for performing a matching algorithm based on the fraction of marked items λ and the rotation phase α. After establishing the policy function α = π(λ), we complete the fixed-phase Grover algorithm, where the phase parameter is selected via the learned policy. Simulation results show that the Q-learning-based Grover search algorithm (QLGA) enables fewer iterations and gives birth to higher success probabilities. Compared with the conventional Grover algorithms, it avoids the optimal local situations, thereby enabling success probabilities to approach one.

Modification Of Learning Rate With Lvq Model Improvement In Learning Backpropagation

NASA Astrophysics Data System (ADS)

Tata Hardinata, Jaya; Zarlis, Muhammad; Budhiarti Nababan, Erna; Hartama, Dedy; Sembiring, Rahmat W.

2017-12-01

One type of artificial neural network is a backpropagation, This algorithm trained with the network architecture used during the training as well as providing the correct output to insert a similar but not the same with the architecture in use at training.The selection of appropriate parameters also affects the outcome, value of learning rate is one of the parameters which influence the process of training, Learning rate affects the speed of learning process on the network architecture.If the learning rate is set too large, then the algorithm will become unstable and otherwise the algorithm will converge in a very long period of time.So this study was made to determine the value of learning rate on the backpropagation algorithm. LVQ models of learning rate is one of the models used in the determination of the value of the learning rate of the algorithm LVQ.By modifying this LVQ model to be applied to the backpropagation algorithm. From the experimental results known to modify the learning rate LVQ models were applied to the backpropagation algorithm learning process becomes faster (epoch less).

Experimental state control by fast non-Abelian holonomic gates with a superconducting qutrit

NASA Astrophysics Data System (ADS)

Danilin, S.; Vepsäläinen, A.; Paraoanu, G. S.

2018-05-01

Quantum state manipulation with gates based on geometric phases acquired during cyclic operations promises inherent fault-tolerance and resilience to local fluctuations in the control parameters. Here we create a general non-Abelian and non-adiabatic holonomic gate acting in the (∣0〉, ∣2〉) subspace of a three-level (qutrit) transmon device fabricated in a fully coplanar design. Experimentally, this is realized by simultaneously coupling the first two transitions by microwave pulses with amplitudes and phases defined such that the condition of parallel transport is fulfilled. We demonstrate the creation of arbitrary superpositions in this subspace by changing the amplitudes of the pulses and the relative phase between them. We use two-photon pulses acting in the holonomic subspace to reveal the coherence of the state created by the geometric gate pulses and to prepare different superposition states. We also test the action of holonomic NOT and Hadamard gates on superpositions in the (| 0> ,| 2> ) subspace.

New subspace methods for ATR

NASA Astrophysics Data System (ADS)

Zhang, Peng; Peng, Jing; Sims, S. Richard F.

2005-05-01

In ATR applications, each feature is a convolution of an image with a filter. It is important to use most discriminant features to produce compact representations. We propose two novel subspace methods for dimension reduction to address limitations associated with Fukunaga-Koontz Transform (FKT). The first method, Scatter-FKT, assumes that target is more homogeneous, while clutter can be anything other than target and anywhere. Thus, instead of estimating a clutter covariance matrix, Scatter-FKT computes a clutter scatter matrix that measures the spread of clutter from the target mean. We choose dimensions along which the difference in variation between target and clutter is most pronounced. When the target follows a Gaussian distribution, Scatter-FKT can be viewed as a generalization of FKT. The second method, Optimal Bayesian Subspace, is derived from the optimal Bayesian classifier. It selects dimensions such that the minimum Bayes error rate can be achieved. When both target and clutter follow Gaussian distributions, OBS computes optimal subspace representations. We compare our methods against FKT using character image as well as IR data.

Improvement in the Accuracy of Matching by Different Feature Subspaces in Traffic Sign Recognition

NASA Astrophysics Data System (ADS)

Ihara, Arihito; Fujiyoshi, Hironobu; Takaki, Masanari; Kumon, Hiroaki; Tamatsu, Yukimasa

A technique for recognizing traffic signs from an image taken with an in-vehicle camera has already been proposed as driver's drive assist. SIFT feature is used for traffic sign recognition, because it is robust to changes in scaling and rotating of the traffic sign. However, it is difficult to process in real-time because the computation cost of the SIFT feature extraction and matching is expensive. This paper presents a method of traffic sign recognition based on keypoint classifier by AdaBoost using PCA-SIFT features in different feature subspaces. Each subspace is constructed from gradients of traffic sign images and general images respectively. A detected keypoint is projected to both subspaces, and then the AdaBoost employs to classy into whether the keypoint is on the traffic sign or not. Experimental results show that the computation cost for keypoint matching can be reduced to about 1/2 compared with the conventional method.

Metastable decoherence-free subspaces and electromagnetically induced transparency in interacting many-body systems

NASA Astrophysics Data System (ADS)

Macieszczak, Katarzyna; Zhou, YanLi; Hofferberth, Sebastian; Garrahan, Juan P.; Li, Weibin; Lesanovsky, Igor

2017-10-01

We investigate the dynamics of a generic interacting many-body system under conditions of electromagnetically induced transparency (EIT). This problem is of current relevance due to its connection to nonlinear optical media realized by Rydberg atoms. In an interacting system the structure of the dynamics and the approach to the stationary state becomes far more complex than in the case of conventional EIT. In particular, we discuss the emergence of a metastable decoherence-free subspace, whose dimension for a single Rydberg excitation grows linearly in the number of atoms. On approach to stationarity this leads to a slow dynamics, which renders the typical assumption of fast relaxation invalid. We derive analytically the effective nonequilibrium dynamics in the decoherence-free subspace, which features coherent and dissipative two-body interactions. We discuss the use of this scenario for the preparation of collective entangled dark states and the realization of general unitary dynamics within the spin-wave subspace.

Krylov Subspace Methods for Complex Non-Hermitian Linear Systems. Thesis

NASA Technical Reports Server (NTRS)

Freund, Roland W.

1991-01-01

We consider Krylov subspace methods for the solution of large sparse linear systems Ax = b with complex non-Hermitian coefficient matrices. Such linear systems arise in important applications, such as inverse scattering, numerical solution of time-dependent Schrodinger equations, underwater acoustics, eddy current computations, numerical computations in quantum chromodynamics, and numerical conformal mapping. Typically, the resulting coefficient matrices A exhibit special structures, such as complex symmetry, or they are shifted Hermitian matrices. In this paper, we first describe a Krylov subspace approach with iterates defined by a quasi-minimal residual property, the QMR method, for solving general complex non-Hermitian linear systems. Then, we study special Krylov subspace methods designed for the two families of complex symmetric respectively shifted Hermitian linear systems. We also include some results concerning the obvious approach to general complex linear systems by solving equivalent real linear systems for the real and imaginary parts of x. Finally, numerical experiments for linear systems arising from the complex Helmholtz equation are reported.

Assessment of time-dependent density functional theory with the restricted excitation space approximation for excited state calculations of large systems

NASA Astrophysics Data System (ADS)

Hanson-Heine, Magnus W. D.; George, Michael W.; Besley, Nicholas A.

2018-06-01

The restricted excitation subspace approximation is explored as a basis to reduce the memory storage required in linear response time-dependent density functional theory (TDDFT) calculations within the Tamm-Dancoff approximation. It is shown that excluding the core orbitals and up to 70% of the virtual orbitals in the construction of the excitation subspace does not result in significant changes in computed UV/vis spectra for large molecules. The reduced size of the excitation subspace greatly reduces the size of the subspace vectors that need to be stored when using the Davidson procedure to determine the eigenvalues of the TDDFT equations. Furthermore, additional screening of the two-electron integrals in combination with a reduction in the size of the numerical integration grid used in the TDDFT calculation leads to significant computational savings. The use of these approximations represents a simple approach to extend TDDFT to the study of large systems and make the calculations increasingly tractable using modest computing resources.

Partially supervised speaker clustering.

PubMed

Tang, Hao; Chu, Stephen Mingyu; Hasegawa-Johnson, Mark; Huang, Thomas S

2012-05-01

Content-based multimedia indexing, retrieval, and processing as well as multimedia databases demand the structuring of the media content (image, audio, video, text, etc.), one significant goal being to associate the identity of the content to the individual segments of the signals. In this paper, we specifically address the problem of speaker clustering, the task of assigning every speech utterance in an audio stream to its speaker. We offer a complete treatment to the idea of partially supervised speaker clustering, which refers to the use of our prior knowledge of speakers in general to assist the unsupervised speaker clustering process. By means of an independent training data set, we encode the prior knowledge at the various stages of the speaker clustering pipeline via 1) learning a speaker-discriminative acoustic feature transformation, 2) learning a universal speaker prior model, and 3) learning a discriminative speaker subspace, or equivalently, a speaker-discriminative distance metric. We study the directional scattering property of the Gaussian mixture model (GMM) mean supervector representation of utterances in the high-dimensional space, and advocate exploiting this property by using the cosine distance metric instead of the euclidean distance metric for speaker clustering in the GMM mean supervector space. We propose to perform discriminant analysis based on the cosine distance metric, which leads to a novel distance metric learning algorithm—linear spherical discriminant analysis (LSDA). We show that the proposed LSDA formulation can be systematically solved within the elegant graph embedding general dimensionality reduction framework. Our speaker clustering experiments on the GALE database clearly indicate that 1) our speaker clustering methods based on the GMM mean supervector representation and vector-based distance metrics outperform traditional speaker clustering methods based on the “bag of acoustic features” representation and statistical model-based distance metrics, 2) our advocated use of the cosine distance metric yields consistent increases in the speaker clustering performance as compared to the commonly used euclidean distance metric, 3) our partially supervised speaker clustering concept and strategies significantly improve the speaker clustering performance over the baselines, and 4) our proposed LSDA algorithm further leads to state-of-the-art speaker clustering performance.

Predicting the survival of diabetes using neural network

NASA Astrophysics Data System (ADS)

Mamuda, Mamman; Sathasivam, Saratha

2017-08-01

Data mining techniques at the present time are used in predicting diseases of health care industries. Neural Network is one among the prevailing method in data mining techniques of an intelligent field for predicting diseases in health care industries. This paper presents a study on the prediction of the survival of diabetes diseases using different learning algorithms from the supervised learning algorithms of neural network. Three learning algorithms are considered in this study: (i) The levenberg-marquardt learning algorithm (ii) The Bayesian regulation learning algorithm and (iii) The scaled conjugate gradient learning algorithm. The network is trained using the Pima Indian Diabetes Dataset with the help of MATLAB R2014(a) software. The performance of each algorithm is further discussed through regression analysis. The prediction accuracy of the best algorithm is further computed to validate the accurate prediction

Multi-frequency subspace migration for imaging of perfectly conducting, arc-like cracks in full- and limited-view inverse scattering problems

NASA Astrophysics Data System (ADS)

Park, Won-Kwang

2015-02-01

Multi-frequency subspace migration imaging techniques are usually adopted for the non-iterative imaging of unknown electromagnetic targets, such as cracks in concrete walls or bridges and anti-personnel mines in the ground, in the inverse scattering problems. It is confirmed that this technique is very fast, effective, robust, and can not only be applied to full- but also to limited-view inverse problems if a suitable number of incidents and corresponding scattered fields are applied and collected. However, in many works, the application of such techniques is heuristic. With the motivation of such heuristic application, this study analyzes the structure of the imaging functional employed in the subspace migration imaging technique in two-dimensional full- and limited-view inverse scattering problems when the unknown targets are arbitrary-shaped, arc-like perfectly conducting cracks located in the two-dimensional homogeneous space. In contrast to the statistical approach based on statistical hypothesis testing, our approach is based on the fact that the subspace migration imaging functional can be expressed by a linear combination of the Bessel functions of integer order of the first kind. This is based on the structure of the Multi-Static Response (MSR) matrix collected in the far-field at nonzero frequency in either Transverse Magnetic (TM) mode (Dirichlet boundary condition) or Transverse Electric (TE) mode (Neumann boundary condition). The investigation of the expression of imaging functionals gives us certain properties of subspace migration and explains why multi-frequency enhances imaging resolution. In particular, we carefully analyze the subspace migration and confirm some properties of imaging when a small number of incident fields are applied. Consequently, we introduce a weighted multi-frequency imaging functional and confirm that it is an improved version of subspace migration in TM mode. Various results of numerical simulations performed on the far-field data affected by large amounts of random noise are similar to the analytical results derived in this study, and they provide a direction for future studies.

Molecular activity prediction by means of supervised subspace projection based ensembles of classifiers.

PubMed

Cerruela García, G; García-Pedrajas, N; Luque Ruiz, I; Gómez-Nieto, M Á

2018-03-01

This paper proposes a method for molecular activity prediction in QSAR studies using ensembles of classifiers constructed by means of two supervised subspace projection methods, namely nonparametric discriminant analysis (NDA) and hybrid discriminant analysis (HDA). We studied the performance of the proposed ensembles compared to classical ensemble methods using four molecular datasets and eight different models for the representation of the molecular structure. Using several measures and statistical tests for classifier comparison, we observe that our proposal improves the classification results with respect to classical ensemble methods. Therefore, we show that ensembles constructed using supervised subspace projections offer an effective way of creating classifiers in cheminformatics.

On spectral synthesis on zero-dimensional Abelian groups

NASA Astrophysics Data System (ADS)

Platonov, S. S.

2013-09-01

Let G be a zero-dimensional locally compact Abelian group all of whose elements are compact, and let C(G) be the space of all complex-valued continuous functions on G. A closed linear subspace \\mathscr H\\subseteq C(G) is said to be an invariant subspace if it is invariant with respect to the translations \\tau_y\\colon f(x)\\mapsto f(x+y), y\\in G. In the paper, it is proved that any invariant subspace \\mathscr H admits spectral synthesis, that is, \\mathscr H coincides with the closed linear span of the characters of G belonging to \\mathscr H. Bibliography: 25 titles.

Conditioned invariant subspaces, disturbance decoupling and solutions of rational matrix equations

NASA Technical Reports Server (NTRS)

Li, Z.; Sastry, S. S.

1986-01-01

Conditioned invariant subspaces are introduced both in terms of output injection and in terms of state estimation. Various properties of these subspaces are explored and the problem of disturbance decoupling by output injection (OIP) is defined. It is then shown that OIP is equivalent to the problem of disturbance decoupled estimation as introduced in Willems (1982) and Willems and Commault (1980). Both solvability conditions and a description of solutions for a class of rational matrix equations of the form X(s)M(s) = Q(s) on several ways are given in state-space form. Finally, the problem of output stabilization with respect to a disturbance is briefly addressed.

Essential uncontrollability of discrete linear, time-invariant, dynamical systems

NASA Technical Reports Server (NTRS)

Cliff, E. M.

1975-01-01

The concept of a 'best approximating m-dimensional subspace' for a given set of vectors in n-dimensional whole space is introduced. Such a subspace is easily described in terms of the eigenvectors of an associated Gram matrix. This technique is used to approximate an achievable set for a discrete linear time-invariant dynamical system. This approximation characterizes the part of the state space that may be reached using modest levels of control. If the achievable set can be closely approximated by a proper subspace of the whole space then the system is 'essentially uncontrollable'. The notion finds application in studies of failure-tolerant systems, and in decoupling.

pyCTQW: A continuous-time quantum walk simulator on distributed memory computers

NASA Astrophysics Data System (ADS)

Izaac, Josh A.; Wang, Jingbo B.

2015-01-01

In the general field of quantum information and computation, quantum walks are playing an increasingly important role in constructing physical models and quantum algorithms. We have recently developed a distributed memory software package pyCTQW, with an object-oriented Python interface, that allows efficient simulation of large multi-particle CTQW (continuous-time quantum walk)-based systems. In this paper, we present an introduction to the Python and Fortran interfaces of pyCTQW, discuss various numerical methods of calculating the matrix exponential, and demonstrate the performance behavior of pyCTQW on a distributed memory cluster. In particular, the Chebyshev and Krylov-subspace methods for calculating the quantum walk propagation are provided, as well as methods for visualization and data analysis.

A computationally tractable version of the collective model

NASA Astrophysics Data System (ADS)

Rowe, D. J.

2004-05-01

A computationally tractable version of the Bohr-Mottelson collective model is presented which makes it possible to diagonalize realistic collective models and obtain convergent results in relatively small appropriately chosen subspaces of the collective model Hilbert space. Special features of the proposed model are that it makes use of the beta wave functions given analytically by the softened-beta version of the Wilets-Jean model, proposed by Elliott et al., and a simple algorithm for computing SO(5)⊃SO(3) spherical harmonics. The latter has much in common with the methods of Chacon, Moshinsky, and Sharp but is conceptually and computationally simpler. Results are presented for collective models ranging from the spherical vibrator to the Wilets-Jean and axially symmetric rotor-vibrator models.

On optimal improvements of classical iterative schemes for Z-matrices

NASA Astrophysics Data System (ADS)

Noutsos, D.; Tzoumas, M.

2006-04-01

Many researchers have considered preconditioners, applied to linear systems, whose matrix coefficient is a Z- or an M-matrix, that make the associated Jacobi and Gauss-Seidel methods converge asymptotically faster than the unpreconditioned ones. Such preconditioners are chosen so that they eliminate the off-diagonal elements of the same column or the elements of the first upper diagonal [Milaszewicz, LAA 93 (1987) 161-170], Gunawardena et al. [LAA 154-156 (1991) 123-143]. In this work we generalize the previous preconditioners to obtain optimal methods. "Good" Jacobi and Gauss-Seidel algorithms are given and preconditioners, that eliminate more than one entry per row, are also proposed and analyzed. Moreover, the behavior of the above preconditioners to the Krylov subspace methods is studied.

An Educational System for Learning Search Algorithms and Automatically Assessing Student Performance

ERIC Educational Resources Information Center

Grivokostopoulou, Foteini; Perikos, Isidoros; Hatzilygeroudis, Ioannis

2017-01-01

In this paper, first we present an educational system that assists students in learning and tutors in teaching search algorithms, an artificial intelligence topic. Learning is achieved through a wide range of learning activities. Algorithm visualizations demonstrate the operational functionality of algorithms according to the principles of active…

Using an improved association rules mining optimization algorithm in web-based mobile-learning system

NASA Astrophysics Data System (ADS)

Huang, Yin; Chen, Jianhua; Xiong, Shaojun

2009-07-01

Mobile-Learning (M-learning) makes many learners get the advantages of both traditional learning and E-learning. Currently, Web-based Mobile-Learning Systems have created many new ways and defined new relationships between educators and learners. Association rule mining is one of the most important fields in data mining and knowledge discovery in databases. Rules explosion is a serious problem which causes great concerns, as conventional mining algorithms often produce too many rules for decision makers to digest. Since Web-based Mobile-Learning System collects vast amounts of student profile data, data mining and knowledge discovery techniques can be applied to find interesting relationships between attributes of learners, assessments, the solution strategies adopted by learners and so on. Therefore ,this paper focus on a new data-mining algorithm, combined with the advantages of genetic algorithm and simulated annealing algorithm , called ARGSA(Association rules based on an improved Genetic Simulated Annealing Algorithm), to mine the association rules. This paper first takes advantage of the Parallel Genetic Algorithm and Simulated Algorithm designed specifically for discovering association rules. Moreover, the analysis and experiment are also made to show the proposed method is superior to the Apriori algorithm in this Mobile-Learning system.

Non-heuristic automatic techniques for overcoming low signal-to-noise-ratio bias of localization microscopy and multiple signal classification algorithm.

PubMed

Agarwal, Krishna; Macháň, Radek; Prasad, Dilip K

2018-03-21

Localization microscopy and multiple signal classification algorithm use temporal stack of image frames of sparse emissions from fluorophores to provide super-resolution images. Localization microscopy localizes emissions in each image independently and later collates the localizations in all the frames, giving same weight to each frame irrespective of its signal-to-noise ratio. This results in a bias towards frames with low signal-to-noise ratio and causes cluttered background in the super-resolved image. User-defined heuristic computational filters are employed to remove a set of localizations in an attempt to overcome this bias. Multiple signal classification performs eigen-decomposition of the entire stack, irrespective of the relative signal-to-noise ratios of the frames, and uses a threshold to classify eigenimages into signal and null subspaces. This results in under-representation of frames with low signal-to-noise ratio in the signal space and over-representation in the null space. Thus, multiple signal classification algorithms is biased against frames with low signal-to-noise ratio resulting into suppression of the corresponding fluorophores. This paper presents techniques to automatically debias localization microscopy and multiple signal classification algorithm of these biases without compromising their resolution and without employing heuristics, user-defined criteria. The effect of debiasing is demonstrated through five datasets of invitro and fixed cell samples.

Validation of Point Clouds Segmentation Algorithms Through Their Application to Several Case Studies for Indoor Building Modelling

NASA Astrophysics Data System (ADS)

Macher, H.; Landes, T.; Grussenmeyer, P.

2016-06-01

Laser scanners are widely used for the modelling of existing buildings and particularly in the creation process of as-built BIM (Building Information Modelling). However, the generation of as-built BIM from point clouds involves mainly manual steps and it is consequently time consuming and error-prone. Along the path to automation, a three steps segmentation approach has been developed. This approach is composed of two phases: a segmentation into sub-spaces namely floors and rooms and a plane segmentation combined with the identification of building elements. In order to assess and validate the developed approach, different case studies are considered. Indeed, it is essential to apply algorithms to several datasets and not to develop algorithms with a unique dataset which could influence the development with its particularities. Indoor point clouds of different types of buildings will be used as input for the developed algorithms, going from an individual house of almost one hundred square meters to larger buildings of several thousand square meters. Datasets provide various space configurations and present numerous different occluding objects as for example desks, computer equipments, home furnishings and even wine barrels. For each dataset, the results will be illustrated. The analysis of the results will provide an insight into the transferability of the developed approach for the indoor modelling of several types of buildings.

A Problem-Centered Approach to Canonical Matrix Forms

ERIC Educational Resources Information Center

Sylvestre, Jeremy

2014-01-01

This article outlines a problem-centered approach to the topic of canonical matrix forms in a second linear algebra course. In this approach, abstract theory, including such topics as eigenvalues, generalized eigenspaces, invariant subspaces, independent subspaces, nilpotency, and cyclic spaces, is developed in response to the patterns discovered…

A Circuit-Based Neural Network with Hybrid Learning of Backpropagation and Random Weight Change Algorithms

PubMed Central

Yang, Changju; Kim, Hyongsuk; Adhikari, Shyam Prasad; Chua, Leon O.

2016-01-01

A hybrid learning method of a software-based backpropagation learning and a hardware-based RWC learning is proposed for the development of circuit-based neural networks. The backpropagation is known as one of the most efficient learning algorithms. A weak point is that its hardware implementation is extremely difficult. The RWC algorithm, which is very easy to implement with respect to its hardware circuits, takes too many iterations for learning. The proposed learning algorithm is a hybrid one of these two. The main learning is performed with a software version of the BP algorithm, firstly, and then, learned weights are transplanted on a hardware version of a neural circuit. At the time of the weight transplantation, a significant amount of output error would occur due to the characteristic difference between the software and the hardware. In the proposed method, such error is reduced via a complementary learning of the RWC algorithm, which is implemented in a simple hardware. The usefulness of the proposed hybrid learning system is verified via simulations upon several classical learning problems. PMID:28025566

Wearable Sensor Localization Considering Mixed Distributed Sources in Health Monitoring Systems

PubMed Central

Wan, Liangtian; Han, Guangjie; Wang, Hao; Shu, Lei; Feng, Nanxing; Peng, Bao

2016-01-01

In health monitoring systems, the base station (BS) and the wearable sensors communicate with each other to construct a virtual multiple input and multiple output (VMIMO) system. In real applications, the signal that the BS received is a distributed source because of the scattering, reflection, diffraction and refraction in the propagation path. In this paper, a 2D direction-of-arrival (DOA) estimation algorithm for incoherently-distributed (ID) and coherently-distributed (CD) sources is proposed based on multiple VMIMO systems. ID and CD sources are separated through the second-order blind identification (SOBI) algorithm. The traditional estimating signal parameters via the rotational invariance technique (ESPRIT)-based algorithm is valid only for one-dimensional (1D) DOA estimation for the ID source. By constructing the signal subspace, two rotational invariant relationships are constructed. Then, we extend the ESPRIT to estimate 2D DOAs for ID sources. For DOA estimation of CD sources, two rational invariance relationships are constructed based on the application of generalized steering vectors (GSVs). Then, the ESPRIT-based algorithm is used for estimating the eigenvalues of two rational invariance matrices, which contain the angular parameters. The expressions of azimuth and elevation for ID and CD sources have closed forms, which means that the spectrum peak searching is avoided. Therefore, compared to the traditional 2D DOA estimation algorithms, the proposed algorithm imposes significantly low computational complexity. The intersecting point of two rays, which come from two different directions measured by two uniform rectangle arrays (URA), can be regarded as the location of the biosensor (wearable sensor). Three BSs adopting the smart antenna (SA) technique cooperate with each other to locate the wearable sensors using the angulation positioning method. Simulation results demonstrate the effectiveness of the proposed algorithm. PMID:26985896

Wearable Sensor Localization Considering Mixed Distributed Sources in Health Monitoring Systems.

PubMed

Wan, Liangtian; Han, Guangjie; Wang, Hao; Shu, Lei; Feng, Nanxing; Peng, Bao

2016-03-12

In health monitoring systems, the base station (BS) and the wearable sensors communicate with each other to construct a virtual multiple input and multiple output (VMIMO) system. In real applications, the signal that the BS received is a distributed source because of the scattering, reflection, diffraction and refraction in the propagation path. In this paper, a 2D direction-of-arrival (DOA) estimation algorithm for incoherently-distributed (ID) and coherently-distributed (CD) sources is proposed based on multiple VMIMO systems. ID and CD sources are separated through the second-order blind identification (SOBI) algorithm. The traditional estimating signal parameters via the rotational invariance technique (ESPRIT)-based algorithm is valid only for one-dimensional (1D) DOA estimation for the ID source. By constructing the signal subspace, two rotational invariant relationships are constructed. Then, we extend the ESPRIT to estimate 2D DOAs for ID sources. For DOA estimation of CD sources, two rational invariance relationships are constructed based on the application of generalized steering vectors (GSVs). Then, the ESPRIT-based algorithm is used for estimating the eigenvalues of two rational invariance matrices, which contain the angular parameters. The expressions of azimuth and elevation for ID and CD sources have closed forms, which means that the spectrum peak searching is avoided. Therefore, compared to the traditional 2D DOA estimation algorithms, the proposed algorithm imposes significantly low computational complexity. The intersecting point of two rays, which come from two different directions measured by two uniform rectangle arrays (URA), can be regarded as the location of the biosensor (wearable sensor). Three BSs adopting the smart antenna (SA) technique cooperate with each other to locate the wearable sensors using the angulation positioning method. Simulation results demonstrate the effectiveness of the proposed algorithm.

Optimal sensor placement for spatial lattice structure based on genetic algorithms

NASA Astrophysics Data System (ADS)

Liu, Wei; Gao, Wei-cheng; Sun, Yi; Xu, Min-jian

2008-10-01

Optimal sensor placement technique plays a key role in structural health monitoring of spatial lattice structures. This paper considers the problem of locating sensors on a spatial lattice structure with the aim of maximizing the data information so that structural dynamic behavior can be fully characterized. Based on the criterion of optimal sensor placement for modal test, an improved genetic algorithm is introduced to find the optimal placement of sensors. The modal strain energy (MSE) and the modal assurance criterion (MAC) have been taken as the fitness function, respectively, so that three placement designs were produced. The decimal two-dimension array coding method instead of binary coding method is proposed to code the solution. Forced mutation operator is introduced when the identical genes appear via the crossover procedure. A computational simulation of a 12-bay plain truss model has been implemented to demonstrate the feasibility of the three optimal algorithms above. The obtained optimal sensor placements using the improved genetic algorithm are compared with those gained by exiting genetic algorithm using the binary coding method. Further the comparison criterion based on the mean square error between the finite element method (FEM) mode shapes and the Guyan expansion mode shapes identified by data-driven stochastic subspace identification (SSI-DATA) method are employed to demonstrate the advantage of the different fitness function. The results showed that some innovations in genetic algorithm proposed in this paper can enlarge the genes storage and improve the convergence of the algorithm. More importantly, the three optimal sensor placement methods can all provide the reliable results and identify the vibration characteristics of the 12-bay plain truss model accurately.

Simultaneous multigrid techniques for nonlinear eigenvalue problems: Solutions of the nonlinear Schrödinger-Poisson eigenvalue problem in two and three dimensions

NASA Astrophysics Data System (ADS)

Costiner, Sorin; Ta'asan, Shlomo

1995-07-01

Algorithms for nonlinear eigenvalue problems (EP's) often require solving self-consistently a large number of EP's. Convergence difficulties may occur if the solution is not sought in an appropriate region, if global constraints have to be satisfied, or if close or equal eigenvalues are present. Multigrid (MG) algorithms for nonlinear problems and for EP's obtained from discretizations of partial differential EP have often been shown to be more efficient than single level algorithms. This paper presents MG techniques and a MG algorithm for nonlinear Schrödinger Poisson EP's. The algorithm overcomes the above mentioned difficulties combining the following techniques: a MG simultaneous treatment of the eigenvectors and nonlinearity, and with the global constrains; MG stable subspace continuation techniques for the treatment of nonlinearity; and a MG projection coupled with backrotations for separation of solutions. These techniques keep the solutions in an appropriate region, where the algorithm converges fast, and reduce the large number of self-consistent iterations to only a few or one MG simultaneous iteration. The MG projection makes it possible to efficiently overcome difficulties related to clusters of close and equal eigenvalues. Computational examples for the nonlinear Schrödinger-Poisson EP in two and three dimensions, presenting special computational difficulties that are due to the nonlinearity and to the equal and closely clustered eigenvalues are demonstrated. For these cases, the algorithm requires O(qN) operations for the calculation of q eigenvectors of size N and for the corresponding eigenvalues. One MG simultaneous cycle per fine level was performed. The total computational cost is equivalent to only a few Gauss-Seidel relaxations per eigenvector. An asymptotic convergence rate of 0.15 per MG cycle is attained.

Experimental implementation of heat-bath algorithmic cooling using solid-state nuclear magnetic resonance.

PubMed

Baugh, J; Moussa, O; Ryan, C A; Nayak, A; Laflamme, R

2005-11-24

The counter-intuitive properties of quantum mechanics have the potential to revolutionize information processing by enabling the development of efficient algorithms with no known classical counterparts. Harnessing this power requires the development of a set of building blocks, one of which is a method to initialize the set of quantum bits (qubits) to a known state. Additionally, fresh ancillary qubits must be available during the course of computation to achieve fault tolerance. In any physical system used to implement quantum computation, one must therefore be able to selectively and dynamically remove entropy from the part of the system that is to be mapped to qubits. One such method is an 'open-system' cooling protocol in which a subset of qubits can be brought into contact with an external system of large heat capacity. Theoretical efforts have led to an implementation-independent cooling procedure, namely heat-bath algorithmic cooling. These efforts have culminated with the proposal of an optimal algorithm, the partner-pairing algorithm, which was used to compute the physical limits of heat-bath algorithmic cooling. Here we report the experimental realization of multi-step cooling of a quantum system via heat-bath algorithmic cooling. The experiment was carried out using nuclear magnetic resonance of a solid-state ensemble three-qubit system. We demonstrate the repeated repolarization of a particular qubit to an effective spin-bath temperature, and alternating logical operations within the three-qubit subspace to ultimately cool a second qubit below this temperature. Demonstration of the control necessary for these operations represents an important step forward in the manipulation of solid-state nuclear magnetic resonance qubits.

Superior Generalization Capability of Hardware-Learing Algorithm Developed for Self-Learning Neuron-MOS Neural Networks

NASA Astrophysics Data System (ADS)

Kondo, Shuhei; Shibata, Tadashi; Ohmi, Tadahiro

1995-02-01

We have investigated the learning performance of the hardware backpropagation (HBP) algorithm, a hardware-oriented learning algorithm developed for the self-learning architecture of neural networks constructed using neuron MOS (metal-oxide-semiconductor) transistors. The solution to finding a mirror symmetry axis in a 4×4 binary pixel array was tested by computer simulation based on the HBP algorithm. Despite the inherent restrictions imposed on the hardware-learning algorithm, HBP exhibits equivalent learning performance to that of the original backpropagation (BP) algorithm when all the pertinent parameters are optimized. Very importantly, we have found that HBP has a superior generalization capability over BP; namely, HBP exhibits higher performance in solving problems that the network has not yet learnt.

Learning control system design based on 2-D theory - An application to parallel link manipulator

NASA Technical Reports Server (NTRS)

Geng, Z.; Carroll, R. L.; Lee, J. D.; Haynes, L. H.

1990-01-01

An approach to iterative learning control system design based on two-dimensional system theory is presented. A two-dimensional model for the iterative learning control system which reveals the connections between learning control systems and two-dimensional system theory is established. A learning control algorithm is proposed, and the convergence of learning using this algorithm is guaranteed by two-dimensional stability. The learning algorithm is applied successfully to the trajectory tracking control problem for a parallel link robot manipulator. The excellent performance of this learning algorithm is demonstrated by the computer simulation results.

Active Learning Using Hint Information.

PubMed

Li, Chun-Liang; Ferng, Chun-Sung; Lin, Hsuan-Tien

2015-08-01

The abundance of real-world data and limited labeling budget calls for active learning, an important learning paradigm for reducing human labeling efforts. Many recently developed active learning algorithms consider both uncertainty and representativeness when making querying decisions. However, exploiting representativeness with uncertainty concurrently usually requires tackling sophisticated and challenging learning tasks, such as clustering. In this letter, we propose a new active learning framework, called hinted sampling, which takes both uncertainty and representativeness into account in a simpler way. We design a novel active learning algorithm within the hinted sampling framework with an extended support vector machine. Experimental results validate that the novel active learning algorithm can result in a better and more stable performance than that achieved by state-of-the-art algorithms. We also show that the hinted sampling framework allows improving another active learning algorithm designed from the transductive support vector machine.

De-biasing the dynamic mode decomposition for applied Koopman spectral analysis of noisy datasets

NASA Astrophysics Data System (ADS)

Hemati, Maziar S.; Rowley, Clarence W.; Deem, Eric A.; Cattafesta, Louis N.

2017-08-01

The dynamic mode decomposition (DMD)—a popular method for performing data-driven Koopman spectral analysis—has gained increased popularity for extracting dynamically meaningful spatiotemporal descriptions of fluid flows from snapshot measurements. Often times, DMD descriptions can be used for predictive purposes as well, which enables informed decision-making based on DMD model forecasts. Despite its widespread use and utility, DMD can fail to yield accurate dynamical descriptions when the measured snapshot data are imprecise due to, e.g., sensor noise. Here, we express DMD as a two-stage algorithm in order to isolate a source of systematic error. We show that DMD's first stage, a subspace projection step, systematically introduces bias errors by processing snapshots asymmetrically. To remove this systematic error, we propose utilizing an augmented snapshot matrix in a subspace projection step, as in problems of total least-squares, in order to account for the error present in all snapshots. The resulting unbiased and noise-aware total DMD (TDMD) formulation reduces to standard DMD in the absence of snapshot errors, while the two-stage perspective generalizes the de-biasing framework to other related methods as well. TDMD's performance is demonstrated in numerical and experimental fluids examples. In particular, in the analysis of time-resolved particle image velocimetry data for a separated flow, TDMD outperforms standard DMD by providing dynamical interpretations that are consistent with alternative analysis techniques. Further, TDMD extracts modes that reveal detailed spatial structures missed by standard DMD.

High-throughput ocular artifact reduction in multichannel electroencephalography (EEG) using component subspace projection.

PubMed

Ma, Junshui; Bayram, Sevinç; Tao, Peining; Svetnik, Vladimir

2011-03-15

After a review of the ocular artifact reduction literature, a high-throughput method designed to reduce the ocular artifacts in multichannel continuous EEG recordings acquired at clinical EEG laboratories worldwide is proposed. The proposed method belongs to the category of component-based methods, and does not rely on any electrooculography (EOG) signals. Based on a concept that all ocular artifact components exist in a signal component subspace, the method can uniformly handle all types of ocular artifacts, including eye-blinks, saccades, and other eye movements, by automatically identifying ocular components from decomposed signal components. This study also proposes an improved strategy to objectively and quantitatively evaluate artifact reduction methods. The evaluation strategy uses real EEG signals to synthesize realistic simulated datasets with different amounts of ocular artifacts. The simulated datasets enable us to objectively demonstrate that the proposed method outperforms some existing methods when no high-quality EOG signals are available. Moreover, the results of the simulated datasets improve our understanding of the involved signal decomposition algorithms, and provide us with insights into the inconsistency regarding the performance of different methods in the literature. The proposed method was also applied to two independent clinical EEG datasets involving 28 volunteers and over 1000 EEG recordings. This effort further confirms that the proposed method can effectively reduce ocular artifacts in large clinical EEG datasets in a high-throughput fashion. Copyright © 2011 Elsevier B.V. All rights reserved.

Multigrid method for stability problems

NASA Technical Reports Server (NTRS)

Taasan, Shlomo

1988-01-01

The problem of calculating the stability of steady state solutions of differential equations is treated. Leading eigenvalues (i.e., having maximal real part) of large matrices that arise from discretization are to be calculated. An efficient multigrid method for solving these problems is presented. The method begins by obtaining an initial approximation for the dominant subspace on a coarse level using a damped Jacobi relaxation. This proceeds until enough accuracy for the dominant subspace has been obtained. The resulting grid functions are then used as an initial approximation for appropriate eigenvalue problems. These problems are being solved first on coarse levels, followed by refinement until a desired accuracy for the eigenvalues has been achieved. The method employs local relaxation on all levels together with a global change on the coarsest level only, which is designed to separate the different eigenfunctions as well as to update their corresponding eigenvalues. Coarsening is done using the FAS formulation in a non-standard way in which the right hand side of the coarse grid equations involves unknown parameters to be solved for on the coarse grid. This in particular leads to a new multigrid method for calculating the eigenvalues of symmetric problems. Numerical experiments with a model problem demonstrate the effectiveness of the method proposed. Using an FMG algorithm a solution to the level of discretization errors is obtained in just a few work units (less than 10), where a work unit is the work involved in one Jacobi relization on the finest level.

Euclidean commute time distance embedding and its application to spectral anomaly detection

NASA Astrophysics Data System (ADS)

Albano, James A.; Messinger, David W.

2012-06-01

Spectral image analysis problems often begin by performing a preprocessing step composed of applying a transformation that generates an alternative representation of the spectral data. In this paper, a transformation based on a Markov-chain model of a random walk on a graph is introduced. More precisely, we quantify the random walk using a quantity known as the average commute time distance and find a nonlinear transformation that embeds the nodes of a graph in a Euclidean space where the separation between them is equal to the square root of this quantity. This has been referred to as the Commute Time Distance (CTD) transformation and it has the important characteristic of increasing when the number of paths between two nodes decreases and/or the lengths of those paths increase. Remarkably, a closed form solution exists for computing the average commute time distance that avoids running an iterative process and is found by simply performing an eigendecomposition on the graph Laplacian matrix. Contained in this paper is a discussion of the particular graph constructed on the spectral data for which the commute time distance is then calculated from, an introduction of some important properties of the graph Laplacian matrix, and a subspace projection that approximately preserves the maximal variance of the square root commute time distance. Finally, RX anomaly detection and Topological Anomaly Detection (TAD) algorithms will be applied to the CTD subspace followed by a discussion of their results.

Active Subspace Methods for Data-Intensive Inverse Problems

DOE Office of Scientific and Technical Information (OSTI.GOV)

Wang, Qiqi

2017-04-27

The project has developed theory and computational tools to exploit active subspaces to reduce the dimension in statistical calibration problems. This dimension reduction enables MCMC methods to calibrate otherwise intractable models. The same theoretical and computational tools can also reduce the measurement dimension for calibration problems that use large stores of data.

Basis adaptation in homogeneous chaos spaces

DOE Office of Scientific and Technical Information (OSTI.GOV)

Tipireddy, Ramakrishna; Ghanem, Roger

2014-02-01

We present a new meth for the characterization of subspaces associated with low-dimensional quantities of interet (QoI). The probability density function of these QoI is found to be concentrated around one-dimensional subspaces for which we develop projection operators. Our approach builds on the properties of Gaussian Hilbert spaces and associated tensor product spaces.

Active subspace uncertainty quantification for a polydomain ferroelectric phase-field model

NASA Astrophysics Data System (ADS)

Leon, Lider S.; Smith, Ralph C.; Miles, Paul; Oates, William S.

2018-03-01

Quantum-informed ferroelectric phase field models capable of predicting material behavior, are necessary for facilitating the development and production of many adaptive structures and intelligent systems. Uncertainty is present in these models, given the quantum scale at which calculations take place. A necessary analysis is to determine how the uncertainty in the response can be attributed to the uncertainty in the model inputs or parameters. A second analysis is to identify active subspaces within the original parameter space, which quantify directions in which the model response varies most dominantly, thus reducing sampling effort and computational cost. In this investigation, we identify an active subspace for a poly-domain ferroelectric phase-field model. Using the active variables as our independent variables, we then construct a surrogate model and perform Bayesian inference. Once we quantify the uncertainties in the active variables, we obtain uncertainties for the original parameters via an inverse mapping. The analysis provides insight into how active subspace methodologies can be used to reduce computational power needed to perform Bayesian inference on model parameters informed by experimental or simulated data.

Data-driven modeling and predictive control for boiler-turbine unit using fuzzy clustering and subspace methods.

PubMed

Wu, Xiao; Shen, Jiong; Li, Yiguo; Lee, Kwang Y

2014-05-01

This paper develops a novel data-driven fuzzy modeling strategy and predictive controller for boiler-turbine unit using fuzzy clustering and subspace identification (SID) methods. To deal with the nonlinear behavior of boiler-turbine unit, fuzzy clustering is used to provide an appropriate division of the operation region and develop the structure of the fuzzy model. Then by combining the input data with the corresponding fuzzy membership functions, the SID method is extended to extract the local state-space model parameters. Owing to the advantages of the both methods, the resulting fuzzy model can represent the boiler-turbine unit very closely, and a fuzzy model predictive controller is designed based on this model. As an alternative approach, a direct data-driven fuzzy predictive control is also developed following the same clustering and subspace methods, where intermediate subspace matrices developed during the identification procedure are utilized directly as the predictor. Simulation results show the advantages and effectiveness of the proposed approach. Copyright © 2014 ISA. Published by Elsevier Ltd. All rights reserved.

Learn the Lagrangian: A Vector-Valued RKHS Approach to Identifying Lagrangian Systems.

PubMed

Cheng, Ching-An; Huang, Han-Pang

2016-12-01

We study the modeling of Lagrangian systems with multiple degrees of freedom. Based on system dynamics, canonical parametric models require ad hoc derivations and sometimes simplification for a computable solution; on the other hand, due to the lack of prior knowledge in the system's structure, modern nonparametric models in machine learning face the curse of dimensionality, especially in learning large systems. In this paper, we bridge this gap by unifying the theories of Lagrangian systems and vector-valued reproducing kernel Hilbert space. We reformulate Lagrangian systems with kernels that embed the governing Euler-Lagrange equation-the Lagrangian kernels-and show that these kernels span a subspace capturing the Lagrangian's projection as inverse dynamics. By such property, our model uses only inputs and outputs as in machine learning and inherits the structured form as in system dynamics, thereby removing the need for the mundane derivations for new systems as well as the generalization problem in learning from scratches. In effect, it learns the system's Lagrangian, a simpler task than directly learning the dynamics. To demonstrate, we applied the proposed kernel to identify the robot inverse dynamics in simulations and experiments. Our results present a competitive novel approach to identifying Lagrangian systems, despite using only inputs and outputs.

A developmental approach to learning causal models for cyber security

NASA Astrophysics Data System (ADS)

Mugan, Jonathan

2013-05-01

To keep pace with our adversaries, we must expand the scope of machine learning and reasoning to address the breadth of possible attacks. One approach is to employ an algorithm to learn a set of causal models that describes the entire cyber network and each host end node. Such a learning algorithm would run continuously on the system and monitor activity in real time. With a set of causal models, the algorithm could anticipate novel attacks, take actions to thwart them, and predict the second-order effects flood of information, and the algorithm would have to determine which streams of that flood were relevant in which situations. This paper will present the results of efforts toward the application of a developmental learning algorithm to the problem of cyber security. The algorithm is modeled on the principles of human developmental learning and is designed to allow an agent to learn about the computer system in which it resides through active exploration. Children are flexible learners who acquire knowledge by actively exploring their environment and making predictions about what they will find,1, 2 and our algorithm is inspired by the work of the developmental psychologist Jean Piaget.3 Piaget described how children construct knowledge in stages and learn new concepts on top of those they already know. Developmental learning allows our algorithm to focus on subsets of the environment that are most helpful for learning given its current knowledge. In experiments, the algorithm was able to learn the conditions for file exfiltration and use that knowledge to protect sensitive files.

Discrete-Time Deterministic $Q$ -Learning: A Novel Convergence Analysis.

PubMed

Wei, Qinglai; Lewis, Frank L; Sun, Qiuye; Yan, Pengfei; Song, Ruizhuo

2017-05-01

In this paper, a novel discrete-time deterministic Q -learning algorithm is developed. In each iteration of the developed Q -learning algorithm, the iterative Q function is updated for all the state and control spaces, instead of updating for a single state and a single control in traditional Q -learning algorithm. A new convergence criterion is established to guarantee that the iterative Q function converges to the optimum, where the convergence criterion of the learning rates for traditional Q -learning algorithms is simplified. During the convergence analysis, the upper and lower bounds of the iterative Q function are analyzed to obtain the convergence criterion, instead of analyzing the iterative Q function itself. For convenience of analysis, the convergence properties for undiscounted case of the deterministic Q -learning algorithm are first developed. Then, considering the discounted factor, the convergence criterion for the discounted case is established. Neural networks are used to approximate the iterative Q function and compute the iterative control law, respectively, for facilitating the implementation of the deterministic Q -learning algorithm. Finally, simulation results and comparisons are given to illustrate the performance of the developed algorithm.

Visual characterization and diversity quantification of chemical libraries: 2. Analysis and selection of size-independent, subspace-specific diversity indices.

PubMed

Colliandre, Lionel; Le Guilloux, Vincent; Bourg, Stephane; Morin-Allory, Luc

2012-02-27

High Throughput Screening (HTS) is a standard technique widely used to find hit compounds in drug discovery projects. The high costs associated with such experiments have highlighted the need to carefully design screening libraries in order to avoid wasting resources. Molecular diversity is an established concept that has been used to this end for many years. In this article, a new approach to quantify the molecular diversity of screening libraries is presented. The approach is based on the Delimited Reference Chemical Subspace (DRCS) methodology, a new method that can be used to delimit the densest subspace spanned by a reference library in a reduced 2D continuous space. A total of 22 diversity indices were implemented or adapted to this methodology, which is used here to remove outliers and obtain a relevant cell-based partition of the subspace. The behavior of these indices was assessed and compared in various extreme situations and with respect to a set of theoretical rules that a diversity function should satisfy when libraries of different sizes have to be compared. Some gold standard indices are found inappropriate in such a context, while none of the tested indices behave perfectly in all cases. Five DRCS-based indices accounting for different aspects of diversity were finally selected, and a simple framework is proposed to use them effectively. Various libraries have been profiled with respect to more specific subspaces, which further illustrate the interest of the method.

Improved Detection of Local Earthquakes in the Vienna Basin (Austria), using Subspace Detectors

NASA Astrophysics Data System (ADS)

Apoloner, Maria-Theresia; Caffagni, Enrico; Bokelmann, Götz

2016-04-01

The Vienna Basin in Eastern Austria is densely populated and highly-developed; it is also a region of low to moderate seismicity, yet the seismological network coverage is relatively sparse. This demands improving our capability of earthquake detection by testing new methods, enlarging the existing local earthquake catalogue. This contributes to imaging tectonic fault zones for better understanding seismic hazard, also through improved earthquake statistics (b-value, magnitude of completeness). Detection of low-magnitude earthquakes or events for which the highest amplitudes slightly exceed the signal-to-noise-ratio (SNR), may be possible by using standard methods like the short-term over long-term average (STA/LTA). However, due to sparse network coverage and high background noise, such a technique may not detect all potentially recoverable events. Yet, earthquakes originating from the same source region and relatively close to each other, should be characterized by similarity in seismic waveforms, at a given station. Therefore, waveform similarity can be exploited by using specific techniques such as correlation-template based (also known as matched filtering) or subspace detection methods (based on the subspace theory). Matching techniques basically require a reference or template event, usually characterized by high waveform coherence in the array receivers, and high SNR, which is cross-correlated with the continuous data. Instead, subspace detection methods overcome in principle the necessity of defining template events as single events, but use a subspace extracted from multiple events. This approach theoretically should be more robust in detecting signals that exhibit a strong variability (e.g. because of source or magnitude). In this study we scan the continuous data recorded in the Vienna Basin with a subspace detector to identify additional events. This will allow us to estimate the increase of the seismicity rate in the local earthquake catalogue, therefore providing an evaluation of network performance and efficiency of the method.

A novel manifold-manifold distance index applied to looseness state assessment of viscoelastic sandwich structures

NASA Astrophysics Data System (ADS)

Sun, Chuang; Zhang, Zhousuo; Guo, Ting; Luo, Xue; Qu, Jinxiu; Zhang, Chenxuan; Cheng, Wei; Li, Bing

2014-06-01

Viscoelastic sandwich structures (VSS) are widely used in mechanical equipment; their state assessment is necessary to detect structural states and to keep equipment running with high reliability. This paper proposes a novel manifold-manifold distance-based assessment (M2DBA) method for assessing the looseness state in VSSs. In the M2DBA method, a manifold-manifold distance is viewed as a health index. To design the index, response signals from the structure are firstly acquired by condition monitoring technology and a Hankel matrix is constructed by using the response signals to describe state patterns of the VSS. Thereafter, a subspace analysis method, that is, principal component analysis (PCA), is performed to extract the condition subspace hidden in the Hankel matrix. From the subspace, pattern changes in dynamic structural properties are characterized. Further, a Grassmann manifold (GM) is formed by organizing a set of subspaces. The manifold is mapped to a reproducing kernel Hilbert space (RKHS), where support vector data description (SVDD) is used to model the manifold as a hypersphere. Finally, a health index is defined as the cosine of the angle between the hypersphere centers corresponding to the structural baseline state and the looseness state. The defined health index contains similarity information existing in the two structural states, so structural looseness states can be effectively identified. Moreover, the health index is derived by analysis of the global properties of subspace sets, which is different from traditional subspace analysis methods. The effectiveness of the health index for state assessment is validated by test data collected from a VSS subjected to different degrees of looseness. The results show that the health index is a very effective metric for detecting the occurrence and extension of structural looseness. Comparison results indicate that the defined index outperforms some existing state-of-the-art ones.

Information theoretic partitioning and confidence based weight assignment for multi-classifier decision level fusion in hyperspectral target recognition applications

NASA Astrophysics Data System (ADS)

Prasad, S.; Bruce, L. M.

2007-04-01

There is a growing interest in using multiple sources for automatic target recognition (ATR) applications. One approach is to take multiple, independent observations of a phenomenon and perform a feature level or a decision level fusion for ATR. This paper proposes a method to utilize these types of multi-source fusion techniques to exploit hyperspectral data when only a small number of training pixels are available. Conventional hyperspectral image based ATR techniques project the high dimensional reflectance signature onto a lower dimensional subspace using techniques such as Principal Components Analysis (PCA), Fisher's linear discriminant analysis (LDA), subspace LDA and stepwise LDA. While some of these techniques attempt to solve the curse of dimensionality, or small sample size problem, these are not necessarily optimal projections. In this paper, we present a divide and conquer approach to address the small sample size problem. The hyperspectral space is partitioned into contiguous subspaces such that the discriminative information within each subspace is maximized, and the statistical dependence between subspaces is minimized. We then treat each subspace as a separate source in a multi-source multi-classifier setup and test various decision fusion schemes to determine their efficacy. Unlike previous approaches which use correlation between variables for band grouping, we study the efficacy of higher order statistical information (using average mutual information) for a bottom up band grouping. We also propose a confidence measure based decision fusion technique, where the weights associated with various classifiers are based on their confidence in recognizing the training data. To this end, training accuracies of all classifiers are used for weight assignment in the fusion process of test pixels. The proposed methods are tested using hyperspectral data with known ground truth, such that the efficacy can be quantitatively measured in terms of target recognition accuracies.

A study on the performance comparison of metaheuristic algorithms on the learning of neural networks

NASA Astrophysics Data System (ADS)

Lai, Kee Huong; Zainuddin, Zarita; Ong, Pauline

2017-08-01

The learning or training process of neural networks entails the task of finding the most optimal set of parameters, which includes translation vectors, dilation parameter, synaptic weights, and bias terms. Apart from the traditional gradient descent-based methods, metaheuristic methods can also be used for this learning purpose. Since the inception of genetic algorithm half a century ago, the last decade witnessed the explosion of a variety of novel metaheuristic algorithms, such as harmony search algorithm, bat algorithm, and whale optimization algorithm. Despite the proof of the no free lunch theorem in the discipline of optimization, a survey in the literature of machine learning gives contrasting results. Some researchers report that certain metaheuristic algorithms are superior to the others, whereas some others argue that different metaheuristic algorithms give comparable performance. As such, this paper aims to investigate if a certain metaheuristic algorithm will outperform the other algorithms. In this work, three metaheuristic algorithms, namely genetic algorithms, particle swarm optimization, and harmony search algorithm are considered. The algorithms are incorporated in the learning of neural networks and their classification results on the benchmark UCI machine learning data sets are compared. It is found that all three metaheuristic algorithms give similar and comparable performance, as captured in the average overall classification accuracy. The results corroborate the findings reported in the works done by previous researchers. Several recommendations are given, which include the need of statistical analysis to verify the results and further theoretical works to support the obtained empirical results.

Impact of JPEG2000 compression on endmember extraction and unmixing of remotely sensed hyperspectral data

NASA Astrophysics Data System (ADS)

Martin, Gabriel; Gonzalez-Ruiz, Vicente; Plaza, Antonio; Ortiz, Juan P.; Garcia, Inmaculada

2010-07-01

Lossy hyperspectral image compression has received considerable interest in recent years due to the extremely high dimensionality of the data. However, the impact of lossy compression on spectral unmixing techniques has not been widely studied. These techniques characterize mixed pixels (resulting from insufficient spatial resolution) in terms of a suitable combination of spectrally pure substances (called endmembers) weighted by their estimated fractional abundances. This paper focuses on the impact of JPEG2000-based lossy compression of hyperspectral images on the quality of the endmembers extracted by different algorithms. The three considered algorithms are the orthogonal subspace projection (OSP), which uses only spatial information, and the automatic morphological endmember extraction (AMEE) and spatial spectral endmember extraction (SSEE), which integrate both spatial and spectral information in the search for endmembers. The impact of compression on the resulting abundance estimation based on the endmembers derived by different methods is also substantiated. Experimental results are conducted using a hyperspectral data set collected by NASA Jet Propulsion Laboratory over the Cuprite mining district in Nevada. The experimental results are quantitatively analyzed using reference information available from U.S. Geological Survey, resulting in recommendations to specialists interested in applying endmember extraction and unmixing algorithms to compressed hyperspectral data.

Detection of mouse liver cancer via a parallel iterative shrinkage method in hybrid optical/microcomputed tomography imaging

NASA Astrophysics Data System (ADS)

Wu, Ping; Liu, Kai; Zhang, Qian; Xue, Zhenwen; Li, Yongbao; Ning, Nannan; Yang, Xin; Li, Xingde; Tian, Jie

2012-12-01

Liver cancer is one of the most common malignant tumors worldwide. In order to enable the noninvasive detection of small liver tumors in mice, we present a parallel iterative shrinkage (PIS) algorithm for dual-modality tomography. It takes advantage of microcomputed tomography and multiview bioluminescence imaging, providing anatomical structure and bioluminescence intensity information to reconstruct the size and location of tumors. By incorporating prior knowledge of signal sparsity, we associate some mathematical strategies including specific smooth convex approximation, an iterative shrinkage operator, and affine subspace with the PIS method, which guarantees the accuracy, efficiency, and reliability for three-dimensional reconstruction. Then an in vivo experiment on the bead-implanted mouse has been performed to validate the feasibility of this method. The findings indicate that a tiny lesion less than 3 mm in diameter can be localized with a position bias no more than 1 mm the computational efficiency is one to three orders of magnitude faster than the existing algorithms; this approach is robust to the different regularization parameters and the lp norms. Finally, we have applied this algorithm to another in vivo experiment on an HCCLM3 orthotopic xenograft mouse model, which suggests the PIS method holds the promise for practical applications of whole-body cancer detection.

Search algorithm complexity modeling with application to image alignment and matching

NASA Astrophysics Data System (ADS)

DelMarco, Stephen

2014-05-01

Search algorithm complexity modeling, in the form of penetration rate estimation, provides a useful way to estimate search efficiency in application domains which involve searching over a hypothesis space of reference templates or models, as in model-based object recognition, automatic target recognition, and biometric recognition. The penetration rate quantifies the expected portion of the database that must be searched, and is useful for estimating search algorithm computational requirements. In this paper we perform mathematical modeling to derive general equations for penetration rate estimates that are applicable to a wide range of recognition problems. We extend previous penetration rate analyses to use more general probabilistic modeling assumptions. In particular we provide penetration rate equations within the framework of a model-based image alignment application domain in which a prioritized hierarchical grid search is used to rank subspace bins based on matching probability. We derive general equations, and provide special cases based on simplifying assumptions. We show how previously-derived penetration rate equations are special cases of the general formulation. We apply the analysis to model-based logo image alignment in which a hierarchical grid search is used over a geometric misalignment transform hypothesis space. We present numerical results validating the modeling assumptions and derived formulation.

Hydraulic tomography of discrete networks of conduits and fractures in a karstic aquifer by using a deterministic inversion algorithm

NASA Astrophysics Data System (ADS)

Fischer, P.; Jardani, A.; Lecoq, N.

2018-02-01

In this paper, we present a novel inverse modeling method called Discrete Network Deterministic Inversion (DNDI) for mapping the geometry and property of the discrete network of conduits and fractures in the karstified aquifers. The DNDI algorithm is based on a coupled discrete-continuum concept to simulate numerically water flows in a model and a deterministic optimization algorithm to invert a set of observed piezometric data recorded during multiple pumping tests. In this method, the model is partioned in subspaces piloted by a set of parameters (matrix transmissivity, and geometry and equivalent transmissivity of the conduits) that are considered as unknown. In this way, the deterministic optimization process can iteratively correct the geometry of the network and the values of the properties, until it converges to a global network geometry in a solution model able to reproduce the set of data. An uncertainty analysis of this result can be performed from the maps of posterior uncertainties on the network geometry or on the property values. This method has been successfully tested for three different theoretical and simplified study cases with hydraulic responses data generated from hypothetical karstic models with an increasing complexity of the network geometry, and of the matrix heterogeneity.

Faceting for direction-dependent spectral deconvolution

NASA Astrophysics Data System (ADS)

Tasse, C.; Hugo, B.; Mirmont, M.; Smirnov, O.; Atemkeng, M.; Bester, L.; Hardcastle, M. J.; Lakhoo, R.; Perkins, S.; Shimwell, T.

2018-04-01

The new generation of radio interferometers is characterized by high sensitivity, wide fields of view and large fractional bandwidth. To synthesize the deepest images enabled by the high dynamic range of these instruments requires us to take into account the direction-dependent Jones matrices, while estimating the spectral properties of the sky in the imaging and deconvolution algorithms. In this paper we discuss and implement a wideband wide-field spectral deconvolution framework (DDFacet) based on image plane faceting, that takes into account generic direction-dependent effects. Specifically, we present a wide-field co-planar faceting scheme, and discuss the various effects that need to be taken into account to solve for the deconvolution problem (image plane normalization, position-dependent Point Spread Function, etc). We discuss two wideband spectral deconvolution algorithms based on hybrid matching pursuit and sub-space optimisation respectively. A few interesting technical features incorporated in our imager are discussed, including baseline dependent averaging, which has the effect of improving computing efficiency. The version of DDFacet presented here can account for any externally defined Jones matrices and/or beam patterns.

An information theory criteria based blind method for enumerating active users in DS-CDMA system

NASA Astrophysics Data System (ADS)

Samsami Khodadad, Farid; Abed Hodtani, Ghosheh

2014-11-01

In this paper, a new and blind algorithm for active user enumeration in asynchronous direct sequence code division multiple access (DS-CDMA) in multipath channel scenario is proposed. The proposed method is based on information theory criteria. There are two main categories of information criteria which are widely used in active user enumeration, Akaike Information Criterion (AIC) and Minimum Description Length (MDL) information theory criteria. The main difference between these two criteria is their penalty functions. Due to this difference, MDL is a consistent enumerator which has better performance in higher signal-to-noise ratios (SNR) but AIC is preferred in lower SNRs. In sequel, we propose a SNR compliance method based on subspace and training genetic algorithm to have the performance of both of them. Moreover, our method uses only a single antenna, in difference to the previous methods which decrease hardware complexity. Simulation results show that the proposed method is capable of estimating the number of active users without any prior knowledge and the efficiency of the method.

Power flow prediction in vibrating systems via model reduction

NASA Astrophysics Data System (ADS)

Li, Xianhui

This dissertation focuses on power flow prediction in vibrating systems. Reduced order models (ROMs) are built based on rational Krylov model reduction which preserve power flow information in the original systems over a specified frequency band. Stiffness and mass matrices of the ROMs are obtained by projecting the original system matrices onto the subspaces spanned by forced responses. A matrix-free algorithm is designed to construct ROMs directly from the power quantities at selected interpolation frequencies. Strategies for parallel implementation of the algorithm via message passing interface are proposed. The quality of ROMs is iteratively refined according to the error estimate based on residual norms. Band capacity is proposed to provide a priori estimate of the sizes of good quality ROMs. Frequency averaging is recast as ensemble averaging and Cauchy distribution is used to simplify the computation. Besides model reduction for deterministic systems, details of constructing ROMs for parametric and nonparametric random systems are also presented. Case studies have been conducted on testbeds from Harwell-Boeing collections. Input and coupling power flow are computed for the original systems and the ROMs. Good agreement is observed in all cases.

Recognizing Activities via Bag of Words for Attribute Dynamics (Open Access)

DTIC Science & Technology

2013-10-03

56.4% 73.4% clean-jerk 83.2% 84.1% 78.2% 79.4% 85.1% 78.2% 85.4% javelin throw 61.1% 74.6% 79.5% 62.1% 87.5% 56.6% 76.7% ham. throw 65.1% 77.5% 70.5...1 ∈ Rτ is the vector of all ones. Each column of C is a basis vector of a latent subspace and the t-th column of X contains the coordinates of the yt...binary PCA is first applied to all attribute score vectors in P ′. The parame- ters of the hidden Gauss-Markov process are then learned by solving a least

A junction-tree based learning algorithm to optimize network wide traffic control: A coordinated multi-agent framework

DOE PAGES

Zhu, Feng; Aziz, H. M. Abdul; Qian, Xinwu; ...

2015-01-31

Our study develops a novel reinforcement learning algorithm for the challenging coordinated signal control problem. Traffic signals are modeled as intelligent agents interacting with the stochastic traffic environment. The model is built on the framework of coordinated reinforcement learning. The Junction Tree Algorithm (JTA) based reinforcement learning is proposed to obtain an exact inference of the best joint actions for all the coordinated intersections. Moreover, the algorithm is implemented and tested with a network containing 18 signalized intersections in VISSIM. Finally, our results show that the JTA based algorithm outperforms independent learning (Q-learning), real-time adaptive learning, and fixed timing plansmore » in terms of average delay, number of stops, and vehicular emissions at the network level.« less

A new Bayesian recursive technique for parameter estimation

NASA Astrophysics Data System (ADS)

Kaheil, Yasir H.; Gill, M. Kashif; McKee, Mac; Bastidas, Luis

2006-08-01

The performance of any model depends on how well its associated parameters are estimated. In the current application, a localized Bayesian recursive estimation (LOBARE) approach is devised for parameter estimation. The LOBARE methodology is an extension of the Bayesian recursive estimation (BARE) method. It is applied in this paper on two different types of models: an artificial intelligence (AI) model in the form of a support vector machine (SVM) application for forecasting soil moisture and a conceptual rainfall-runoff (CRR) model represented by the Sacramento soil moisture accounting (SAC-SMA) model. Support vector machines, based on statistical learning theory (SLT), represent the modeling task as a quadratic optimization problem and have already been used in various applications in hydrology. They require estimation of three parameters. SAC-SMA is a very well known model that estimates runoff. It has a 13-dimensional parameter space. In the LOBARE approach presented here, Bayesian inference is used in an iterative fashion to estimate the parameter space that will most likely enclose a best parameter set. This is done by narrowing the sampling space through updating the "parent" bounds based on their fitness. These bounds are actually the parameter sets that were selected by BARE runs on subspaces of the initial parameter space. The new approach results in faster convergence toward the optimal parameter set using minimum training/calibration data and fewer sets of parameter values. The efficacy of the localized methodology is also compared with the previously used BARE algorithm.

Density scaling for multiplets

NASA Astrophysics Data System (ADS)

Nagy, Á.

2011-02-01

Generalized Kohn-Sham equations are presented for lowest-lying multiplets. The way of treating non-integer particle numbers is coupled with an earlier method of the author. The fundamental quantity of the theory is the subspace density. The Kohn-Sham equations are similar to the conventional Kohn-Sham equations. The difference is that the subspace density is used instead of the density and the Kohn-Sham potential is different for different subspaces. The exchange-correlation functional is studied using density scaling. It is shown that there exists a value of the scaling factor ζ for which the correlation energy disappears. Generalized OPM and Krieger-Li-Iafrate (KLI) methods incorporating correlation are presented. The ζKLI method, being as simple as the original KLI method, is proposed for multiplets.

Switching LPV Control for High Performance Tactical Aircraft

NASA Technical Reports Server (NTRS)

Lu, Bei; Wu, Fen; Kim, SungWan

2004-01-01

This paper examines a switching Linear Parameter-Varying (LPV) control approach to determine if it is practical to use for flight control designs within a wide angle of attack region. The approach is based on multiple parameter-dependent Lyapunov functions. The full parameter space is partitioned into overlapping subspaces and a family of LPV controllers are designed, each suitable for a specific parameter subspace. The hysteresis switching logic is used to accomplish the transition among different parameter subspaces. The proposed switching LPV control scheme is applied to an F-16 aircraft model with different actuator dynamics in low and high angle of attack regions. The nonlinear simulation results show that the aircraft performs well when switching among different angle of attack regions.

Folded concave penalized sparse linear regression: sparsity, statistical performance, and algorithmic theory for local solutions.

PubMed

Liu, Hongcheng; Yao, Tao; Li, Runze; Ye, Yinyu

2017-11-01

This paper concerns the folded concave penalized sparse linear regression (FCPSLR), a class of popular sparse recovery methods. Although FCPSLR yields desirable recovery performance when solved globally, computing a global solution is NP-complete. Despite some existing statistical performance analyses on local minimizers or on specific FCPSLR-based learning algorithms, it still remains open questions whether local solutions that are known to admit fully polynomial-time approximation schemes (FPTAS) may already be sufficient to ensure the statistical performance, and whether that statistical performance can be non-contingent on the specific designs of computing procedures. To address the questions, this paper presents the following threefold results: (i) Any local solution (stationary point) is a sparse estimator, under some conditions on the parameters of the folded concave penalties. (ii) Perhaps more importantly, any local solution satisfying a significant subspace second-order necessary condition (S 3 ONC), which is weaker than the second-order KKT condition, yields a bounded error in approximating the true parameter with high probability. In addition, if the minimal signal strength is sufficient, the S 3 ONC solution likely recovers the oracle solution. This result also explicates that the goal of improving the statistical performance is consistent with the optimization criteria of minimizing the suboptimality gap in solving the non-convex programming formulation of FCPSLR. (iii) We apply (ii) to the special case of FCPSLR with minimax concave penalty (MCP) and show that under the restricted eigenvalue condition, any S 3 ONC solution with a better objective value than the Lasso solution entails the strong oracle property. In addition, such a solution generates a model error (ME) comparable to the optimal but exponential-time sparse estimator given a sufficient sample size, while the worst-case ME is comparable to the Lasso in general. Furthermore, to guarantee the S 3 ONC admits FPTAS.

Modified conjugate gradient method for diagonalizing large matrices.

PubMed

Jie, Quanlin; Liu, Dunhuan

2003-11-01

We present an iterative method to diagonalize large matrices. The basic idea is the same as the conjugate gradient (CG) method, i.e, minimizing the Rayleigh quotient via its gradient and avoiding reintroducing errors to the directions of previous gradients. Each iteration step is to find lowest eigenvector of the matrix in a subspace spanned by the current trial vector and the corresponding gradient of the Rayleigh quotient, as well as some previous trial vectors. The gradient, together with the previous trial vectors, play a similar role as the conjugate gradient of the original CG algorithm. Our numeric tests indicate that this method converges significantly faster than the original CG method. And the computational cost of one iteration step is about the same as the original CG method. It is suitable for first principle calculations.

Novel palmprint representations for palmprint recognition

NASA Astrophysics Data System (ADS)

Li, Hengjian; Dong, Jiwen; Li, Jinping; Wang, Lei

2015-02-01

In this paper, we propose a novel palmprint recognition algorithm. Firstly, the palmprint images are represented by the anisotropic filter. The filters are built on Gaussian functions along one direction, and on second derivative of Gaussian functions in the orthogonal direction. Also, this choice is motivated by the optimal joint spatial and frequency localization of the Gaussian kernel. Therefore,they can better approximate the edge or line of palmprint images. A palmprint image is processed with a bank of anisotropic filters at different scales and rotations for robust palmprint features extraction. Once these features are extracted, subspace analysis is then applied to the feature vectors for dimension reduction as well as class separability. Experimental results on a public palmprint database show that the accuracy could be improved by the proposed novel representations, compared with Gabor.

An efficient classification method based on principal component and sparse representation.

PubMed

Zhai, Lin; Fu, Shujun; Zhang, Caiming; Liu, Yunxian; Wang, Lu; Liu, Guohua; Yang, Mingqiang

2016-01-01

As an important application in optical imaging, palmprint recognition is interfered by many unfavorable factors. An effective fusion of blockwise bi-directional two-dimensional principal component analysis and grouping sparse classification is presented. The dimension reduction and normalizing are implemented by the blockwise bi-directional two-dimensional principal component analysis for palmprint images to extract feature matrixes, which are assembled into an overcomplete dictionary in sparse classification. A subspace orthogonal matching pursuit algorithm is designed to solve the grouping sparse representation. Finally, the classification result is gained by comparing the residual between testing and reconstructed images. Experiments are carried out on a palmprint database, and the results show that this method has better robustness against position and illumination changes of palmprint images, and can get higher rate of palmprint recognition.

MC ray-tracing optimization of lobster-eye focusing devices with RESTRAX

NASA Astrophysics Data System (ADS)

Šaroun, Jan; Kulda, Jiří

2006-11-01

The enhanced functionalities of the latest version of the RESTRAX software, providing a high-speed Monte Carlo (MC) ray-tracing code to represent a virtual three-axis neutron spectrometer, include representation of parabolic and elliptic guide profiles and facilities for numerical optimization of parameter values, characterizing the instrument components. As examples, we present simulations of a doubly focusing monochromator in combination with cold neutron guides and lobster-eye supermirror devices, concentrating a monochromatic beam to small sample volumes. A Levenberg-Marquardt minimization algorithm is used to optimize simultaneously several parameters of the monochromator and lobster-eye guides. We compare the performance of optimized configurations in terms of monochromatic neutron flux and energy spread and demonstrate the effect of lobster-eye optics on beam transformations in real and momentum subspaces.

Voronoi Diagram Based Optimization of Dynamic Reactive Power Sources

DOE Office of Scientific and Technical Information (OSTI.GOV)

Huang, Weihong; Sun, Kai; Qi, Junjian

2015-01-01

Dynamic var sources can effectively mitigate fault-induced delayed voltage recovery (FIDVR) issues or even voltage collapse. This paper proposes a new approach to optimization of the sizes of dynamic var sources at candidate locations by a Voronoi diagram based algorithm. It first disperses sample points of potential solutions in a searching space, evaluates a cost function at each point by barycentric interpolation for the subspaces around the point, and then constructs a Voronoi diagram about cost function values over the entire space. Accordingly, the final optimal solution can be obtained. Case studies on the WSCC 9-bus system and NPCC 140-busmore » system have validated that the new approach can quickly identify the boundary of feasible solutions in searching space and converge to the global optimal solution.« less

A cancelable biometric scheme based on multi-lead ECGs.

PubMed

Peng-Tzu Chen; Shun-Chi Wu; Jui-Hsuan Hsieh

2017-07-01

Biometric technologies offer great advantages over other recognition methods, but there are concerns that they may compromise the privacy of individuals. In this paper, an electrocardiogram (ECG)-based cancelable biometric scheme is proposed to relieve such concerns. In this scheme, distinct biometric templates for a given beat bundle are constructed via "subspace collapsing." To determine the identity of any unknown beat bundle, the multiple signal classification (MUSIC) algorithm, incorporating a "suppression and poll" strategy, is adopted. Unlike the existing cancelable biometric schemes, knowledge of the distortion transform is not required for recognition. Experiments with real ECGs from 285 subjects are presented to illustrate the efficacy of the proposed scheme. The best recognition rate of 97.58 % was achieved under the test condition N train = 10 and N test = 10.