Deep learning for computational chemistry.

PubMed

Goh, Garrett B; Hodas, Nathan O; Vishnu, Abhinav

2017-06-15

The rise and fall of artificial neural networks is well documented in the scientific literature of both computer science and computational chemistry. Yet almost two decades later, we are now seeing a resurgence of interest in deep learning, a machine learning algorithm based on multilayer neural networks. Within the last few years, we have seen the transformative impact of deep learning in many domains, particularly in speech recognition and computer vision, to the extent that the majority of expert practitioners in those field are now regularly eschewing prior established models in favor of deep learning models. In this review, we provide an introductory overview into the theory of deep neural networks and their unique properties that distinguish them from traditional machine learning algorithms used in cheminformatics. By providing an overview of the variety of emerging applications of deep neural networks, we highlight its ubiquity and broad applicability to a wide range of challenges in the field, including quantitative structure activity relationship, virtual screening, protein structure prediction, quantum chemistry, materials design, and property prediction. In reviewing the performance of deep neural networks, we observed a consistent outperformance against non-neural networks state-of-the-art models across disparate research topics, and deep neural network-based models often exceeded the "glass ceiling" expectations of their respective tasks. Coupled with the maturity of GPU-accelerated computing for training deep neural networks and the exponential growth of chemical data on which to train these networks on, we anticipate that deep learning algorithms will be a valuable tool for computational chemistry. © 2017 Wiley Periodicals, Inc. © 2017 Wiley Periodicals, Inc.

Deep learning for computational chemistry

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

Goh, Garrett B.; Hodas, Nathan O.; Vishnu, Abhinav

The rise and fall of artificial neural networks is well documented in the scientific literature of both the fields of computer science and computational chemistry. Yet almost two decades later, we are now seeing a resurgence of interest in deep learning, a machine learning algorithm based on “deep” neural networks. Within the last few years, we have seen the transformative impact of deep learning the computer science domain, notably in speech recognition and computer vision, to the extent that the majority of practitioners in those field are now regularly eschewing prior established models in favor of deep learning models. Inmore » this review, we provide an introductory overview into the theory of deep neural networks and their unique properties as compared to traditional machine learning algorithms used in cheminformatics. By providing an overview of the variety of emerging applications of deep neural networks, we highlight its ubiquity and broad applicability to a wide range of challenges in the field, including QSAR, virtual screening, protein structure modeling, QM calculations, materials synthesis and property prediction. In reviewing the performance of deep neural networks, we observed a consistent outperformance against non neural networks state-of-the-art models across disparate research topics, and deep neural network based models often exceeded the “glass ceiling” expectations of their respective tasks. Coupled with the maturity of GPU-accelerated computing for training deep neural networks and the exponential growth of chemical data on which to train these networks on, we anticipate that deep learning algorithms will be a useful tool and may grow into a pivotal role for various challenges in the computational chemistry field.« less

Optimal mapping of neural-network learning on message-passing multicomputers

NASA Technical Reports Server (NTRS)

Chu, Lon-Chan; Wah, Benjamin W.

1992-01-01

A minimization of learning-algorithm completion time is sought in the present optimal-mapping study of the learning process in multilayer feed-forward artificial neural networks (ANNs) for message-passing multicomputers. A novel approximation algorithm for mappings of this kind is derived from observations of the dominance of a parallel ANN algorithm over its communication time. Attention is given to both static and dynamic mapping schemes for systems with static and dynamic background workloads, as well as to experimental results obtained for simulated mappings on multicomputers with dynamic background workloads.

Bio-Inspired Neural Model for Learning Dynamic Models

NASA Technical Reports Server (NTRS)

Duong, Tuan; Duong, Vu; Suri, Ronald

2009-01-01

A neural-network mathematical model that, relative to prior such models, places greater emphasis on some of the temporal aspects of real neural physical processes, has been proposed as a basis for massively parallel, distributed algorithms that learn dynamic models of possibly complex external processes by means of learning rules that are local in space and time. The algorithms could be made to perform such functions as recognition and prediction of words in speech and of objects depicted in video images. The approach embodied in this model is said to be "hardware-friendly" in the following sense: The algorithms would be amenable to execution by special-purpose computers implemented as very-large-scale integrated (VLSI) circuits that would operate at relatively high speeds and low power demands.

Application of dynamic recurrent neural networks in nonlinear system identification

NASA Astrophysics Data System (ADS)

Du, Yun; Wu, Xueli; Sun, Huiqin; Zhang, Suying; Tian, Qiang

2006-11-01

An adaptive identification method of simple dynamic recurrent neural network (SRNN) for nonlinear dynamic systems is presented in this paper. This method based on the theory that by using the inner-states feed-back of dynamic network to describe the nonlinear kinetic characteristics of system can reflect the dynamic characteristics more directly, deduces the recursive prediction error (RPE) learning algorithm of SRNN, and improves the algorithm by studying topological structure on recursion layer without the weight values. The simulation results indicate that this kind of neural network can be used in real-time control, due to its less weight values, simpler learning algorithm, higher identification speed, and higher precision of model. It solves the problems of intricate in training algorithm and slow rate in convergence caused by the complicate topological structure in usual dynamic recurrent neural network.

Learning control of inverted pendulum system by neural network driven fuzzy reasoning: The learning function of NN-driven fuzzy reasoning under changes of reasoning environment

NASA Technical Reports Server (NTRS)

Hayashi, Isao; Nomura, Hiroyoshi; Wakami, Noboru

1991-01-01

Whereas conventional fuzzy reasonings are associated with tuning problems, which are lack of membership functions and inference rule designs, a neural network driven fuzzy reasoning (NDF) capable of determining membership functions by neural network is formulated. In the antecedent parts of the neural network driven fuzzy reasoning, the optimum membership function is determined by a neural network, while in the consequent parts, an amount of control for each rule is determined by other plural neural networks. By introducing an algorithm of neural network driven fuzzy reasoning, inference rules for making a pendulum stand up from its lowest suspended point are determined for verifying the usefulness of the algorithm.

Comparative Analysis of Soft Computing Models in Prediction of Bending Rigidity of Cotton Woven Fabrics

NASA Astrophysics Data System (ADS)

Guruprasad, R.; Behera, B. K.

2015-10-01

Quantitative prediction of fabric mechanical properties is an essential requirement for design engineering of textile and apparel products. In this work, the possibility of prediction of bending rigidity of cotton woven fabrics has been explored with the application of Artificial Neural Network (ANN) and two hybrid methodologies, namely Neuro-genetic modeling and Adaptive Neuro-Fuzzy Inference System (ANFIS) modeling. For this purpose, a set of cotton woven grey fabrics was desized, scoured and relaxed. The fabrics were then conditioned and tested for bending properties. With the database thus created, a neural network model was first developed using back propagation as the learning algorithm. The second model was developed by applying a hybrid learning strategy, in which genetic algorithm was first used as a learning algorithm to optimize the number of neurons and connection weights of the neural network. The Genetic algorithm optimized network structure was further allowed to learn using back propagation algorithm. In the third model, an ANFIS modeling approach was attempted to map the input-output data. The prediction performances of the models were compared and a sensitivity analysis was reported. The results show that the prediction by neuro-genetic and ANFIS models were better in comparison with that of back propagation neural network model.

Towards representation of a perceptual color manifold using associative memory for color constancy.

PubMed

Seow, Ming-Jung; Asari, Vijayan K

2009-01-01

In this paper, we propose the concept of a manifold of color perception through empirical observation that the center-surround properties of images in a perceptually similar environment define a manifold in the high dimensional space. Such a manifold representation can be learned using a novel recurrent neural network based learning algorithm. Unlike the conventional recurrent neural network model in which the memory is stored in an attractive fixed point at discrete locations in the state space, the dynamics of the proposed learning algorithm represent memory as a nonlinear line of attraction. The region of convergence around the nonlinear line is defined by the statistical characteristics of the training data. This learned manifold can then be used as a basis for color correction of the images having different color perception to the learned color perception. Experimental results show that the proposed recurrent neural network learning algorithm is capable of color balance the lighting variations in images captured in different environments successfully.

Single-hidden-layer feed-forward quantum neural network based on Grover learning.

PubMed

Liu, Cheng-Yi; Chen, Chein; Chang, Ching-Ter; Shih, Lun-Min

2013-09-01

In this paper, a novel single-hidden-layer feed-forward quantum neural network model is proposed based on some concepts and principles in the quantum theory. By combining the quantum mechanism with the feed-forward neural network, we defined quantum hidden neurons and connected quantum weights, and used them as the fundamental information processing unit in a single-hidden-layer feed-forward neural network. The quantum neurons make a wide range of nonlinear functions serve as the activation functions in the hidden layer of the network, and the Grover searching algorithm outstands the optimal parameter setting iteratively and thus makes very efficient neural network learning possible. The quantum neuron and weights, along with a Grover searching algorithm based learning, result in a novel and efficient neural network characteristic of reduced network, high efficient training and prospect application in future. Some simulations are taken to investigate the performance of the proposed quantum network and the result show that it can achieve accurate learning. Copyright © 2013 Elsevier Ltd. All rights reserved.

LiteNet: Lightweight Neural Network for Detecting Arrhythmias at Resource-Constrained Mobile Devices.

PubMed

He, Ziyang; Zhang, Xiaoqing; Cao, Yangjie; Liu, Zhi; Zhang, Bo; Wang, Xiaoyan

2018-04-17

By running applications and services closer to the user, edge processing provides many advantages, such as short response time and reduced network traffic. Deep-learning based algorithms provide significantly better performances than traditional algorithms in many fields but demand more resources, such as higher computational power and more memory. Hence, designing deep learning algorithms that are more suitable for resource-constrained mobile devices is vital. In this paper, we build a lightweight neural network, termed LiteNet which uses a deep learning algorithm design to diagnose arrhythmias, as an example to show how we design deep learning schemes for resource-constrained mobile devices. Compare to other deep learning models with an equivalent accuracy, LiteNet has several advantages. It requires less memory, incurs lower computational cost, and is more feasible for deployment on resource-constrained mobile devices. It can be trained faster than other neural network algorithms and requires less communication across different processing units during distributed training. It uses filters of heterogeneous size in a convolutional layer, which contributes to the generation of various feature maps. The algorithm was tested using the MIT-BIH electrocardiogram (ECG) arrhythmia database; the results showed that LiteNet outperforms comparable schemes in diagnosing arrhythmias, and in its feasibility for use at the mobile devices.

LiteNet: Lightweight Neural Network for Detecting Arrhythmias at Resource-Constrained Mobile Devices

PubMed Central

Zhang, Xiaoqing; Cao, Yangjie; Liu, Zhi; Zhang, Bo; Wang, Xiaoyan

2018-01-01

By running applications and services closer to the user, edge processing provides many advantages, such as short response time and reduced network traffic. Deep-learning based algorithms provide significantly better performances than traditional algorithms in many fields but demand more resources, such as higher computational power and more memory. Hence, designing deep learning algorithms that are more suitable for resource-constrained mobile devices is vital. In this paper, we build a lightweight neural network, termed LiteNet which uses a deep learning algorithm design to diagnose arrhythmias, as an example to show how we design deep learning schemes for resource-constrained mobile devices. Compare to other deep learning models with an equivalent accuracy, LiteNet has several advantages. It requires less memory, incurs lower computational cost, and is more feasible for deployment on resource-constrained mobile devices. It can be trained faster than other neural network algorithms and requires less communication across different processing units during distributed training. It uses filters of heterogeneous size in a convolutional layer, which contributes to the generation of various feature maps. The algorithm was tested using the MIT-BIH electrocardiogram (ECG) arrhythmia database; the results showed that LiteNet outperforms comparable schemes in diagnosing arrhythmias, and in its feasibility for use at the mobile devices. PMID:29673171

Utilising reinforcement learning to develop strategies for driving auditory neural implants.

PubMed

Lee, Geoffrey W; Zambetta, Fabio; Li, Xiaodong; Paolini, Antonio G

2016-08-01

In this paper we propose a novel application of reinforcement learning to the area of auditory neural stimulation. We aim to develop a simulation environment which is based off real neurological responses to auditory and electrical stimulation in the cochlear nucleus (CN) and inferior colliculus (IC) of an animal model. Using this simulator we implement closed loop reinforcement learning algorithms to determine which methods are most effective at learning effective acoustic neural stimulation strategies. By recording a comprehensive set of acoustic frequency presentations and neural responses from a set of animals we created a large database of neural responses to acoustic stimulation. Extensive electrical stimulation in the CN and the recording of neural responses in the IC provides a mapping of how the auditory system responds to electrical stimuli. The combined dataset is used as the foundation for the simulator, which is used to implement and test learning algorithms. Reinforcement learning, utilising a modified n-Armed Bandit solution, is implemented to demonstrate the model's function. We show the ability to effectively learn stimulation patterns which mimic the cochlea's ability to covert acoustic frequencies to neural activity. Time taken to learn effective replication using neural stimulation takes less than 20 min under continuous testing. These results show the utility of reinforcement learning in the field of neural stimulation. These results can be coupled with existing sound processing technologies to develop new auditory prosthetics that are adaptable to the recipients current auditory pathway. The same process can theoretically be abstracted to other sensory and motor systems to develop similar electrical replication of neural signals.

Genetic algorithm for neural networks optimization

NASA Astrophysics Data System (ADS)

Setyawati, Bina R.; Creese, Robert C.; Sahirman, Sidharta

2004-11-01

This paper examines the forecasting performance of multi-layer feed forward neural networks in modeling a particular foreign exchange rates, i.e. Japanese Yen/US Dollar. The effects of two learning methods, Back Propagation and Genetic Algorithm, in which the neural network topology and other parameters fixed, were investigated. The early results indicate that the application of this hybrid system seems to be well suited for the forecasting of foreign exchange rates. The Neural Networks and Genetic Algorithm were programmed using MATLAB«.

Neural model of gene regulatory network: a survey on supportive meta-heuristics.

PubMed

Biswas, Surama; Acharyya, Sriyankar

2016-06-01

Gene regulatory network (GRN) is produced as a result of regulatory interactions between different genes through their coded proteins in cellular context. Having immense importance in disease detection and drug finding, GRN has been modelled through various mathematical and computational schemes and reported in survey articles. Neural and neuro-fuzzy models have been the focus of attraction in bioinformatics. Predominant use of meta-heuristic algorithms in training neural models has proved its excellence. Considering these facts, this paper is organized to survey neural modelling schemes of GRN and the efficacy of meta-heuristic algorithms towards parameter learning (i.e. weighting connections) within the model. This survey paper renders two different structure-related approaches to infer GRN which are global structure approach and substructure approach. It also describes two neural modelling schemes, such as artificial neural network/recurrent neural network based modelling and neuro-fuzzy modelling. The meta-heuristic algorithms applied so far to learn the structure and parameters of neutrally modelled GRN have been reviewed here.

Learning Perfectly Secure Cryptography to Protect Communications with Adversarial Neural Cryptography

PubMed Central

2018-01-01

Researches in Artificial Intelligence (AI) have achieved many important breakthroughs, especially in recent years. In some cases, AI learns alone from scratch and performs human tasks faster and better than humans. With the recent advances in AI, it is natural to wonder whether Artificial Neural Networks will be used to successfully create or break cryptographic algorithms. Bibliographic review shows the main approach to this problem have been addressed throughout complex Neural Networks, but without understanding or proving the security of the generated model. This paper presents an analysis of the security of cryptographic algorithms generated by a new technique called Adversarial Neural Cryptography (ANC). Using the proposed network, we show limitations and directions to improve the current approach of ANC. Training the proposed Artificial Neural Network with the improved model of ANC, we show that artificially intelligent agents can learn the unbreakable One-Time Pad (OTP) algorithm, without human knowledge, to communicate securely through an insecure communication channel. This paper shows in which conditions an AI agent can learn a secure encryption scheme. However, it also shows that, without a stronger adversary, it is more likely to obtain an insecure one. PMID:29695066

Learning Perfectly Secure Cryptography to Protect Communications with Adversarial Neural Cryptography.

PubMed

Coutinho, Murilo; de Oliveira Albuquerque, Robson; Borges, Fábio; García Villalba, Luis Javier; Kim, Tai-Hoon

2018-04-24

Researches in Artificial Intelligence (AI) have achieved many important breakthroughs, especially in recent years. In some cases, AI learns alone from scratch and performs human tasks faster and better than humans. With the recent advances in AI, it is natural to wonder whether Artificial Neural Networks will be used to successfully create or break cryptographic algorithms. Bibliographic review shows the main approach to this problem have been addressed throughout complex Neural Networks, but without understanding or proving the security of the generated model. This paper presents an analysis of the security of cryptographic algorithms generated by a new technique called Adversarial Neural Cryptography (ANC). Using the proposed network, we show limitations and directions to improve the current approach of ANC. Training the proposed Artificial Neural Network with the improved model of ANC, we show that artificially intelligent agents can learn the unbreakable One-Time Pad (OTP) algorithm, without human knowledge, to communicate securely through an insecure communication channel. This paper shows in which conditions an AI agent can learn a secure encryption scheme. However, it also shows that, without a stronger adversary, it is more likely to obtain an insecure one.

On the fusion of tuning parameters of fuzzy rules and neural network

NASA Astrophysics Data System (ADS)

Mamuda, Mamman; Sathasivam, Saratha

2017-08-01

Learning fuzzy rule-based system with neural network can lead to a precise valuable empathy of several problems. Fuzzy logic offers a simple way to reach at a definite conclusion based upon its vague, ambiguous, imprecise, noisy or missing input information. Conventional learning algorithm for tuning parameters of fuzzy rules using training input-output data usually end in a weak firing state, this certainly powers the fuzzy rule and makes it insecure for a multiple-input fuzzy system. In this paper, we introduce a new learning algorithm for tuning the parameters of the fuzzy rules alongside with radial basis function neural network (RBFNN) in training input-output data based on the gradient descent method. By the new learning algorithm, the problem of weak firing using the conventional method was addressed. We illustrated the efficiency of our new learning algorithm by means of numerical examples. MATLAB R2014(a) software was used in simulating our result The result shows that the new learning method has the best advantage of training the fuzzy rules without tempering with the fuzzy rule table which allowed a membership function of the rule to be used more than one time in the fuzzy rule base.

Estimation of State Transition Probabilities: A Neural Network Model

NASA Astrophysics Data System (ADS)

Saito, Hiroshi; Takiyama, Ken; Okada, Masato

2015-12-01

Humans and animals can predict future states on the basis of acquired knowledge. This prediction of the state transition is important for choosing the best action, and the prediction is only possible if the state transition probability has already been learned. However, how our brains learn the state transition probability is unknown. Here, we propose a simple algorithm for estimating the state transition probability by utilizing the state prediction error. We analytically and numerically confirmed that our algorithm is able to learn the probability completely with an appropriate learning rate. Furthermore, our learning rule reproduced experimentally reported psychometric functions and neural activities in the lateral intraparietal area in a decision-making task. Thus, our algorithm might describe the manner in which our brains learn state transition probabilities and predict future states.

Time-oriented hierarchical method for computation of principal components using subspace learning algorithm.

PubMed

Jankovic, Marko; Ogawa, Hidemitsu

2004-10-01

Principal Component Analysis (PCA) and Principal Subspace Analysis (PSA) are classic techniques in statistical data analysis, feature extraction and data compression. Given a set of multivariate measurements, PCA and PSA provide a smaller set of "basis vectors" with less redundancy, and a subspace spanned by them, respectively. Artificial neurons and neural networks have been shown to perform PSA and PCA when gradient ascent (descent) learning rules are used, which is related to the constrained maximization (minimization) of statistical objective functions. Due to their low complexity, such algorithms and their implementation in neural networks are potentially useful in cases of tracking slow changes of correlations in the input data or in updating eigenvectors with new samples. In this paper we propose PCA learning algorithm that is fully homogeneous with respect to neurons. The algorithm is obtained by modification of one of the most famous PSA learning algorithms--Subspace Learning Algorithm (SLA). Modification of the algorithm is based on Time-Oriented Hierarchical Method (TOHM). The method uses two distinct time scales. On a faster time scale PSA algorithm is responsible for the "behavior" of all output neurons. On a slower scale, output neurons will compete for fulfillment of their "own interests". On this scale, basis vectors in the principal subspace are rotated toward the principal eigenvectors. At the end of the paper it will be briefly analyzed how (or why) time-oriented hierarchical method can be used for transformation of any of the existing neural network PSA method, into PCA method.

Cardiac Arrhythmia Classification by Multi-Layer Perceptron and Convolution Neural Networks.

PubMed

Savalia, Shalin; Emamian, Vahid

2018-05-04

The electrocardiogram (ECG) plays an imperative role in the medical field, as it records heart signal over time and is used to discover numerous cardiovascular diseases. If a documented ECG signal has a certain irregularity in its predefined features, this is called arrhythmia, the types of which include tachycardia, bradycardia, supraventricular arrhythmias, and ventricular, etc. This has encouraged us to do research that consists of distinguishing between several arrhythmias by using deep neural network algorithms such as multi-layer perceptron (MLP) and convolution neural network (CNN). The TensorFlow library that was established by Google for deep learning and machine learning is used in python to acquire the algorithms proposed here. The ECG databases accessible at PhysioBank.com and kaggle.com were used for training, testing, and validation of the MLP and CNN algorithms. The proposed algorithm consists of four hidden layers with weights, biases in MLP, and four-layer convolution neural networks which map ECG samples to the different classes of arrhythmia. The accuracy of the algorithm surpasses the performance of the current algorithms that have been developed by other cardiologists in both sensitivity and precision.

Optimizing the learning rate for adaptive estimation of neural encoding models

PubMed Central

2018-01-01

Closed-loop neurotechnologies often need to adaptively learn an encoding model that relates the neural activity to the brain state, and is used for brain state decoding. The speed and accuracy of adaptive learning algorithms are critically affected by the learning rate, which dictates how fast model parameters are updated based on new observations. Despite the importance of the learning rate, currently an analytical approach for its selection is largely lacking and existing signal processing methods vastly tune it empirically or heuristically. Here, we develop a novel analytical calibration algorithm for optimal selection of the learning rate in adaptive Bayesian filters. We formulate the problem through a fundamental trade-off that learning rate introduces between the steady-state error and the convergence time of the estimated model parameters. We derive explicit functions that predict the effect of learning rate on error and convergence time. Using these functions, our calibration algorithm can keep the steady-state parameter error covariance smaller than a desired upper-bound while minimizing the convergence time, or keep the convergence time faster than a desired value while minimizing the error. We derive the algorithm both for discrete-valued spikes modeled as point processes nonlinearly dependent on the brain state, and for continuous-valued neural recordings modeled as Gaussian processes linearly dependent on the brain state. Using extensive closed-loop simulations, we show that the analytical solution of the calibration algorithm accurately predicts the effect of learning rate on parameter error and convergence time. Moreover, the calibration algorithm allows for fast and accurate learning of the encoding model and for fast convergence of decoding to accurate performance. Finally, larger learning rates result in inaccurate encoding models and decoders, and smaller learning rates delay their convergence. The calibration algorithm provides a novel analytical approach to predictably achieve a desired level of error and convergence time in adaptive learning, with application to closed-loop neurotechnologies and other signal processing domains. PMID:29813069

Optimizing the learning rate for adaptive estimation of neural encoding models.

PubMed

Hsieh, Han-Lin; Shanechi, Maryam M

2018-05-01

Closed-loop neurotechnologies often need to adaptively learn an encoding model that relates the neural activity to the brain state, and is used for brain state decoding. The speed and accuracy of adaptive learning algorithms are critically affected by the learning rate, which dictates how fast model parameters are updated based on new observations. Despite the importance of the learning rate, currently an analytical approach for its selection is largely lacking and existing signal processing methods vastly tune it empirically or heuristically. Here, we develop a novel analytical calibration algorithm for optimal selection of the learning rate in adaptive Bayesian filters. We formulate the problem through a fundamental trade-off that learning rate introduces between the steady-state error and the convergence time of the estimated model parameters. We derive explicit functions that predict the effect of learning rate on error and convergence time. Using these functions, our calibration algorithm can keep the steady-state parameter error covariance smaller than a desired upper-bound while minimizing the convergence time, or keep the convergence time faster than a desired value while minimizing the error. We derive the algorithm both for discrete-valued spikes modeled as point processes nonlinearly dependent on the brain state, and for continuous-valued neural recordings modeled as Gaussian processes linearly dependent on the brain state. Using extensive closed-loop simulations, we show that the analytical solution of the calibration algorithm accurately predicts the effect of learning rate on parameter error and convergence time. Moreover, the calibration algorithm allows for fast and accurate learning of the encoding model and for fast convergence of decoding to accurate performance. Finally, larger learning rates result in inaccurate encoding models and decoders, and smaller learning rates delay their convergence. The calibration algorithm provides a novel analytical approach to predictably achieve a desired level of error and convergence time in adaptive learning, with application to closed-loop neurotechnologies and other signal processing domains.

On the improvement of neural cryptography using erroneous transmitted information with error prediction.

PubMed

Allam, Ahmed M; Abbas, Hazem M

2010-12-01

Neural cryptography deals with the problem of "key exchange" between two neural networks using the mutual learning concept. The two networks exchange their outputs (in bits) and the key between the two communicating parties is eventually represented in the final learned weights, when the two networks are said to be synchronized. Security of neural synchronization is put at risk if an attacker is capable of synchronizing with any of the two parties during the training process. Therefore, diminishing the probability of such a threat improves the reliability of exchanging the output bits through a public channel. The synchronization with feedback algorithm is one of the existing algorithms that enhances the security of neural cryptography. This paper proposes three new algorithms to enhance the mutual learning process. They mainly depend on disrupting the attacker confidence in the exchanged outputs and input patterns during training. The first algorithm is called "Do not Trust My Partner" (DTMP), which relies on one party sending erroneous output bits, with the other party being capable of predicting and correcting this error. The second algorithm is called "Synchronization with Common Secret Feedback" (SCSFB), where inputs are kept partially secret and the attacker has to train its network on input patterns that are different from the training sets used by the communicating parties. The third algorithm is a hybrid technique combining the features of the DTMP and SCSFB. The proposed approaches are shown to outperform the synchronization with feedback algorithm in the time needed for the parties to synchronize.

Quick fuzzy backpropagation algorithm.

PubMed

Nikov, A; Stoeva, S

2001-03-01

A modification of the fuzzy backpropagation (FBP) algorithm called QuickFBP algorithm is proposed, where the computation of the net function is significantly quicker. It is proved that the FBP algorithm is of exponential time complexity, while the QuickFBP algorithm is of polynomial time complexity. Convergence conditions of the QuickFBP, resp. the FBP algorithm are defined and proved for: (1) single output neural networks in case of training patterns with different targets; and (2) multiple output neural networks in case of training patterns with equivalued target vector. They support the automation of the weights training process (quasi-unsupervised learning) establishing the target value(s) depending on the network's input values. In these cases the simulation results confirm the convergence of both algorithms. An example with a large-sized neural network illustrates the significantly greater training speed of the QuickFBP rather than the FBP algorithm. The adaptation of an interactive web system to users on the basis of the QuickFBP algorithm is presented. Since the QuickFBP algorithm ensures quasi-unsupervised learning, this implies its broad applicability in areas of adaptive and adaptable interactive systems, data mining, etc. applications.

Pragmatically Framed Cross-Situational Noun Learning Using Computational Reinforcement Models

PubMed Central

Najnin, Shamima; Banerjee, Bonny

2018-01-01

Cross-situational learning and social pragmatic theories are prominent mechanisms for learning word meanings (i.e., word-object pairs). In this paper, the role of reinforcement is investigated for early word-learning by an artificial agent. When exposed to a group of speakers, the agent comes to understand an initial set of vocabulary items belonging to the language used by the group. Both cross-situational learning and social pragmatic theory are taken into account. As social cues, joint attention and prosodic cues in caregiver's speech are considered. During agent-caregiver interaction, the agent selects a word from the caregiver's utterance and learns the relations between that word and the objects in its visual environment. The “novel words to novel objects” language-specific constraint is assumed for computing rewards. The models are learned by maximizing the expected reward using reinforcement learning algorithms [i.e., table-based algorithms: Q-learning, SARSA, SARSA-λ, and neural network-based algorithms: Q-learning for neural network (Q-NN), neural-fitted Q-network (NFQ), and deep Q-network (DQN)]. Neural network-based reinforcement learning models are chosen over table-based models for better generalization and quicker convergence. Simulations are carried out using mother-infant interaction CHILDES dataset for learning word-object pairings. Reinforcement is modeled in two cross-situational learning cases: (1) with joint attention (Attentional models), and (2) with joint attention and prosodic cues (Attentional-prosodic models). Attentional-prosodic models manifest superior performance to Attentional ones for the task of word-learning. The Attentional-prosodic DQN outperforms existing word-learning models for the same task. PMID:29441027

Low-cost autonomous perceptron neural network inspired by quantum computation

NASA Astrophysics Data System (ADS)

Zidan, Mohammed; Abdel-Aty, Abdel-Haleem; El-Sadek, Alaa; Zanaty, E. A.; Abdel-Aty, Mahmoud

2017-11-01

Achieving low cost learning with reliable accuracy is one of the important goals to achieve intelligent machines to save time, energy and perform learning process over limited computational resources machines. In this paper, we propose an efficient algorithm for a perceptron neural network inspired by quantum computing composite from a single neuron to classify inspirable linear applications after a single training iteration O(1). The algorithm is applied over a real world data set and the results are outer performs the other state-of-the art algorithms.

Supervised learning of probability distributions by neural networks

NASA Technical Reports Server (NTRS)

Baum, Eric B.; Wilczek, Frank

1988-01-01

Supervised learning algorithms for feedforward neural networks are investigated analytically. The back-propagation algorithm described by Werbos (1974), Parker (1985), and Rumelhart et al. (1986) is generalized by redefining the values of the input and output neurons as probabilities. The synaptic weights are then varied to follow gradients in the logarithm of likelihood rather than in the error. This modification is shown to provide a more rigorous theoretical basis for the algorithm and to permit more accurate predictions. A typical application involving a medical-diagnosis expert system is discussed.

Automated sleep stage detection with a classical and a neural learning algorithm--methodological aspects.

PubMed

Schwaibold, M; Schöchlin, J; Bolz, A

2002-01-01

For classification tasks in biosignal processing, several strategies and algorithms can be used. Knowledge-based systems allow prior knowledge about the decision process to be integrated, both by the developer and by self-learning capabilities. For the classification stages in a sleep stage detection framework, three inference strategies were compared regarding their specific strengths: a classical signal processing approach, artificial neural networks and neuro-fuzzy systems. Methodological aspects were assessed to attain optimum performance and maximum transparency for the user. Due to their effective and robust learning behavior, artificial neural networks could be recommended for pattern recognition, while neuro-fuzzy systems performed best for the processing of contextual information.

Adaptive control of nonlinear system using online error minimum neural networks.

PubMed

Jia, Chao; Li, Xiaoli; Wang, Kang; Ding, Dawei

2016-11-01

In this paper, a new learning algorithm named OEM-ELM (Online Error Minimized-ELM) is proposed based on ELM (Extreme Learning Machine) neural network algorithm and the spreading of its main structure. The core idea of this OEM-ELM algorithm is: online learning, evaluation of network performance, and increasing of the number of hidden nodes. It combines the advantages of OS-ELM and EM-ELM, which can improve the capability of identification and avoid the redundancy of networks. The adaptive control based on the proposed algorithm OEM-ELM is set up which has stronger adaptive capability to the change of environment. The adaptive control of chemical process Continuous Stirred Tank Reactor (CSTR) is also given for application. The simulation results show that the proposed algorithm with respect to the traditional ELM algorithm can avoid network redundancy and improve the control performance greatly. Copyright © 2016 ISA. Published by Elsevier Ltd. All rights reserved.

Geometrical structure of Neural Networks: Geodesics, Jeffrey's Prior and Hyper-ribbons

NASA Astrophysics Data System (ADS)

Hayden, Lorien; Alemi, Alex; Sethna, James

2014-03-01

Neural networks are learning algorithms which are employed in a host of Machine Learning problems including speech recognition, object classification and data mining. In practice, neural networks learn a low dimensional representation of high dimensional data and define a model manifold which is an embedding of this low dimensional structure in the higher dimensional space. In this work, we explore the geometrical structure of a neural network model manifold. A Stacked Denoising Autoencoder and a Deep Belief Network are trained on handwritten digits from the MNIST database. Construction of geodesics along the surface and of slices taken from the high dimensional manifolds reveal a hierarchy of widths corresponding to a hyper-ribbon structure. This property indicates that neural networks fall into the class of sloppy models, in which certain parameter combinations dominate the behavior. Employing this information could prove valuable in designing both neural network architectures and training algorithms. This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant No . DGE-1144153.

SortNet: learning to rank by a neural preference function.

PubMed

Rigutini, Leonardo; Papini, Tiziano; Maggini, Marco; Scarselli, Franco

2011-09-01

Relevance ranking consists in sorting a set of objects with respect to a given criterion. However, in personalized retrieval systems, the relevance criteria may usually vary among different users and may not be predefined. In this case, ranking algorithms that adapt their behavior from users' feedbacks must be devised. Two main approaches are proposed in the literature for learning to rank: the use of a scoring function, learned by examples, that evaluates a feature-based representation of each object yielding an absolute relevance score, a pairwise approach, where a preference function is learned to determine the object that has to be ranked first in a given pair. In this paper, we present a preference learning method for learning to rank. A neural network, the comparative neural network (CmpNN), is trained from examples to approximate the comparison function for a pair of objects. The CmpNN adopts a particular architecture designed to implement the symmetries naturally present in a preference function. The learned preference function can be embedded as the comparator into a classical sorting algorithm to provide a global ranking of a set of objects. To improve the ranking performances, an active-learning procedure is devised, that aims at selecting the most informative patterns in the training set. The proposed algorithm is evaluated on the LETOR dataset showing promising performances in comparison with other state-of-the-art algorithms.

On adaptive learning rate that guarantees convergence in feedforward networks.

PubMed

Behera, Laxmidhar; Kumar, Swagat; Patnaik, Awhan

2006-09-01

This paper investigates new learning algorithms (LF I and LF II) based on Lyapunov function for the training of feedforward neural networks. It is observed that such algorithms have interesting parallel with the popular backpropagation (BP) algorithm where the fixed learning rate is replaced by an adaptive learning rate computed using convergence theorem based on Lyapunov stability theory. LF II, a modified version of LF I, has been introduced with an aim to avoid local minima. This modification also helps in improving the convergence speed in some cases. Conditions for achieving global minimum for these kind of algorithms have been studied in detail. The performances of the proposed algorithms are compared with BP algorithm and extended Kalman filtering (EKF) on three bench-mark function approximation problems: XOR, 3-bit parity, and 8-3 encoder. The comparisons are made in terms of number of learning iterations and computational time required for convergence. It is found that the proposed algorithms (LF I and II) are much faster in convergence than other two algorithms to attain same accuracy. Finally, the comparison is made on a complex two-dimensional (2-D) Gabor function and effect of adaptive learning rate for faster convergence is verified. In a nutshell, the investigations made in this paper help us better understand the learning procedure of feedforward neural networks in terms of adaptive learning rate, convergence speed, and local minima.

Analog Delta-Back-Propagation Neural-Network Circuitry

NASA Technical Reports Server (NTRS)

Eberhart, Silvio

1990-01-01

Changes in synapse weights due to circuit drifts suppressed. Proposed fully parallel analog version of electronic neural-network processor based on delta-back-propagation algorithm. Processor able to "learn" when provided with suitable combinations of inputs and enforced outputs. Includes programmable resistive memory elements (corresponding to synapses), conductances (synapse weights) adjusted during learning. Buffer amplifiers, summing circuits, and sample-and-hold circuits arranged in layers of electronic neurons in accordance with delta-back-propagation algorithm.

Least square neural network model of the crude oil blending process.

PubMed

Rubio, José de Jesús

2016-06-01

In this paper, the recursive least square algorithm is designed for the big data learning of a feedforward neural network. The proposed method as the combination of the recursive least square and feedforward neural network obtains four advantages over the alone algorithms: it requires less number of regressors, it is fast, it has the learning ability, and it is more compact. Stability, convergence, boundedness of parameters, and local minimum avoidance of the proposed technique are guaranteed. The introduced strategy is applied for the modeling of the crude oil blending process. Copyright © 2016 Elsevier Ltd. All rights reserved.

A globally convergent MC algorithm with an adaptive learning rate.

PubMed

Peng, Dezhong; Yi, Zhang; Xiang, Yong; Zhang, Haixian

2012-02-01

This brief deals with the problem of minor component analysis (MCA). Artificial neural networks can be exploited to achieve the task of MCA. Recent research works show that convergence of neural networks based MCA algorithms can be guaranteed if the learning rates are less than certain thresholds. However, the computation of these thresholds needs information about the eigenvalues of the autocorrelation matrix of data set, which is unavailable in online extraction of minor component from input data stream. In this correspondence, we introduce an adaptive learning rate into the OJAn MCA algorithm, such that its convergence condition does not depend on any unobtainable information, and can be easily satisfied in practical applications.

Efficient Training of Supervised Spiking Neural Network via Accurate Synaptic-Efficiency Adjustment Method.

PubMed

Xie, Xiurui; Qu, Hong; Yi, Zhang; Kurths, Jurgen

2017-06-01

The spiking neural network (SNN) is the third generation of neural networks and performs remarkably well in cognitive tasks, such as pattern recognition. The temporal neural encode mechanism found in biological hippocampus enables SNN to possess more powerful computation capability than networks with other encoding schemes. However, this temporal encoding approach requires neurons to process information serially on time, which reduces learning efficiency significantly. To keep the powerful computation capability of the temporal encoding mechanism and to overcome its low efficiency in the training of SNNs, a new training algorithm, the accurate synaptic-efficiency adjustment method is proposed in this paper. Inspired by the selective attention mechanism of the primate visual system, our algorithm selects only the target spike time as attention areas, and ignores voltage states of the untarget ones, resulting in a significant reduction of training time. Besides, our algorithm employs a cost function based on the voltage difference between the potential of the output neuron and the firing threshold of the SNN, instead of the traditional precise firing time distance. A normalized spike-timing-dependent-plasticity learning window is applied to assigning this error to different synapses for instructing their training. Comprehensive simulations are conducted to investigate the learning properties of our algorithm, with input neurons emitting both single spike and multiple spikes. Simulation results indicate that our algorithm possesses higher learning performance than the existing other methods and achieves the state-of-the-art efficiency in the training of SNN.

Selected Flight Test Results for Online Learning Neural Network-Based Flight Control System

NASA Technical Reports Server (NTRS)

Williams-Hayes, Peggy S.

2004-01-01

The NASA F-15 Intelligent Flight Control System project team developed a series of flight control concepts designed to demonstrate neural network-based adaptive controller benefits, with the objective to develop and flight-test control systems using neural network technology to optimize aircraft performance under nominal conditions and stabilize the aircraft under failure conditions. This report presents flight-test results for an adaptive controller using stability and control derivative values from an online learning neural network. A dynamic cell structure neural network is used in conjunction with a real-time parameter identification algorithm to estimate aerodynamic stability and control derivative increments to baseline aerodynamic derivatives in flight. This open-loop flight test set was performed in preparation for a future phase in which the learning neural network and parameter identification algorithm output would provide the flight controller with aerodynamic stability and control derivative updates in near real time. Two flight maneuvers are analyzed - pitch frequency sweep and automated flight-test maneuver designed to optimally excite the parameter identification algorithm in all axes. Frequency responses generated from flight data are compared to those obtained from nonlinear simulation runs. Flight data examination shows that addition of flight-identified aerodynamic derivative increments into the simulation improved aircraft pitch handling qualities.

The graph neural network model.

PubMed

Scarselli, Franco; Gori, Marco; Tsoi, Ah Chung; Hagenbuchner, Markus; Monfardini, Gabriele

2009-01-01

Many underlying relationships among data in several areas of science and engineering, e.g., computer vision, molecular chemistry, molecular biology, pattern recognition, and data mining, can be represented in terms of graphs. In this paper, we propose a new neural network model, called graph neural network (GNN) model, that extends existing neural network methods for processing the data represented in graph domains. This GNN model, which can directly process most of the practically useful types of graphs, e.g., acyclic, cyclic, directed, and undirected, implements a function tau(G,n) is an element of IR(m) that maps a graph G and one of its nodes n into an m-dimensional Euclidean space. A supervised learning algorithm is derived to estimate the parameters of the proposed GNN model. The computational cost of the proposed algorithm is also considered. Some experimental results are shown to validate the proposed learning algorithm, and to demonstrate its generalization capabilities.

A Deep Learning Algorithm of Neural Network for the Parameterization of Typhoon-Ocean Feedback in Typhoon Forecast Models

NASA Astrophysics Data System (ADS)

Jiang, Guo-Qing; Xu, Jing; Wei, Jun

2018-04-01

Two algorithms based on machine learning neural networks are proposed—the shallow learning (S-L) and deep learning (D-L) algorithms—that can potentially be used in atmosphere-only typhoon forecast models to provide flow-dependent typhoon-induced sea surface temperature cooling (SSTC) for improving typhoon predictions. The major challenge of existing SSTC algorithms in forecast models is how to accurately predict SSTC induced by an upcoming typhoon, which requires information not only from historical data but more importantly also from the target typhoon itself. The S-L algorithm composes of a single layer of neurons with mixed atmospheric and oceanic factors. Such a structure is found to be unable to represent correctly the physical typhoon-ocean interaction. It tends to produce an unstable SSTC distribution, for which any perturbations may lead to changes in both SSTC pattern and strength. The D-L algorithm extends the neural network to a 4 × 5 neuron matrix with atmospheric and oceanic factors being separated in different layers of neurons, so that the machine learning can determine the roles of atmospheric and oceanic factors in shaping the SSTC. Therefore, it produces a stable crescent-shaped SSTC distribution, with its large-scale pattern determined mainly by atmospheric factors (e.g., winds) and small-scale features by oceanic factors (e.g., eddies). Sensitivity experiments reveal that the D-L algorithms improve maximum wind intensity errors by 60-70% for four case study simulations, compared to their atmosphere-only model runs.

Machine Learning and Quantum Mechanics

NASA Astrophysics Data System (ADS)

Chapline, George

The author has previously pointed out some similarities between selforganizing neural networks and quantum mechanics. These types of neural networks were originally conceived of as away of emulating the cognitive capabilities of the human brain. Recently extensions of these networks, collectively referred to as deep learning networks, have strengthened the connection between self-organizing neural networks and human cognitive capabilities. In this note we consider whether hardware quantum devices might be useful for emulating neural networks with human-like cognitive capabilities, or alternatively whether implementations of deep learning neural networks using conventional computers might lead to better algorithms for solving the many body Schrodinger equation.

Modeling development of natural multi-sensory integration using neural self-organisation and probabilistic population codes

NASA Astrophysics Data System (ADS)

Bauer, Johannes; Dávila-Chacón, Jorge; Wermter, Stefan

2015-10-01

Humans and other animals have been shown to perform near-optimally in multi-sensory integration tasks. Probabilistic population codes (PPCs) have been proposed as a mechanism by which optimal integration can be accomplished. Previous approaches have focussed on how neural networks might produce PPCs from sensory input or perform calculations using them, like combining multiple PPCs. Less attention has been given to the question of how the necessary organisation of neurons can arise and how the required knowledge about the input statistics can be learned. In this paper, we propose a model of learning multi-sensory integration based on an unsupervised learning algorithm in which an artificial neural network learns the noise characteristics of each of its sources of input. Our algorithm borrows from the self-organising map the ability to learn latent-variable models of the input and extends it to learning to produce a PPC approximating a probability density function over the latent variable behind its (noisy) input. The neurons in our network are only required to perform simple calculations and we make few assumptions about input noise properties and tuning functions. We report on a neurorobotic experiment in which we apply our algorithm to multi-sensory integration in a humanoid robot to demonstrate its effectiveness and compare it to human multi-sensory integration on the behavioural level. We also show in simulations that our algorithm performs near-optimally under certain plausible conditions, and that it reproduces important aspects of natural multi-sensory integration on the neural level.

Training Feedforward Neural Networks Using Symbiotic Organisms Search Algorithm.

PubMed

Wu, Haizhou; Zhou, Yongquan; Luo, Qifang; Basset, Mohamed Abdel

2016-01-01

Symbiotic organisms search (SOS) is a new robust and powerful metaheuristic algorithm, which stimulates the symbiotic interaction strategies adopted by organisms to survive and propagate in the ecosystem. In the supervised learning area, it is a challenging task to present a satisfactory and efficient training algorithm for feedforward neural networks (FNNs). In this paper, SOS is employed as a new method for training FNNs. To investigate the performance of the aforementioned method, eight different datasets selected from the UCI machine learning repository are employed for experiment and the results are compared among seven metaheuristic algorithms. The results show that SOS performs better than other algorithms for training FNNs in terms of converging speed. It is also proven that an FNN trained by the method of SOS has better accuracy than most algorithms compared.

The Behavioral and Neural Mechanisms Underlying the Tracking of Expertise

PubMed Central

Boorman, Erie D.; O’Doherty, John P.; Adolphs, Ralph; Rangel, Antonio

2013-01-01

Summary Evaluating the abilities of others is fundamental for successful economic and social behavior. We investigated the computational and neurobiological basis of ability tracking by designing an fMRI task that required participants to use and update estimates of both people and algorithms’ expertise through observation of their predictions. Behaviorally, we find a model-based algorithm characterized subject predictions better than several alternative models. Notably, when the agent’s prediction was concordant rather than discordant with the subject’s own likely prediction, participants credited people more than algorithms for correct predictions and penalized them less for incorrect predictions. Neurally, many components of the mentalizing network—medial prefrontal cortex, anterior cingulate gyrus, temporoparietal junction, and precuneus—represented or updated expertise beliefs about both people and algorithms. Moreover, activity in lateral orbitofrontal and medial prefrontal cortex reflected behavioral differences in learning about people and algorithms. These findings provide basic insights into the neural basis of social learning. PMID:24360551

Relabeling exchange method (REM) for learning in neural networks

NASA Astrophysics Data System (ADS)

Wu, Wen; Mammone, Richard J.

1994-02-01

The supervised training of neural networks require the use of output labels which are usually arbitrarily assigned. In this paper it is shown that there is a significant difference in the rms error of learning when `optimal' label assignment schemes are used. We have investigated two efficient random search algorithms to solve the relabeling problem: the simulated annealing and the genetic algorithm. However, we found them to be computationally expensive. Therefore we shall introduce a new heuristic algorithm called the Relabeling Exchange Method (REM) which is computationally more attractive and produces optimal performance. REM has been used to organize the optimal structure for multi-layered perceptrons and neural tree networks. The method is a general one and can be implemented as a modification to standard training algorithms. The motivation of the new relabeling strategy is based on the present interpretation of dyslexia as an encoding problem.

Self Improving Methods for Materials and Process Design

DTIC Science & Technology

1998-08-31

using inductive coupling techniques. The first phase of the work focuses on developing an artificial neural network learning for function approximation...developing an artificial neural network learning algorithm for time-series prediction. The third phase of the work focuses on model selection. We have

Mastering the game of Go without human knowledge.

PubMed

Silver, David; Schrittwieser, Julian; Simonyan, Karen; Antonoglou, Ioannis; Huang, Aja; Guez, Arthur; Hubert, Thomas; Baker, Lucas; Lai, Matthew; Bolton, Adrian; Chen, Yutian; Lillicrap, Timothy; Hui, Fan; Sifre, Laurent; van den Driessche, George; Graepel, Thore; Hassabis, Demis

2017-10-18

A long-standing goal of artificial intelligence is an algorithm that learns, tabula rasa, superhuman proficiency in challenging domains. Recently, AlphaGo became the first program to defeat a world champion in the game of Go. The tree search in AlphaGo evaluated positions and selected moves using deep neural networks. These neural networks were trained by supervised learning from human expert moves, and by reinforcement learning from self-play. Here we introduce an algorithm based solely on reinforcement learning, without human data, guidance or domain knowledge beyond game rules. AlphaGo becomes its own teacher: a neural network is trained to predict AlphaGo's own move selections and also the winner of AlphaGo's games. This neural network improves the strength of the tree search, resulting in higher quality move selection and stronger self-play in the next iteration. Starting tabula rasa, our new program AlphaGo Zero achieved superhuman performance, winning 100-0 against the previously published, champion-defeating AlphaGo.

Mastering the game of Go without human knowledge

NASA Astrophysics Data System (ADS)

Silver, David; Schrittwieser, Julian; Simonyan, Karen; Antonoglou, Ioannis; Huang, Aja; Guez, Arthur; Hubert, Thomas; Baker, Lucas; Lai, Matthew; Bolton, Adrian; Chen, Yutian; Lillicrap, Timothy; Hui, Fan; Sifre, Laurent; van den Driessche, George; Graepel, Thore; Hassabis, Demis

2017-10-01

A long-standing goal of artificial intelligence is an algorithm that learns, tabula rasa, superhuman proficiency in challenging domains. Recently, AlphaGo became the first program to defeat a world champion in the game of Go. The tree search in AlphaGo evaluated positions and selected moves using deep neural networks. These neural networks were trained by supervised learning from human expert moves, and by reinforcement learning from self-play. Here we introduce an algorithm based solely on reinforcement learning, without human data, guidance or domain knowledge beyond game rules. AlphaGo becomes its own teacher: a neural network is trained to predict AlphaGo’s own move selections and also the winner of AlphaGo’s games. This neural network improves the strength of the tree search, resulting in higher quality move selection and stronger self-play in the next iteration. Starting tabula rasa, our new program AlphaGo Zero achieved superhuman performance, winning 100-0 against the previously published, champion-defeating AlphaGo.

Resolution of Singularities Introduced by Hierarchical Structure in Deep Neural Networks.

PubMed

Nitta, Tohru

2017-10-01

We present a theoretical analysis of singular points of artificial deep neural networks, resulting in providing deep neural network models having no critical points introduced by a hierarchical structure. It is considered that such deep neural network models have good nature for gradient-based optimization. First, we show that there exist a large number of critical points introduced by a hierarchical structure in deep neural networks as straight lines, depending on the number of hidden layers and the number of hidden neurons. Second, we derive a sufficient condition for deep neural networks having no critical points introduced by a hierarchical structure, which can be applied to general deep neural networks. It is also shown that the existence of critical points introduced by a hierarchical structure is determined by the rank and the regularity of weight matrices for a specific class of deep neural networks. Finally, two kinds of implementation methods of the sufficient conditions to have no critical points are provided. One is a learning algorithm that can avoid critical points introduced by the hierarchical structure during learning (called avoidant learning algorithm). The other is a neural network that does not have some critical points introduced by the hierarchical structure as an inherent property (called avoidant neural network).

Neural network error correction for solving coupled ordinary differential equations

NASA Technical Reports Server (NTRS)

Shelton, R. O.; Darsey, J. A.; Sumpter, B. G.; Noid, D. W.

1992-01-01

A neural network is presented to learn errors generated by a numerical algorithm for solving coupled nonlinear differential equations. The method is based on using a neural network to correctly learn the error generated by, for example, Runge-Kutta on a model molecular dynamics (MD) problem. The neural network programs used in this study were developed by NASA. Comparisons are made for training the neural network using backpropagation and a new method which was found to converge with fewer iterations. The neural net programs, the MD model and the calculations are discussed.

Designing the Architecture of Hierachical Neural Networks Model Attention, Learning and Goal-Oriented Behavior

DTIC Science & Technology

1993-12-31

19,23,25,26,27,28,32,33,35,41]) - A new cost function is postulated and an algorithm that employs this cost function is proposed for the learning of...updates the controller parameters from time to time [53]. The learning control algorithm consist of updating the parameter estimates as used in the...proposed cost function with the other learning type algorithms , such as based upon learning of iterative tasks [Kawamura-85], variable structure

Self-learning Monte Carlo with deep neural networks

NASA Astrophysics Data System (ADS)

Shen, Huitao; Liu, Junwei; Fu, Liang

2018-05-01

The self-learning Monte Carlo (SLMC) method is a general algorithm to speedup MC simulations. Its efficiency has been demonstrated in various systems by introducing an effective model to propose global moves in the configuration space. In this paper, we show that deep neural networks can be naturally incorporated into SLMC, and without any prior knowledge can learn the original model accurately and efficiently. Demonstrated in quantum impurity models, we reduce the complexity for a local update from O (β2) in Hirsch-Fye algorithm to O (β lnβ ) , which is a significant speedup especially for systems at low temperatures.

Learning and optimization with cascaded VLSI neural network building-block chips

NASA Technical Reports Server (NTRS)

Duong, T.; Eberhardt, S. P.; Tran, M.; Daud, T.; Thakoor, A. P.

1992-01-01

To demonstrate the versatility of the building-block approach, two neural network applications were implemented on cascaded analog VLSI chips. Weights were implemented using 7-b multiplying digital-to-analog converter (MDAC) synapse circuits, with 31 x 32 and 32 x 32 synapses per chip. A novel learning algorithm compatible with analog VLSI was applied to the two-input parity problem. The algorithm combines dynamically evolving architecture with limited gradient-descent backpropagation for efficient and versatile supervised learning. To implement the learning algorithm in hardware, synapse circuits were paralleled for additional quantization levels. The hardware-in-the-loop learning system allocated 2-5 hidden neurons for parity problems. Also, a 7 x 7 assignment problem was mapped onto a cascaded 64-neuron fully connected feedback network. In 100 randomly selected problems, the network found optimal or good solutions in most cases, with settling times in the range of 7-100 microseconds.

Bringing Interpretability and Visualization with Artificial Neural Networks

ERIC Educational Resources Information Center

Gritsenko, Andrey

2017-01-01

Extreme Learning Machine (ELM) is a training algorithm for Single-Layer Feed-forward Neural Network (SLFN). The difference in theory of ELM from other training algorithms is in the existence of explicitly-given solution due to the immutability of initialed weights. In practice, ELMs achieve performance similar to that of other state-of-the-art…

A novel neural-inspired learning algorithm with application to clinical risk prediction.

PubMed

Tay, Darwin; Poh, Chueh Loo; Kitney, Richard I

2015-04-01

Clinical risk prediction - the estimation of the likelihood an individual is at risk of a disease - is a coveted and exigent clinical task, and a cornerstone to the recommendation of life saving management strategies. This is especially important for individuals at risk of cardiovascular disease (CVD) given the fact that it is the leading causes of death in many developed counties. To this end, we introduce a novel learning algorithm - a key factor that influences the performance of machine learning-based prediction models - and utilities it to develop CVD risk prediction tool. This novel neural-inspired algorithm, called the Artificial Neural Cell System for classification (ANCSc), is inspired by mechanisms that develop the brain and empowering it with capabilities such as information processing/storage and recall, decision making and initiating actions on external environment. Specifically, we exploit on 3 natural neural mechanisms responsible for developing and enriching the brain - namely neurogenesis, neuroplasticity via nurturing and apoptosis - when implementing ANCSc algorithm. Benchmark testing was conducted using the Honolulu Heart Program (HHP) dataset and results are juxtaposed with 2 other algorithms - i.e. Support Vector Machine (SVM) and Evolutionary Data-Conscious Artificial Immune Recognition System (EDC-AIRS). Empirical experiments indicate that ANCSc algorithm (statistically) outperforms both SVM and EDC-AIRS algorithms. Key clinical markers identified by ANCSc algorithm include risk factors related to diet/lifestyle, pulmonary function, personal/family/medical history, blood data, blood pressure, and electrocardiography. These clinical markers, in general, are also found to be clinically significant - providing a promising avenue for identifying potential cardiovascular risk factors to be evaluated in clinical trials. Copyright © 2015 Elsevier Inc. All rights reserved.

Mortality risk score prediction in an elderly population using machine learning.

PubMed

Rose, Sherri

2013-03-01

Standard practice for prediction often relies on parametric regression methods. Interesting new methods from the machine learning literature have been introduced in epidemiologic studies, such as random forest and neural networks. However, a priori, an investigator will not know which algorithm to select and may wish to try several. Here I apply the super learner, an ensembling machine learning approach that combines multiple algorithms into a single algorithm and returns a prediction function with the best cross-validated mean squared error. Super learning is a generalization of stacking methods. I used super learning in the Study of Physical Performance and Age-Related Changes in Sonomans (SPPARCS) to predict death among 2,066 residents of Sonoma, California, aged 54 years or more during the period 1993-1999. The super learner for predicting death (risk score) improved upon all single algorithms in the collection of algorithms, although its performance was similar to that of several algorithms. Super learner outperformed the worst algorithm (neural networks) by 44% with respect to estimated cross-validated mean squared error and had an R2 value of 0.201. The improvement of super learner over random forest with respect to R2 was approximately 2-fold. Alternatives for risk score prediction include the super learner, which can provide improved performance.

Training Feedforward Neural Networks Using Symbiotic Organisms Search Algorithm

PubMed Central

Wu, Haizhou; Luo, Qifang

2016-01-01

Symbiotic organisms search (SOS) is a new robust and powerful metaheuristic algorithm, which stimulates the symbiotic interaction strategies adopted by organisms to survive and propagate in the ecosystem. In the supervised learning area, it is a challenging task to present a satisfactory and efficient training algorithm for feedforward neural networks (FNNs). In this paper, SOS is employed as a new method for training FNNs. To investigate the performance of the aforementioned method, eight different datasets selected from the UCI machine learning repository are employed for experiment and the results are compared among seven metaheuristic algorithms. The results show that SOS performs better than other algorithms for training FNNs in terms of converging speed. It is also proven that an FNN trained by the method of SOS has better accuracy than most algorithms compared. PMID:28105044

On structure-exploiting trust-region regularized nonlinear least squares algorithms for neural-network learning.

PubMed

Mizutani, Eiji; Demmel, James W

2003-01-01

This paper briefly introduces our numerical linear algebra approaches for solving structured nonlinear least squares problems arising from 'multiple-output' neural-network (NN) models. Our algorithms feature trust-region regularization, and exploit sparsity of either the 'block-angular' residual Jacobian matrix or the 'block-arrow' Gauss-Newton Hessian (or Fisher information matrix in statistical sense) depending on problem scale so as to render a large class of NN-learning algorithms 'efficient' in both memory and operation costs. Using a relatively large real-world nonlinear regression application, we shall explain algorithmic strengths and weaknesses, analyzing simulation results obtained by both direct and iterative trust-region algorithms with two distinct NN models: 'multilayer perceptrons' (MLP) and 'complementary mixtures of MLP-experts' (or neuro-fuzzy modular networks).

Bidirectional extreme learning machine for regression problem and its learning effectiveness.

PubMed

Yang, Yimin; Wang, Yaonan; Yuan, Xiaofang

2012-09-01

It is clear that the learning effectiveness and learning speed of neural networks are in general far slower than required, which has been a major bottleneck for many applications. Recently, a simple and efficient learning method, referred to as extreme learning machine (ELM), was proposed by Huang , which has shown that, compared to some conventional methods, the training time of neural networks can be reduced by a thousand times. However, one of the open problems in ELM research is whether the number of hidden nodes can be further reduced without affecting learning effectiveness. This brief proposes a new learning algorithm, called bidirectional extreme learning machine (B-ELM), in which some hidden nodes are not randomly selected. In theory, this algorithm tends to reduce network output error to 0 at an extremely early learning stage. Furthermore, we find a relationship between the network output error and the network output weights in the proposed B-ELM. Simulation results demonstrate that the proposed method can be tens to hundreds of times faster than other incremental ELM algorithms.

Can machine-learning improve cardiovascular risk prediction using routine clinical data?

PubMed Central

Kai, Joe; Garibaldi, Jonathan M.; Qureshi, Nadeem

2017-01-01

Background Current approaches to predict cardiovascular risk fail to identify many people who would benefit from preventive treatment, while others receive unnecessary intervention. Machine-learning offers opportunity to improve accuracy by exploiting complex interactions between risk factors. We assessed whether machine-learning can improve cardiovascular risk prediction. Methods Prospective cohort study using routine clinical data of 378,256 patients from UK family practices, free from cardiovascular disease at outset. Four machine-learning algorithms (random forest, logistic regression, gradient boosting machines, neural networks) were compared to an established algorithm (American College of Cardiology guidelines) to predict first cardiovascular event over 10-years. Predictive accuracy was assessed by area under the ‘receiver operating curve’ (AUC); and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) to predict 7.5% cardiovascular risk (threshold for initiating statins). Findings 24,970 incident cardiovascular events (6.6%) occurred. Compared to the established risk prediction algorithm (AUC 0.728, 95% CI 0.723–0.735), machine-learning algorithms improved prediction: random forest +1.7% (AUC 0.745, 95% CI 0.739–0.750), logistic regression +3.2% (AUC 0.760, 95% CI 0.755–0.766), gradient boosting +3.3% (AUC 0.761, 95% CI 0.755–0.766), neural networks +3.6% (AUC 0.764, 95% CI 0.759–0.769). The highest achieving (neural networks) algorithm predicted 4,998/7,404 cases (sensitivity 67.5%, PPV 18.4%) and 53,458/75,585 non-cases (specificity 70.7%, NPV 95.7%), correctly predicting 355 (+7.6%) more patients who developed cardiovascular disease compared to the established algorithm. Conclusions Machine-learning significantly improves accuracy of cardiovascular risk prediction, increasing the number of patients identified who could benefit from preventive treatment, while avoiding unnecessary treatment of others. PMID:28376093

Can machine-learning improve cardiovascular risk prediction using routine clinical data?

PubMed

Weng, Stephen F; Reps, Jenna; Kai, Joe; Garibaldi, Jonathan M; Qureshi, Nadeem

2017-01-01

Current approaches to predict cardiovascular risk fail to identify many people who would benefit from preventive treatment, while others receive unnecessary intervention. Machine-learning offers opportunity to improve accuracy by exploiting complex interactions between risk factors. We assessed whether machine-learning can improve cardiovascular risk prediction. Prospective cohort study using routine clinical data of 378,256 patients from UK family practices, free from cardiovascular disease at outset. Four machine-learning algorithms (random forest, logistic regression, gradient boosting machines, neural networks) were compared to an established algorithm (American College of Cardiology guidelines) to predict first cardiovascular event over 10-years. Predictive accuracy was assessed by area under the 'receiver operating curve' (AUC); and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) to predict 7.5% cardiovascular risk (threshold for initiating statins). 24,970 incident cardiovascular events (6.6%) occurred. Compared to the established risk prediction algorithm (AUC 0.728, 95% CI 0.723-0.735), machine-learning algorithms improved prediction: random forest +1.7% (AUC 0.745, 95% CI 0.739-0.750), logistic regression +3.2% (AUC 0.760, 95% CI 0.755-0.766), gradient boosting +3.3% (AUC 0.761, 95% CI 0.755-0.766), neural networks +3.6% (AUC 0.764, 95% CI 0.759-0.769). The highest achieving (neural networks) algorithm predicted 4,998/7,404 cases (sensitivity 67.5%, PPV 18.4%) and 53,458/75,585 non-cases (specificity 70.7%, NPV 95.7%), correctly predicting 355 (+7.6%) more patients who developed cardiovascular disease compared to the established algorithm. Machine-learning significantly improves accuracy of cardiovascular risk prediction, increasing the number of patients identified who could benefit from preventive treatment, while avoiding unnecessary treatment of others.

An Adaptive Deghosting Method in Neural Network-Based Infrared Detectors Nonuniformity Correction

PubMed Central

Li, Yiyang; Jin, Weiqi; Zhu, Jin; Zhang, Xu; Li, Shuo

2018-01-01

The problems of the neural network-based nonuniformity correction algorithm for infrared focal plane arrays mainly concern slow convergence speed and ghosting artifacts. In general, the more stringent the inhibition of ghosting, the slower the convergence speed. The factors that affect these two problems are the estimated desired image and the learning rate. In this paper, we propose a learning rate rule that combines adaptive threshold edge detection and a temporal gate. Through the noise estimation algorithm, the adaptive spatial threshold is related to the residual nonuniformity noise in the corrected image. The proposed learning rate is used to effectively and stably suppress ghosting artifacts without slowing down the convergence speed. The performance of the proposed technique was thoroughly studied with infrared image sequences with both simulated nonuniformity and real nonuniformity. The results show that the deghosting performance of the proposed method is superior to that of other neural network-based nonuniformity correction algorithms and that the convergence speed is equivalent to the tested deghosting methods. PMID:29342857

An Adaptive Deghosting Method in Neural Network-Based Infrared Detectors Nonuniformity Correction.

PubMed

Li, Yiyang; Jin, Weiqi; Zhu, Jin; Zhang, Xu; Li, Shuo

2018-01-13

The problems of the neural network-based nonuniformity correction algorithm for infrared focal plane arrays mainly concern slow convergence speed and ghosting artifacts. In general, the more stringent the inhibition of ghosting, the slower the convergence speed. The factors that affect these two problems are the estimated desired image and the learning rate. In this paper, we propose a learning rate rule that combines adaptive threshold edge detection and a temporal gate. Through the noise estimation algorithm, the adaptive spatial threshold is related to the residual nonuniformity noise in the corrected image. The proposed learning rate is used to effectively and stably suppress ghosting artifacts without slowing down the convergence speed. The performance of the proposed technique was thoroughly studied with infrared image sequences with both simulated nonuniformity and real nonuniformity. The results show that the deghosting performance of the proposed method is superior to that of other neural network-based nonuniformity correction algorithms and that the convergence speed is equivalent to the tested deghosting methods.

Fault-tolerant nonlinear adaptive flight control using sliding mode online learning.

PubMed

Krüger, Thomas; Schnetter, Philipp; Placzek, Robin; Vörsmann, Peter

2012-08-01

An expanded nonlinear model inversion flight control strategy using sliding mode online learning for neural networks is presented. The proposed control strategy is implemented for a small unmanned aircraft system (UAS). This class of aircraft is very susceptible towards nonlinearities like atmospheric turbulence, model uncertainties and of course system failures. Therefore, these systems mark a sensible testbed to evaluate fault-tolerant, adaptive flight control strategies. Within this work the concept of feedback linearization is combined with feed forward neural networks to compensate for inversion errors and other nonlinear effects. Backpropagation-based adaption laws of the network weights are used for online training. Within these adaption laws the standard gradient descent backpropagation algorithm is augmented with the concept of sliding mode control (SMC). Implemented as a learning algorithm, this nonlinear control strategy treats the neural network as a controlled system and allows a stable, dynamic calculation of the learning rates. While considering the system's stability, this robust online learning method therefore offers a higher speed of convergence, especially in the presence of external disturbances. The SMC-based flight controller is tested and compared with the standard gradient descent backpropagation algorithm in the presence of system failures. Copyright © 2012 Elsevier Ltd. All rights reserved.

Deep learning improves prediction of CRISPR-Cpf1 guide RNA activity.

PubMed

Kim, Hui Kwon; Min, Seonwoo; Song, Myungjae; Jung, Soobin; Choi, Jae Woo; Kim, Younggwang; Lee, Sangeun; Yoon, Sungroh; Kim, Hyongbum Henry

2018-03-01

We present two algorithms to predict the activity of AsCpf1 guide RNAs. Indel frequencies for 15,000 target sequences were used in a deep-learning framework based on a convolutional neural network to train Seq-deepCpf1. We then incorporated chromatin accessibility information to create the better-performing DeepCpf1 algorithm for cell lines for which such information is available and show that both algorithms outperform previous machine learning algorithms on our own and published data sets.

A neural network model for credit risk evaluation.

PubMed

Khashman, Adnan

2009-08-01

Credit scoring is one of the key analytical techniques in credit risk evaluation which has been an active research area in financial risk management. This paper presents a credit risk evaluation system that uses a neural network model based on the back propagation learning algorithm. We train and implement the neural network to decide whether to approve or reject a credit application, using seven learning schemes and real world credit applications from the Australian credit approval datasets. A comparison of the system performance under the different learning schemes is provided, furthermore, we compare the performance of two neural networks; with one and two hidden layers following the ideal learning scheme. Experimental results suggest that neural networks can be effectively used in automatic processing of credit applications.

Development and application of deep convolutional neural network in target detection

NASA Astrophysics Data System (ADS)

Jiang, Xiaowei; Wang, Chunping; Fu, Qiang

2018-04-01

With the development of big data and algorithms, deep convolution neural networks with more hidden layers have more powerful feature learning and feature expression ability than traditional machine learning methods, making artificial intelligence surpass human level in many fields. This paper first reviews the development and application of deep convolutional neural networks in the field of object detection in recent years, then briefly summarizes and ponders some existing problems in the current research, and the future development of deep convolutional neural network is prospected.

Spiking neuron network Helmholtz machine.

PubMed

Sountsov, Pavel; Miller, Paul

2015-01-01

An increasing amount of behavioral and neurophysiological data suggests that the brain performs optimal (or near-optimal) probabilistic inference and learning during perception and other tasks. Although many machine learning algorithms exist that perform inference and learning in an optimal way, the complete description of how one of those algorithms (or a novel algorithm) can be implemented in the brain is currently incomplete. There have been many proposed solutions that address how neurons can perform optimal inference but the question of how synaptic plasticity can implement optimal learning is rarely addressed. This paper aims to unify the two fields of probabilistic inference and synaptic plasticity by using a neuronal network of realistic model spiking neurons to implement a well-studied computational model called the Helmholtz Machine. The Helmholtz Machine is amenable to neural implementation as the algorithm it uses to learn its parameters, called the wake-sleep algorithm, uses a local delta learning rule. Our spiking-neuron network implements both the delta rule and a small example of a Helmholtz machine. This neuronal network can learn an internal model of continuous-valued training data sets without supervision. The network can also perform inference on the learned internal models. We show how various biophysical features of the neural implementation constrain the parameters of the wake-sleep algorithm, such as the duration of the wake and sleep phases of learning and the minimal sample duration. We examine the deviations from optimal performance and tie them to the properties of the synaptic plasticity rule.

Spiking neuron network Helmholtz machine

PubMed Central

Sountsov, Pavel; Miller, Paul

2015-01-01

An increasing amount of behavioral and neurophysiological data suggests that the brain performs optimal (or near-optimal) probabilistic inference and learning during perception and other tasks. Although many machine learning algorithms exist that perform inference and learning in an optimal way, the complete description of how one of those algorithms (or a novel algorithm) can be implemented in the brain is currently incomplete. There have been many proposed solutions that address how neurons can perform optimal inference but the question of how synaptic plasticity can implement optimal learning is rarely addressed. This paper aims to unify the two fields of probabilistic inference and synaptic plasticity by using a neuronal network of realistic model spiking neurons to implement a well-studied computational model called the Helmholtz Machine. The Helmholtz Machine is amenable to neural implementation as the algorithm it uses to learn its parameters, called the wake-sleep algorithm, uses a local delta learning rule. Our spiking-neuron network implements both the delta rule and a small example of a Helmholtz machine. This neuronal network can learn an internal model of continuous-valued training data sets without supervision. The network can also perform inference on the learned internal models. We show how various biophysical features of the neural implementation constrain the parameters of the wake-sleep algorithm, such as the duration of the wake and sleep phases of learning and the minimal sample duration. We examine the deviations from optimal performance and tie them to the properties of the synaptic plasticity rule. PMID:25954191

Making predictions in a changing world-inference, uncertainty, and learning.

PubMed

O'Reilly, Jill X

2013-01-01

To function effectively, brains need to make predictions about their environment based on past experience, i.e., they need to learn about their environment. The algorithms by which learning occurs are of interest to neuroscientists, both in their own right (because they exist in the brain) and as a tool to model participants' incomplete knowledge of task parameters and hence, to better understand their behavior. This review focusses on a particular challenge for learning algorithms-how to match the rate at which they learn to the rate of change in the environment, so that they use as much observed data as possible whilst disregarding irrelevant, old observations. To do this algorithms must evaluate whether the environment is changing. We discuss the concepts of likelihood, priors and transition functions, and how these relate to change detection. We review expected and estimation uncertainty, and how these relate to change detection and learning rate. Finally, we consider the neural correlates of uncertainty and learning. We argue that the neural correlates of uncertainty bear a resemblance to neural systems that are active when agents actively explore their environments, suggesting that the mechanisms by which the rate of learning is set may be subject to top down control (in circumstances when agents actively seek new information) as well as bottom up control (by observations that imply change in the environment).

A novel joint-processing adaptive nonlinear equalizer using a modular recurrent neural network for chaotic communication systems.

PubMed

Zhao, Haiquan; Zeng, Xiangping; Zhang, Jiashu; Liu, Yangguang; Wang, Xiaomin; Li, Tianrui

2011-01-01

To eliminate nonlinear channel distortion in chaotic communication systems, a novel joint-processing adaptive nonlinear equalizer based on a pipelined recurrent neural network (JPRNN) is proposed, using a modified real-time recurrent learning (RTRL) algorithm. Furthermore, an adaptive amplitude RTRL algorithm is adopted to overcome the deteriorating effect introduced by the nesting process. Computer simulations illustrate that the proposed equalizer outperforms the pipelined recurrent neural network (PRNN) and recurrent neural network (RNN) equalizers. Copyright © 2010 Elsevier Ltd. All rights reserved.

Learning polynomial feedforward neural networks by genetic programming and backpropagation.

PubMed

Nikolaev, N Y; Iba, H

2003-01-01

This paper presents an approach to learning polynomial feedforward neural networks (PFNNs). The approach suggests, first, finding the polynomial network structure by means of a population-based search technique relying on the genetic programming paradigm, and second, further adjustment of the best discovered network weights by an especially derived backpropagation algorithm for higher order networks with polynomial activation functions. These two stages of the PFNN learning process enable us to identify networks with good training as well as generalization performance. Empirical results show that this approach finds PFNN which outperform considerably some previous constructive polynomial network algorithms on processing benchmark time series.

Neural Networks and other Techniques for Fault Identification and Isolation of Aircraft Systems

NASA Technical Reports Server (NTRS)

Innocenti, M.; Napolitano, M.

2003-01-01

Fault identification, isolation, and accomodation have become critical issues in the overall performance of advanced aircraft systems. Neural Networks have shown to be a very attractive alternative to classic adaptation methods for identification and control of non-linear dynamic systems. The purpose of this paper is to show the improvements in neural network applications achievable through the use of learning algorithms more efficient than the classic Back-Propagation, and through the implementation of the neural schemes in parallel hardware. The results of the analysis of a scheme for Sensor Failure, Detection, Identification and Accommodation (SFDIA) using experimental flight data of a research aircraft model are presented. Conventional approaches to the problem are based on observers and Kalman Filters while more recent methods are based on neural approximators. The work described in this paper is based on the use of neural networks (NNs) as on-line learning non-linear approximators. The performances of two different neural architectures were compared. The first architecture is based on a Multi Layer Perceptron (MLP) NN trained with the Extended Back Propagation algorithm (EBPA). The second architecture is based on a Radial Basis Function (RBF) NN trained with the Extended-MRAN (EMRAN) algorithms. In addition, alternative methods for communications links fault detection and accomodation are presented, relative to multiple unmanned aircraft applications.

Gradient descent learning algorithm overview: a general dynamical systems perspective.

PubMed

Baldi, P

1995-01-01

Gives a unified treatment of gradient descent learning algorithms for neural networks using a general framework of dynamical systems. This general approach organizes and simplifies all the known algorithms and results which have been originally derived for different problems (fixed point/trajectory learning), for different models (discrete/continuous), for different architectures (forward/recurrent), and using different techniques (backpropagation, variational calculus, adjoint methods, etc.). The general approach can also be applied to derive new algorithms. The author then briefly examines some of the complexity issues and limitations intrinsic to gradient descent learning. Throughout the paper, the author focuses on the problem of trajectory learning.

Control of magnetic bearing systems via the Chebyshev polynomial-based unified model (CPBUM) neural network.

PubMed

Jeng, J T; Lee, T T

2000-01-01

A Chebyshev polynomial-based unified model (CPBUM) neural network is introduced and applied to control a magnetic bearing systems. First, we show that the CPBUM neural network not only has the same capability of universal approximator, but also has faster learning speed than conventional feedforward/recurrent neural network. It turns out that the CPBUM neural network is more suitable in the design of controller than the conventional feedforward/recurrent neural network. Second, we propose the inverse system method, based on the CPBUM neural networks, to control a magnetic bearing system. The proposed controller has two structures; namely, off-line and on-line learning structures. We derive a new learning algorithm for each proposed structure. The experimental results show that the proposed neural network architecture provides a greater flexibility and better performance in controlling magnetic bearing systems.

Automatic Classification Using Supervised Learning in a Medical Document Filtering Application.

ERIC Educational Resources Information Center

Mostafa, J.; Lam, W.

2000-01-01

Presents a multilevel model of the information filtering process that permits document classification. Evaluates a document classification approach based on a supervised learning algorithm, measures the accuracy of the algorithm in a neural network that was trained to classify medical documents on cell biology, and discusses filtering…

Neural signal processing and closed-loop control algorithm design for an implanted neural recording and stimulation system.

PubMed

Hamilton, Lei; McConley, Marc; Angermueller, Kai; Goldberg, David; Corba, Massimiliano; Kim, Louis; Moran, James; Parks, Philip D; Sang Chin; Widge, Alik S; Dougherty, Darin D; Eskandar, Emad N

2015-08-01

A fully autonomous intracranial device is built to continually record neural activities in different parts of the brain, process these sampled signals, decode features that correlate to behaviors and neuropsychiatric states, and use these features to deliver brain stimulation in a closed-loop fashion. In this paper, we describe the sampling and stimulation aspects of such a device. We first describe the signal processing algorithms of two unsupervised spike sorting methods. Next, we describe the LFP time-frequency analysis and feature derivation from the two spike sorting methods. Spike sorting includes a novel approach to constructing a dictionary learning algorithm in a Compressed Sensing (CS) framework. We present a joint prediction scheme to determine the class of neural spikes in the dictionary learning framework; and, the second approach is a modified OSort algorithm which is implemented in a distributed system optimized for power efficiency. Furthermore, sorted spikes and time-frequency analysis of LFP signals can be used to generate derived features (including cross-frequency coupling, spike-field coupling). We then show how these derived features can be used in the design and development of novel decode and closed-loop control algorithms that are optimized to apply deep brain stimulation based on a patient's neuropsychiatric state. For the control algorithm, we define the state vector as representative of a patient's impulsivity, avoidance, inhibition, etc. Controller parameters are optimized to apply stimulation based on the state vector's current state as well as its historical values. The overall algorithm and software design for our implantable neural recording and stimulation system uses an innovative, adaptable, and reprogrammable architecture that enables advancement of the state-of-the-art in closed-loop neural control while also meeting the challenges of system power constraints and concurrent development with ongoing scientific research designed to define brain network connectivity and neural network dynamics that vary at the individual patient level and vary over time.

An H(∞) control approach to robust learning of feedforward neural networks.

PubMed

Jing, Xingjian

2011-09-01

A novel H(∞) robust control approach is proposed in this study to deal with the learning problems of feedforward neural networks (FNNs). The analysis and design of a desired weight update law for the FNN is transformed into a robust controller design problem for a discrete dynamic system in terms of the estimation error. The drawbacks of some existing learning algorithms can therefore be revealed, especially for the case that the output data is fast changing with respect to the input or the output data is corrupted by noise. Based on this approach, the optimal learning parameters can be found by utilizing the linear matrix inequality (LMI) optimization techniques to achieve a predefined H(∞) "noise" attenuation level. Several existing BP-type algorithms are shown to be special cases of the new H(∞)-learning algorithm. Theoretical analysis and several examples are provided to show the advantages of the new method. Copyright © 2011 Elsevier Ltd. All rights reserved.

Using Fuzzy Logic for Performance Evaluation in Reinforcement Learning

NASA Technical Reports Server (NTRS)

Berenji, Hamid R.; Khedkar, Pratap S.

1992-01-01

Current reinforcement learning algorithms require long training periods which generally limit their applicability to small size problems. A new architecture is described which uses fuzzy rules to initialize its two neural networks: a neural network for performance evaluation and another for action selection. This architecture is applied to control of dynamic systems and it is demonstrated that it is possible to start with an approximate prior knowledge and learn to refine it through experiments using reinforcement learning.

Evaluation of supervised machine-learning algorithms to distinguish between inflammatory bowel disease and alimentary lymphoma in cats.

PubMed

Awaysheh, Abdullah; Wilcke, Jeffrey; Elvinger, François; Rees, Loren; Fan, Weiguo; Zimmerman, Kurt L

2016-11-01

Inflammatory bowel disease (IBD) and alimentary lymphoma (ALA) are common gastrointestinal diseases in cats. The very similar clinical signs and histopathologic features of these diseases make the distinction between them diagnostically challenging. We tested the use of supervised machine-learning algorithms to differentiate between the 2 diseases using data generated from noninvasive diagnostic tests. Three prediction models were developed using 3 machine-learning algorithms: naive Bayes, decision trees, and artificial neural networks. The models were trained and tested on data from complete blood count (CBC) and serum chemistry (SC) results for the following 3 groups of client-owned cats: normal, inflammatory bowel disease (IBD), or alimentary lymphoma (ALA). Naive Bayes and artificial neural networks achieved higher classification accuracy (sensitivities of 70.8% and 69.2%, respectively) than the decision tree algorithm (63%, p < 0.0001). The areas under the receiver-operating characteristic curve for classifying cases into the 3 categories was 83% by naive Bayes, 79% by decision tree, and 82% by artificial neural networks. Prediction models using machine learning provided a method for distinguishing between ALA-IBD, ALA-normal, and IBD-normal. The naive Bayes and artificial neural networks classifiers used 10 and 4 of the CBC and SC variables, respectively, to outperform the C4.5 decision tree, which used 5 CBC and SC variables in classifying cats into the 3 classes. These models can provide another noninvasive diagnostic tool to assist clinicians with differentiating between IBD and ALA, and between diseased and nondiseased cats. © 2016 The Author(s).

Adaptive Optimization of Aircraft Engine Performance Using Neural Networks

NASA Technical Reports Server (NTRS)

Simon, Donald L.; Long, Theresa W.

1995-01-01

Preliminary results are presented on the development of an adaptive neural network based control algorithm to enhance aircraft engine performance. This work builds upon a previous National Aeronautics and Space Administration (NASA) effort known as Performance Seeking Control (PSC). PSC is an adaptive control algorithm which contains a model of the aircraft's propulsion system which is updated on-line to match the operation of the aircraft's actual propulsion system. Information from the on-line model is used to adapt the control system during flight to allow optimal operation of the aircraft's propulsion system (inlet, engine, and nozzle) to improve aircraft engine performance without compromising reliability or operability. Performance Seeking Control has been shown to yield reductions in fuel flow, increases in thrust, and reductions in engine fan turbine inlet temperature. The neural network based adaptive control, like PSC, will contain a model of the propulsion system which will be used to calculate optimal control commands on-line. Hopes are that it will be able to provide some additional benefits above and beyond those of PSC. The PSC algorithm is computationally intensive, it is valid only at near steady-state flight conditions, and it has no way to adapt or learn on-line. These issues are being addressed in the development of the optimal neural controller. Specialized neural network processing hardware is being developed to run the software, the algorithm will be valid at steady-state and transient conditions, and will take advantage of the on-line learning capability of neural networks. Future plans include testing the neural network software and hardware prototype against an aircraft engine simulation. In this paper, the proposed neural network software and hardware is described and preliminary neural network training results are presented.

Are Student Evaluations of Teaching Effectiveness Valid for Measuring Student Learning Outcomes in Business Related Classes? A Neural Network and Bayesian Analyses

ERIC Educational Resources Information Center

Galbraith, Craig S.; Merrill, Gregory B.; Kline, Doug M.

2012-01-01

In this study we investigate the underlying relational structure between student evaluations of teaching effectiveness (SETEs) and achievement of student learning outcomes in 116 business related courses. Utilizing traditional statistical techniques, a neural network analysis and a Bayesian data reduction and classification algorithm, we find…

Using an Extended Kalman Filter Learning Algorithm for Feed-Forward Neural Networks to Describe Tracer Correlations

NASA Technical Reports Server (NTRS)

Lary, David J.; Mussa, Yussuf

2004-01-01

In this study a new extended Kalman filter (EKF) learning algorithm for feed-forward neural networks (FFN) is used. With the EKF approach, the training of the FFN can be seen as state estimation for a non-linear stationary process. The EKF method gives excellent convergence performances provided that there is enough computer core memory and that the machine precision is high. Neural networks are ideally suited to describe the spatial and temporal dependence of tracer-tracer correlations. The neural network performs well even in regions where the correlations are less compact and normally a family of correlation curves would be required. For example, the CH4-N2O correlation can be well described using a neural network trained with the latitude, pressure, time of year, and CH4 volume mixing ratio (v.m.r.). The neural network was able to reproduce the CH4-N2O correlation with a correlation coefficient between simulated and training values of 0.9997. The neural network Fortran code used is available for download.

Program Helps Simulate Neural Networks

NASA Technical Reports Server (NTRS)

Villarreal, James; Mcintire, Gary

1993-01-01

Neural Network Environment on Transputer System (NNETS) computer program provides users high degree of flexibility in creating and manipulating wide variety of neural-network topologies at processing speeds not found in conventional computing environments. Supports back-propagation and back-propagation-related algorithms. Back-propagation algorithm used is implementation of Rumelhart's generalized delta rule. NNETS developed on INMOS Transputer(R). Predefines back-propagation network, Jordan network, and reinforcement network to assist users in learning and defining own networks. Also enables users to configure other neural-network paradigms from NNETS basic architecture. Small portion of software written in OCCAM(R) language.

Supervised Machine Learning for Regionalization of Environmental Data: Distribution of Uranium in Groundwater in Ukraine

NASA Astrophysics Data System (ADS)

Govorov, Michael; Gienko, Gennady; Putrenko, Viktor

2018-05-01

In this paper, several supervised machine learning algorithms were explored to define homogeneous regions of con-centration of uranium in surface waters in Ukraine using multiple environmental parameters. The previous study was focused on finding the primary environmental parameters related to uranium in ground waters using several methods of spatial statistics and unsupervised classification. At this step, we refined the regionalization using Artifi-cial Neural Networks (ANN) techniques including Multilayer Perceptron (MLP), Radial Basis Function (RBF), and Convolutional Neural Network (CNN). The study is focused on building local ANN models which may significantly improve the prediction results of machine learning algorithms by taking into considerations non-stationarity and autocorrelation in spatial data.

Genetic attack on neural cryptography.

PubMed

Ruttor, Andreas; Kinzel, Wolfgang; Naeh, Rivka; Kanter, Ido

2006-03-01

Different scaling properties for the complexity of bidirectional synchronization and unidirectional learning are essential for the security of neural cryptography. Incrementing the synaptic depth of the networks increases the synchronization time only polynomially, but the success of the geometric attack is reduced exponentially and it clearly fails in the limit of infinite synaptic depth. This method is improved by adding a genetic algorithm, which selects the fittest neural networks. The probability of a successful genetic attack is calculated for different model parameters using numerical simulations. The results show that scaling laws observed in the case of other attacks hold for the improved algorithm, too. The number of networks needed for an effective attack grows exponentially with increasing synaptic depth. In addition, finite-size effects caused by Hebbian and anti-Hebbian learning are analyzed. These learning rules converge to the random walk rule if the synaptic depth is small compared to the square root of the system size.

Genetic attack on neural cryptography

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

Ruttor, Andreas; Kinzel, Wolfgang; Naeh, Rivka

2006-03-15

Different scaling properties for the complexity of bidirectional synchronization and unidirectional learning are essential for the security of neural cryptography. Incrementing the synaptic depth of the networks increases the synchronization time only polynomially, but the success of the geometric attack is reduced exponentially and it clearly fails in the limit of infinite synaptic depth. This method is improved by adding a genetic algorithm, which selects the fittest neural networks. The probability of a successful genetic attack is calculated for different model parameters using numerical simulations. The results show that scaling laws observed in the case of other attacks hold formore » the improved algorithm, too. The number of networks needed for an effective attack grows exponentially with increasing synaptic depth. In addition, finite-size effects caused by Hebbian and anti-Hebbian learning are analyzed. These learning rules converge to the random walk rule if the synaptic depth is small compared to the square root of the system size.« less

Genetic attack on neural cryptography

NASA Astrophysics Data System (ADS)

Ruttor, Andreas; Kinzel, Wolfgang; Naeh, Rivka; Kanter, Ido

2006-03-01

Different scaling properties for the complexity of bidirectional synchronization and unidirectional learning are essential for the security of neural cryptography. Incrementing the synaptic depth of the networks increases the synchronization time only polynomially, but the success of the geometric attack is reduced exponentially and it clearly fails in the limit of infinite synaptic depth. This method is improved by adding a genetic algorithm, which selects the fittest neural networks. The probability of a successful genetic attack is calculated for different model parameters using numerical simulations. The results show that scaling laws observed in the case of other attacks hold for the improved algorithm, too. The number of networks needed for an effective attack grows exponentially with increasing synaptic depth. In addition, finite-size effects caused by Hebbian and anti-Hebbian learning are analyzed. These learning rules converge to the random walk rule if the synaptic depth is small compared to the square root of the system size.

Using input feature information to improve ultraviolet retrieval in neural networks

NASA Astrophysics Data System (ADS)

Sun, Zhibin; Chang, Ni-Bin; Gao, Wei; Chen, Maosi; Zempila, Melina

2017-09-01

In neural networks, the training/predicting accuracy and algorithm efficiency can be improved significantly via accurate input feature extraction. In this study, some spatial features of several important factors in retrieving surface ultraviolet (UV) are extracted. An extreme learning machine (ELM) is used to retrieve the surface UV of 2014 in the continental United States, using the extracted features. The results conclude that more input weights can improve the learning capacities of neural networks.

Predicting the random drift of MEMS gyroscope based on K-means clustering and OLS RBF Neural Network

NASA Astrophysics Data System (ADS)

Wang, Zhen-yu; Zhang, Li-jie

2017-10-01

Measure error of the sensor can be effectively compensated with prediction. Aiming at large random drift error of MEMS(Micro Electro Mechanical System))gyroscope, an improved learning algorithm of Radial Basis Function(RBF) Neural Network(NN) based on K-means clustering and Orthogonal Least-Squares (OLS) is proposed in this paper. The algorithm selects the typical samples as the initial cluster centers of RBF NN firstly, candidates centers with K-means algorithm secondly, and optimizes the candidate centers with OLS algorithm thirdly, which makes the network structure simpler and makes the prediction performance better. Experimental results show that the proposed K-means clustering OLS learning algorithm can predict the random drift of MEMS gyroscope effectively, the prediction error of which is 9.8019e-007°/s and the prediction time of which is 2.4169e-006s

LensFlow: A Convolutional Neural Network in Search of Strong Gravitational Lenses

NASA Astrophysics Data System (ADS)

Pourrahmani, Milad; Nayyeri, Hooshang; Cooray, Asantha

2018-03-01

In this work, we present our machine learning classification algorithm for identifying strong gravitational lenses from wide-area surveys using convolutional neural networks; LENSFLOW. We train and test the algorithm using a wide variety of strong gravitational lens configurations from simulations of lensing events. Images are processed through multiple convolutional layers that extract feature maps necessary to assign a lens probability to each image. LENSFLOW provides a ranking scheme for all sources that could be used to identify potential gravitational lens candidates by significantly reducing the number of images that have to be visually inspected. We apply our algorithm to the HST/ACS i-band observations of the COSMOS field and present our sample of identified lensing candidates. The developed machine learning algorithm is more computationally efficient and complimentary to classical lens identification algorithms and is ideal for discovering such events across wide areas from current and future surveys such as LSST and WFIRST.

Approximately adaptive neural cooperative control for nonlinear multiagent systems with performance guarantee

NASA Astrophysics Data System (ADS)

Wang, Jing; Yang, Tianyu; Staskevich, Gennady; Abbe, Brian

2017-04-01

This paper studies the cooperative control problem for a class of multiagent dynamical systems with partially unknown nonlinear system dynamics. In particular, the control objective is to solve the state consensus problem for multiagent systems based on the minimisation of certain cost functions for individual agents. Under the assumption that there exist admissible cooperative controls for such class of multiagent systems, the formulated problem is solved through finding the optimal cooperative control using the approximate dynamic programming and reinforcement learning approach. With the aid of neural network parameterisation and online adaptive learning, our method renders a practically implementable approximately adaptive neural cooperative control for multiagent systems. Specifically, based on the Bellman's principle of optimality, the Hamilton-Jacobi-Bellman (HJB) equation for multiagent systems is first derived. We then propose an approximately adaptive policy iteration algorithm for multiagent cooperative control based on neural network approximation of the value functions. The convergence of the proposed algorithm is rigorously proved using the contraction mapping method. The simulation results are included to validate the effectiveness of the proposed algorithm.

Recursive least-squares learning algorithms for neural networks

NASA Astrophysics Data System (ADS)

Lewis, Paul S.; Hwang, Jenq N.

1990-11-01

This paper presents the development of a pair of recursive least squares (ItLS) algorithms for online training of multilayer perceptrons which are a class of feedforward artificial neural networks. These algorithms incorporate second order information about the training error surface in order to achieve faster learning rates than are possible using first order gradient descent algorithms such as the generalized delta rule. A least squares formulation is derived from a linearization of the training error function. Individual training pattern errors are linearized about the network parameters that were in effect when the pattern was presented. This permits the recursive solution of the least squares approximation either via conventional RLS recursions or by recursive QR decomposition-based techniques. The computational complexity of the update is 0(N2) where N is the number of network parameters. This is due to the estimation of the N x N inverse Hessian matrix. Less computationally intensive approximations of the ilLS algorithms can be easily derived by using only block diagonal elements of this matrix thereby partitioning the learning into independent sets. A simulation example is presented in which a neural network is trained to approximate a two dimensional Gaussian bump. In this example RLS training required an order of magnitude fewer iterations on average (527) than did training with the generalized delta rule (6 1 BACKGROUND Artificial neural networks (ANNs) offer an interesting and potentially useful paradigm for signal processing and pattern recognition. The majority of ANN applications employ the feed-forward multilayer perceptron (MLP) network architecture in which network parameters are " trained" by a supervised learning algorithm employing the generalized delta rule (GDIt) [1 2]. The GDR algorithm approximates a fixed step steepest descent algorithm using derivatives computed by error backpropagatiori. The GDII algorithm is sometimes referred to as the backpropagation algorithm. However in this paper we will use the term backpropagation to refer only to the process of computing error derivatives. While multilayer perceptrons provide a very powerful nonlinear modeling capability GDR training can be very slow and inefficient. In linear adaptive filtering the analog of the GDR algorithm is the leastmean- squares (LMS) algorithm. Steepest descent-based algorithms such as GDR or LMS are first order because they use only first derivative or gradient information about the training error to be minimized. To speed up the training process second order algorithms may be employed that take advantage of second derivative or Hessian matrix information. Second order information can be incorporated into MLP training in different ways. In many applications especially in the area of pattern recognition the training set is finite. In these cases block learning can be applied using standard nonlinear optimization techniques [3 4 5].

Gross domestic product estimation based on electricity utilization by artificial neural network

NASA Astrophysics Data System (ADS)

Stevanović, Mirjana; Vujičić, Slađana; Gajić, Aleksandar M.

2018-01-01

The main goal of the paper was to estimate gross domestic product (GDP) based on electricity estimation by artificial neural network (ANN). The electricity utilization was analyzed based on different sources like renewable, coal and nuclear sources. The ANN network was trained with two training algorithms namely extreme learning method and back-propagation algorithm in order to produce the best prediction results of the GDP. According to the results it can be concluded that the ANN model with extreme learning method could produce the acceptable prediction of the GDP based on the electricity utilization.

Optical implementation of neural learning algorithms based on cross-gain modulation in a semiconductor optical amplifier

NASA Astrophysics Data System (ADS)

Li, Qiang; Wang, Zhi; Le, Yansi; Sun, Chonghui; Song, Xiaojia; Wu, Chongqing

2016-10-01

Neuromorphic engineering has a wide range of applications in the fields of machine learning, pattern recognition, adaptive control, etc. Photonics, characterized by its high speed, wide bandwidth, low power consumption and massive parallelism, is an ideal way to realize ultrafast spiking neural networks (SNNs). Synaptic plasticity is believed to be critical for learning, memory and development in neural circuits. Experimental results have shown that changes of synapse are highly dependent on the relative timing of pre- and postsynaptic spikes. Synaptic plasticity in which presynaptic spikes preceding postsynaptic spikes results in strengthening, while the opposite timing results in weakening is called antisymmetric spike-timing-dependent plasticity (STDP) learning rule. And synaptic plasticity has the opposite effect under the same conditions is called antisymmetric anti-STDP learning rule. We proposed and experimentally demonstrated an optical implementation of neural learning algorithms, which can achieve both of antisymmetric STDP and anti-STDP learning rule, based on the cross-gain modulation (XGM) within a single semiconductor optical amplifier (SOA). The weight and height of the potentitation and depression window can be controlled by adjusting the injection current of the SOA, to mimic the biological antisymmetric STDP and anti-STDP learning rule more realistically. As the injection current increases, the width of depression and potentitation window decreases and height increases, due to the decreasing of recovery time and increasing of gain under a stronger injection current. Based on the demonstrated optical STDP circuit, ultrafast learning in optical SNNs can be realized.

Electronic Neural Networks

NASA Technical Reports Server (NTRS)

Thakoor, Anil

1990-01-01

Viewgraphs on electronic neural networks for space station are presented. Topics covered include: electronic neural networks; electronic implementations; VLSI/thin film hybrid hardware for neurocomputing; computations with analog parallel processing; features of neuroprocessors; applications of neuroprocessors; neural network hardware for terrain trafficability determination; a dedicated processor for path planning; neural network system interface; neural network for robotic control; error backpropagation algorithm for learning; resource allocation matrix; global optimization neuroprocessor; and electrically programmable read only thin-film synaptic array.

Computational neural learning formalisms for manipulator inverse kinematics

NASA Technical Reports Server (NTRS)

Gulati, Sandeep; Barhen, Jacob; Iyengar, S. Sitharama

1989-01-01

An efficient, adaptive neural learning paradigm for addressing the inverse kinematics of redundant manipulators is presented. The proposed methodology exploits the infinite local stability of terminal attractors - a new class of mathematical constructs which provide unique information processing capabilities to artificial neural systems. For robotic applications, synaptic elements of such networks can rapidly acquire the kinematic invariances embedded within the presented samples. Subsequently, joint-space configurations, required to follow arbitrary end-effector trajectories, can readily be computed. In a significant departure from prior neuromorphic learning algorithms, this methodology provides mechanisms for incorporating an in-training skew to handle kinematics and environmental constraints.

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).

Neural Decoder for Topological Codes

NASA Astrophysics Data System (ADS)

Torlai, Giacomo; Melko, Roger G.

2017-07-01

We present an algorithm for error correction in topological codes that exploits modern machine learning techniques. Our decoder is constructed from a stochastic neural network called a Boltzmann machine, of the type extensively used in deep learning. We provide a general prescription for the training of the network and a decoding strategy that is applicable to a wide variety of stabilizer codes with very little specialization. We demonstrate the neural decoder numerically on the well-known two-dimensional toric code with phase-flip errors.

Efficient self-organizing multilayer neural network for nonlinear system modeling.

PubMed

Han, Hong-Gui; Wang, Li-Dan; Qiao, Jun-Fei

2013-07-01

It has been shown extensively that the dynamic behaviors of a neural system are strongly influenced by the network architecture and learning process. To establish an artificial neural network (ANN) with self-organizing architecture and suitable learning algorithm for nonlinear system modeling, an automatic axon-neural network (AANN) is investigated in the following respects. First, the network architecture is constructed automatically to change both the number of hidden neurons and topologies of the neural network during the training process. The approach introduced in adaptive connecting-and-pruning algorithm (ACP) is a type of mixed mode operation, which is equivalent to pruning or adding the connecting of the neurons, as well as inserting some required neurons directly. Secondly, the weights are adjusted, using a feedforward computation (FC) to obtain the information for the gradient during learning computation. Unlike most of the previous studies, AANN is able to self-organize the architecture and weights, and to improve the network performances. Also, the proposed AANN has been tested on a number of benchmark problems, ranging from nonlinear function approximating to nonlinear systems modeling. The experimental results show that AANN can have better performances than that of some existing neural networks. Crown Copyright © 2013. Published by Elsevier Ltd. All rights reserved.

Selected Flight Test Results for Online Learning Neural Network-Based Flight Control System

NASA Technical Reports Server (NTRS)

Williams, Peggy S.

2004-01-01

The NASA F-15 Intelligent Flight Control System project team has developed a series of flight control concepts designed to demonstrate the benefits of a neural network-based adaptive controller. The objective of the team is to develop and flight-test control systems that use neural network technology to optimize the performance of the aircraft under nominal conditions as well as stabilize the aircraft under failure conditions. Failure conditions include locked or failed control surfaces as well as unforeseen damage that might occur to the aircraft in flight. This report presents flight-test results for an adaptive controller using stability and control derivative values from an online learning neural network. A dynamic cell structure neural network is used in conjunction with a real-time parameter identification algorithm to estimate aerodynamic stability and control derivative increments to the baseline aerodynamic derivatives in flight. This set of open-loop flight tests was performed in preparation for a future phase of flights in which the learning neural network and parameter identification algorithm output would provide the flight controller with aerodynamic stability and control derivative updates in near real time. Two flight maneuvers are analyzed a pitch frequency sweep and an automated flight-test maneuver designed to optimally excite the parameter identification algorithm in all axes. Frequency responses generated from flight data are compared to those obtained from nonlinear simulation runs. An examination of flight data shows that addition of the flight-identified aerodynamic derivative increments into the simulation improved the pitch handling qualities of the aircraft.

Research on particle swarm optimization algorithm based on optimal movement probability

NASA Astrophysics Data System (ADS)

Ma, Jianhong; Zhang, Han; He, Baofeng

2017-01-01

The particle swarm optimization algorithm to improve the control precision, and has great application value training neural network and fuzzy system control fields etc.The traditional particle swarm algorithm is used for the training of feed forward neural networks,the search efficiency is low, and easy to fall into local convergence.An improved particle swarm optimization algorithm is proposed based on error back propagation gradient descent. Particle swarm optimization for Solving Least Squares Problems to meme group, the particles in the fitness ranking, optimization problem of the overall consideration, the error back propagation gradient descent training BP neural network, particle to update the velocity and position according to their individual optimal and global optimization, make the particles more to the social optimal learning and less to its optimal learning, it can avoid the particles fall into local optimum, by using gradient information can accelerate the PSO local search ability, improve the multi beam particle swarm depth zero less trajectory information search efficiency, the realization of improved particle swarm optimization algorithm. Simulation results show that the algorithm in the initial stage of rapid convergence to the global optimal solution can be near to the global optimal solution and keep close to the trend, the algorithm has faster convergence speed and search performance in the same running time, it can improve the convergence speed of the algorithm, especially the later search efficiency.

Neuromorphic implementations of neurobiological learning algorithms for spiking neural networks.

PubMed

Walter, Florian; Röhrbein, Florian; Knoll, Alois

2015-12-01

The application of biologically inspired methods in design and control has a long tradition in robotics. Unlike previous approaches in this direction, the emerging field of neurorobotics not only mimics biological mechanisms at a relatively high level of abstraction but employs highly realistic simulations of actual biological nervous systems. Even today, carrying out these simulations efficiently at appropriate timescales is challenging. Neuromorphic chip designs specially tailored to this task therefore offer an interesting perspective for neurorobotics. Unlike Von Neumann CPUs, these chips cannot be simply programmed with a standard programming language. Like real brains, their functionality is determined by the structure of neural connectivity and synaptic efficacies. Enabling higher cognitive functions for neurorobotics consequently requires the application of neurobiological learning algorithms to adjust synaptic weights in a biologically plausible way. In this paper, we therefore investigate how to program neuromorphic chips by means of learning. First, we provide an overview over selected neuromorphic chip designs and analyze them in terms of neural computation, communication systems and software infrastructure. On the theoretical side, we review neurobiological learning techniques. Based on this overview, we then examine on-die implementations of these learning algorithms on the considered neuromorphic chips. A final discussion puts the findings of this work into context and highlights how neuromorphic hardware can potentially advance the field of autonomous robot systems. The paper thus gives an in-depth overview of neuromorphic implementations of basic mechanisms of synaptic plasticity which are required to realize advanced cognitive capabilities with spiking neural networks. Copyright © 2015 Elsevier Ltd. All rights reserved.

Efficient and self-adaptive in-situ learning in multilayer memristor neural networks.

PubMed

Li, Can; Belkin, Daniel; Li, Yunning; Yan, Peng; Hu, Miao; Ge, Ning; Jiang, Hao; Montgomery, Eric; Lin, Peng; Wang, Zhongrui; Song, Wenhao; Strachan, John Paul; Barnell, Mark; Wu, Qing; Williams, R Stanley; Yang, J Joshua; Xia, Qiangfei

2018-06-19

Memristors with tunable resistance states are emerging building blocks of artificial neural networks. However, in situ learning on a large-scale multiple-layer memristor network has yet to be demonstrated because of challenges in device property engineering and circuit integration. Here we monolithically integrate hafnium oxide-based memristors with a foundry-made transistor array into a multiple-layer neural network. We experimentally demonstrate in situ learning capability and achieve competitive classification accuracy on a standard machine learning dataset, which further confirms that the training algorithm allows the network to adapt to hardware imperfections. Our simulation using the experimental parameters suggests that a larger network would further increase the classification accuracy. The memristor neural network is a promising hardware platform for artificial intelligence with high speed-energy efficiency.

Learning Traffic as Images: A Deep Convolutional Neural Network for Large-Scale Transportation Network Speed Prediction.

PubMed

Ma, Xiaolei; Dai, Zhuang; He, Zhengbing; Ma, Jihui; Wang, Yong; Wang, Yunpeng

2017-04-10

This paper proposes a convolutional neural network (CNN)-based method that learns traffic as images and predicts large-scale, network-wide traffic speed with a high accuracy. Spatiotemporal traffic dynamics are converted to images describing the time and space relations of traffic flow via a two-dimensional time-space matrix. A CNN is applied to the image following two consecutive steps: abstract traffic feature extraction and network-wide traffic speed prediction. The effectiveness of the proposed method is evaluated by taking two real-world transportation networks, the second ring road and north-east transportation network in Beijing, as examples, and comparing the method with four prevailing algorithms, namely, ordinary least squares, k-nearest neighbors, artificial neural network, and random forest, and three deep learning architectures, namely, stacked autoencoder, recurrent neural network, and long-short-term memory network. The results show that the proposed method outperforms other algorithms by an average accuracy improvement of 42.91% within an acceptable execution time. The CNN can train the model in a reasonable time and, thus, is suitable for large-scale transportation networks.

Learning Traffic as Images: A Deep Convolutional Neural Network for Large-Scale Transportation Network Speed Prediction

PubMed Central

Ma, Xiaolei; Dai, Zhuang; He, Zhengbing; Ma, Jihui; Wang, Yong; Wang, Yunpeng

2017-01-01

This paper proposes a convolutional neural network (CNN)-based method that learns traffic as images and predicts large-scale, network-wide traffic speed with a high accuracy. Spatiotemporal traffic dynamics are converted to images describing the time and space relations of traffic flow via a two-dimensional time-space matrix. A CNN is applied to the image following two consecutive steps: abstract traffic feature extraction and network-wide traffic speed prediction. The effectiveness of the proposed method is evaluated by taking two real-world transportation networks, the second ring road and north-east transportation network in Beijing, as examples, and comparing the method with four prevailing algorithms, namely, ordinary least squares, k-nearest neighbors, artificial neural network, and random forest, and three deep learning architectures, namely, stacked autoencoder, recurrent neural network, and long-short-term memory network. The results show that the proposed method outperforms other algorithms by an average accuracy improvement of 42.91% within an acceptable execution time. The CNN can train the model in a reasonable time and, thus, is suitable for large-scale transportation networks. PMID:28394270

Artificial neuron-glia networks learning approach based on cooperative coevolution.

PubMed

Mesejo, Pablo; Ibáñez, Oscar; Fernández-Blanco, Enrique; Cedrón, Francisco; Pazos, Alejandro; Porto-Pazos, Ana B

2015-06-01

Artificial Neuron-Glia Networks (ANGNs) are a novel bio-inspired machine learning approach. They extend classical Artificial Neural Networks (ANNs) by incorporating recent findings and suppositions about the way information is processed by neural and astrocytic networks in the most evolved living organisms. Although ANGNs are not a consolidated method, their performance against the traditional approach, i.e. without artificial astrocytes, was already demonstrated on classification problems. However, the corresponding learning algorithms developed so far strongly depends on a set of glial parameters which are manually tuned for each specific problem. As a consequence, previous experimental tests have to be done in order to determine an adequate set of values, making such manual parameter configuration time-consuming, error-prone, biased and problem dependent. Thus, in this paper, we propose a novel learning approach for ANGNs that fully automates the learning process, and gives the possibility of testing any kind of reasonable parameter configuration for each specific problem. This new learning algorithm, based on coevolutionary genetic algorithms, is able to properly learn all the ANGNs parameters. Its performance is tested on five classification problems achieving significantly better results than ANGN and competitive results with ANN approaches.

Autonomous reinforcement learning with experience replay.

PubMed

Wawrzyński, Paweł; Tanwani, Ajay Kumar

2013-05-01

This paper considers the issues of efficiency and autonomy that are required to make reinforcement learning suitable for real-life control tasks. A real-time reinforcement learning algorithm is presented that repeatedly adjusts the control policy with the use of previously collected samples, and autonomously estimates the appropriate step-sizes for the learning updates. The algorithm is based on the actor-critic with experience replay whose step-sizes are determined on-line by an enhanced fixed point algorithm for on-line neural network training. An experimental study with simulated octopus arm and half-cheetah demonstrates the feasibility of the proposed algorithm to solve difficult learning control problems in an autonomous way within reasonably short time. Copyright © 2012 Elsevier Ltd. All rights reserved.

Evolutionary image simplification for lung nodule classification with convolutional neural networks.

PubMed

Lückehe, Daniel; von Voigt, Gabriele

2018-05-29

Understanding decisions of deep learning techniques is important. Especially in the medical field, the reasons for a decision in a classification task are as crucial as the pure classification results. In this article, we propose a new approach to compute relevant parts of a medical image. Knowing the relevant parts makes it easier to understand decisions. In our approach, a convolutional neural network is employed to learn structures of images of lung nodules. Then, an evolutionary algorithm is applied to compute a simplified version of an unknown image based on the learned structures by the convolutional neural network. In the simplified version, irrelevant parts are removed from the original image. In the results, we show simplified images which allow the observer to focus on the relevant parts. In these images, more than 50% of the pixels are simplified. The simplified pixels do not change the meaning of the images based on the learned structures by the convolutional neural network. An experimental analysis shows the potential of the approach. Besides the examples of simplified images, we analyze the run time development. Simplified images make it easier to focus on relevant parts and to find reasons for a decision. The combination of an evolutionary algorithm employing a learned convolutional neural network is well suited for the simplification task. From a research perspective, it is interesting which areas of the images are simplified and which parts are taken as relevant.

Using Reinforcement Learning to Provide Stable Brain-Machine Interface Control Despite Neural Input Reorganization

PubMed Central

Pohlmeyer, Eric A.; Mahmoudi, Babak; Geng, Shijia; Prins, Noeline W.; Sanchez, Justin C.

2014-01-01

Brain-machine interface (BMI) systems give users direct neural control of robotic, communication, or functional electrical stimulation systems. As BMI systems begin transitioning from laboratory settings into activities of daily living, an important goal is to develop neural decoding algorithms that can be calibrated with a minimal burden on the user, provide stable control for long periods of time, and can be responsive to fluctuations in the decoder’s neural input space (e.g. neurons appearing or being lost amongst electrode recordings). These are significant challenges for static neural decoding algorithms that assume stationary input/output relationships. Here we use an actor-critic reinforcement learning architecture to provide an adaptive BMI controller that can successfully adapt to dramatic neural reorganizations, can maintain its performance over long time periods, and which does not require the user to produce specific kinetic or kinematic activities to calibrate the BMI. Two marmoset monkeys used the Reinforcement Learning BMI (RLBMI) to successfully control a robotic arm during a two-target reaching task. The RLBMI was initialized using random initial conditions, and it quickly learned to control the robot from brain states using only a binary evaluative feedback regarding whether previously chosen robot actions were good or bad. The RLBMI was able to maintain control over the system throughout sessions spanning multiple weeks. Furthermore, the RLBMI was able to quickly adapt and maintain control of the robot despite dramatic perturbations to the neural inputs, including a series of tests in which the neuron input space was deliberately halved or doubled. PMID:24498055

Comparison of artificial intelligence classifiers for SIP attack data

NASA Astrophysics Data System (ADS)

Safarik, Jakub; Slachta, Jiri

2016-05-01

Honeypot application is a source of valuable data about attacks on the network. We run several SIP honeypots in various computer networks, which are separated geographically and logically. Each honeypot runs on public IP address and uses standard SIP PBX ports. All information gathered via honeypot is periodically sent to the centralized server. This server classifies all attack data by neural network algorithm. The paper describes optimizations of a neural network classifier, which lower the classification error. The article contains the comparison of two neural network algorithm used for the classification of validation data. The first is the original implementation of the neural network described in recent work; the second neural network uses further optimizations like input normalization or cross-entropy cost function. We also use other implementations of neural networks and machine learning classification algorithms. The comparison test their capabilities on validation data to find the optimal classifier. The article result shows promise for further development of an accurate SIP attack classification engine.

Brain-Machine Interface control of a robot arm using actor-critic rainforcement learning.

PubMed

Pohlmeyer, Eric A; Mahmoudi, Babak; Geng, Shijia; Prins, Noeline; Sanchez, Justin C

2012-01-01

Here we demonstrate how a marmoset monkey can use a reinforcement learning (RL) Brain-Machine Interface (BMI) to effectively control the movements of a robot arm for a reaching task. In this work, an actor-critic RL algorithm used neural ensemble activity in the monkey's motor cortext to control the robot movements during a two-target decision task. This novel approach to decoding offers unique advantages for BMI control applications. Compared to supervised learning decoding methods, the actor-critic RL algorithm does not require an explicit set of training data to create a static control model, but rather it incrementally adapts the model parameters according to its current performance, in this case requiring only a very basic feedback signal. We show how this algorithm achieved high performance when mapping the monkey's neural states (94%) to robot actions, and only needed to experience a few trials before obtaining accurate real-time control of the robot arm. Since RL methods responsively adapt and adjust their parameters, they can provide a method to create BMIs that are robust against perturbations caused by changes in either the neural input space or the output actions they generate under different task requirements or goals.

"FORCE" learning in recurrent neural networks as data assimilation

NASA Astrophysics Data System (ADS)

Duane, Gregory S.

2017-12-01

It is shown that the "FORCE" algorithm for learning in arbitrarily connected networks of simple neuronal units can be cast as a Kalman Filter, with a particular state-dependent form for the background error covariances. The resulting interpretation has implications for initialization of the learning algorithm, leads to an extension to include interactions between the weight updates for different neurons, and can represent relationships within groups of multiple target output signals.

Scalable training of artificial neural networks with adaptive sparse connectivity inspired by network science.

PubMed

Mocanu, Decebal Constantin; Mocanu, Elena; Stone, Peter; Nguyen, Phuong H; Gibescu, Madeleine; Liotta, Antonio

2018-06-19

Through the success of deep learning in various domains, artificial neural networks are currently among the most used artificial intelligence methods. Taking inspiration from the network properties of biological neural networks (e.g. sparsity, scale-freeness), we argue that (contrary to general practice) artificial neural networks, too, should not have fully-connected layers. Here we propose sparse evolutionary training of artificial neural networks, an algorithm which evolves an initial sparse topology (Erdős-Rényi random graph) of two consecutive layers of neurons into a scale-free topology, during learning. Our method replaces artificial neural networks fully-connected layers with sparse ones before training, reducing quadratically the number of parameters, with no decrease in accuracy. We demonstrate our claims on restricted Boltzmann machines, multi-layer perceptrons, and convolutional neural networks for unsupervised and supervised learning on 15 datasets. Our approach has the potential to enable artificial neural networks to scale up beyond what is currently possible.

Modeling of mass transfer of Phospholipids in separation process with supercritical CO2 fluid by RBF artificial neural networks

USDA-ARS?s Scientific Manuscript database

An artificial Radial Basis Function (RBF) neural network model was developed for the prediction of mass transfer of the phospholipids from canola meal in supercritical CO2 fluid. The RBF kind of artificial neural networks (ANN) with orthogonal least squares (OLS) learning algorithm were used for mod...

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.

A Deep Learning-Based Method for Similar Patient Question Retrieval in Chinese.

PubMed

Tang, Guo Yu; Ni, Yuan; Xie, Guo Tong; Fan, Xin Li; Shi, Yan Ling

2017-01-01

The online patient question and answering (Q&A) system, either as a website or a mobile application, attracts an increasing number of users in China. Patients will post their questions and the registered doctors then provide the corresponding answers. A large amount of questions with answers from doctors are accumulated. Instead of awaiting the response from a doctor, the newly posted question could be quickly answered by finding a semantically equivalent question from the Q&A achive. In this study, we investigated a novel deep learning based method to retrieve the similar patient question in Chinese. An unsupervised learning algorithm using deep neural network is performed on the corpus to generate the word embedding. The word embedding was then used as the input to a supervised learning algorithm using a designed deep neural network, i.e. the supervised neural attention model (SNA), to predict the similarity between two questions. The experimental results showed that our SNA method achieved P@1 = 77% and P@5 = 84%, which outperformed all other compared methods.

A forecast-based STDP rule suitable for neuromorphic implementation.

PubMed

Davies, S; Galluppi, F; Rast, A D; Furber, S B

2012-08-01

Artificial neural networks increasingly involve spiking dynamics to permit greater computational efficiency. This becomes especially attractive for on-chip implementation using dedicated neuromorphic hardware. However, both spiking neural networks and neuromorphic hardware have historically found difficulties in implementing efficient, effective learning rules. The best-known spiking neural network learning paradigm is Spike Timing Dependent Plasticity (STDP) which adjusts the strength of a connection in response to the time difference between the pre- and post-synaptic spikes. Approaches that relate learning features to the membrane potential of the post-synaptic neuron have emerged as possible alternatives to the more common STDP rule, with various implementations and approximations. Here we use a new type of neuromorphic hardware, SpiNNaker, which represents the flexible "neuromimetic" architecture, to demonstrate a new approach to this problem. Based on the standard STDP algorithm with modifications and approximations, a new rule, called STDP TTS (Time-To-Spike) relates the membrane potential with the Long Term Potentiation (LTP) part of the basic STDP rule. Meanwhile, we use the standard STDP rule for the Long Term Depression (LTD) part of the algorithm. We show that on the basis of the membrane potential it is possible to make a statistical prediction of the time needed by the neuron to reach the threshold, and therefore the LTP part of the STDP algorithm can be triggered when the neuron receives a spike. In our system these approximations allow efficient memory access, reducing the overall computational time and the memory bandwidth required. The improvements here presented are significant for real-time applications such as the ones for which the SpiNNaker system has been designed. We present simulation results that show the efficacy of this algorithm using one or more input patterns repeated over the whole time of the simulation. On-chip results show that the STDP TTS algorithm allows the neural network to adapt and detect the incoming pattern with improvements both in the reliability of, and the time required for, consistent output. Through the approximations we suggest in this paper, we introduce a learning rule that is easy to implement both in event-driven simulators and in dedicated hardware, reducing computational complexity relative to the standard STDP rule. Such a rule offers a promising solution, complementary to standard STDP evaluation algorithms, for real-time learning using spiking neural networks in time-critical applications. Copyright © 2012 Elsevier Ltd. All rights reserved.

Artificial Intelligence in Prediction of Secondary Protein Structure Using CB513 Database

PubMed Central

Avdagic, Zikrija; Purisevic, Elvir; Omanovic, Samir; Coralic, Zlatan

2009-01-01

In this paper we describe CB513 a non-redundant dataset, suitable for development of algorithms for prediction of secondary protein structure. A program was made in Borland Delphi for transforming data from our dataset to make it suitable for learning of neural network for prediction of secondary protein structure implemented in MATLAB Neural-Network Toolbox. Learning (training and testing) of neural network is researched with different sizes of windows, different number of neurons in the hidden layer and different number of training epochs, while using dataset CB513. PMID:21347158

Criteria for Choosing the Best Neural Network: Part 1

DTIC Science & Technology

1991-07-24

Touretzky, pp. 177-185. San Mateo: Morgan Kaufmann. Harp, S.A., Samad , T., and Guha, A . (1990). Designing application-specific neural networks using genetic...determining a parsimonious neural network for use in prediction/generalization based on a given fixed learning sample. Both the classification and...statistical settings, algorithms for selecting the number of hidden layer nodes in a three layer, feedforward neural network are presented. The selection

Neural network based load and price forecasting and confidence interval estimation in deregulated power markets

NASA Astrophysics Data System (ADS)

Zhang, Li

With the deregulation of the electric power market in New England, an independent system operator (ISO) has been separated from the New England Power Pool (NEPOOL). The ISO provides a regional spot market, with bids on various electricity-related products and services submitted by utilities and independent power producers. A utility can bid on the spot market and buy or sell electricity via bilateral transactions. Good estimation of market clearing prices (MCP) will help utilities and independent power producers determine bidding and transaction strategies with low risks, and this is crucial for utilities to compete in the deregulated environment. MCP prediction, however, is difficult since bidding strategies used by participants are complicated and MCP is a non-stationary process. The main objective of this research is to provide efficient short-term load and MCP forecasting and corresponding confidence interval estimation methodologies. In this research, the complexity of load and MCP with other factors is investigated, and neural networks are used to model the complex relationship between input and output. With improved learning algorithm and on-line update features for load forecasting, a neural network based load forecaster was developed, and has been in daily industry use since summer 1998 with good performance. MCP is volatile because of the complexity of market behaviors. In practice, neural network based MCP predictors usually have a cascaded structure, as several key input factors need to be estimated first. In this research, the uncertainties involved in a cascaded neural network structure for MCP prediction are analyzed, and prediction distribution under the Bayesian framework is developed. A fast algorithm to evaluate the confidence intervals by using the memoryless Quasi-Newton method is also developed. The traditional back-propagation algorithm for neural network learning needs to be improved since MCP is a non-stationary process. The extended Kalman filter (EKF) can be used as an integrated adaptive learning and confidence interval estimation algorithm for neural networks, with fast convergence and small confidence intervals. However, EKF learning is computationally expensive because it involves high dimensional matrix manipulations. A modified U-D factorization within the decoupled EKF (DEKF-UD) framework is developed in this research. The computational efficiency and numerical stability are significantly improved.

DANoC: An Efficient Algorithm and Hardware Codesign of Deep Neural Networks on Chip.

PubMed

Zhou, Xichuan; Li, Shengli; Tang, Fang; Hu, Shengdong; Lin, Zhi; Zhang, Lei

2017-07-18

Deep neural networks (NNs) are the state-of-the-art models for understanding the content of images and videos. However, implementing deep NNs in embedded systems is a challenging task, e.g., a typical deep belief network could exhaust gigabytes of memory and result in bandwidth and computational bottlenecks. To address this challenge, this paper presents an algorithm and hardware codesign for efficient deep neural computation. A hardware-oriented deep learning algorithm, named the deep adaptive network, is proposed to explore the sparsity of neural connections. By adaptively removing the majority of neural connections and robustly representing the reserved connections using binary integers, the proposed algorithm could save up to 99.9% memory utility and computational resources without undermining classification accuracy. An efficient sparse-mapping-memory-based hardware architecture is proposed to fully take advantage of the algorithmic optimization. Different from traditional Von Neumann architecture, the deep-adaptive network on chip (DANoC) brings communication and computation in close proximity to avoid power-hungry parameter transfers between on-board memory and on-chip computational units. Experiments over different image classification benchmarks show that the DANoC system achieves competitively high accuracy and efficiency comparing with the state-of-the-art approaches.

Convolutional Neural Network Based on Extreme Learning Machine for Maritime Ships Recognition in Infrared Images.

PubMed

Khellal, Atmane; Ma, Hongbin; Fei, Qing

2018-05-09

The success of Deep Learning models, notably convolutional neural networks (CNNs), makes them the favorable solution for object recognition systems in both visible and infrared domains. However, the lack of training data in the case of maritime ships research leads to poor performance due to the problem of overfitting. In addition, the back-propagation algorithm used to train CNN is very slow and requires tuning many hyperparameters. To overcome these weaknesses, we introduce a new approach fully based on Extreme Learning Machine (ELM) to learn useful CNN features and perform a fast and accurate classification, which is suitable for infrared-based recognition systems. The proposed approach combines an ELM based learning algorithm to train CNN for discriminative features extraction and an ELM based ensemble for classification. The experimental results on VAIS dataset, which is the largest dataset of maritime ships, confirm that the proposed approach outperforms the state-of-the-art models in term of generalization performance and training speed. For instance, the proposed model is up to 950 times faster than the traditional back-propagation based training of convolutional neural networks, primarily for low-level features extraction.

Decoding of finger trajectory from ECoG using deep learning.

PubMed

Xie, Ziqian; Schwartz, Odelia; Prasad, Abhishek

2018-06-01

Conventional decoding pipeline for brain-machine interfaces (BMIs) consists of chained different stages of feature extraction, time-frequency analysis and statistical learning models. Each of these stages uses a different algorithm trained in a sequential manner, which makes it difficult to make the whole system adaptive. The goal was to create an adaptive online system with a single objective function and a single learning algorithm so that the whole system can be trained in parallel to increase the decoding performance. Here, we used deep neural networks consisting of convolutional neural networks (CNN) and a special kind of recurrent neural network (RNN) called long short term memory (LSTM) to address these needs. We used electrocorticography (ECoG) data collected by Kubanek et al. The task consisted of individual finger flexions upon a visual cue. Our model combined a hierarchical feature extractor CNN and a RNN that was able to process sequential data and recognize temporal dynamics in the neural data. CNN was used as the feature extractor and LSTM was used as the regression algorithm to capture the temporal dynamics of the signal. We predicted the finger trajectory using ECoG signals and compared results for the least angle regression (LARS), CNN-LSTM, random forest, LSTM model (LSTM_HC, for using hard-coded features) and a decoding pipeline consisting of band-pass filtering, energy extraction, feature selection and linear regression. The results showed that the deep learning models performed better than the commonly used linear model. The deep learning models not only gave smoother and more realistic trajectories but also learned the transition between movement and rest state. This study demonstrated a decoding network for BMI that involved a convolutional and recurrent neural network model. It integrated the feature extraction pipeline into the convolution and pooling layer and used LSTM layer to capture the state transitions. The discussed network eliminated the need to separately train the model at each step in the decoding pipeline. The whole system can be jointly optimized using stochastic gradient descent and is capable of online learning.

Decoding of finger trajectory from ECoG using deep learning

NASA Astrophysics Data System (ADS)

Xie, Ziqian; Schwartz, Odelia; Prasad, Abhishek

2018-06-01

Objective. Conventional decoding pipeline for brain-machine interfaces (BMIs) consists of chained different stages of feature extraction, time-frequency analysis and statistical learning models. Each of these stages uses a different algorithm trained in a sequential manner, which makes it difficult to make the whole system adaptive. The goal was to create an adaptive online system with a single objective function and a single learning algorithm so that the whole system can be trained in parallel to increase the decoding performance. Here, we used deep neural networks consisting of convolutional neural networks (CNN) and a special kind of recurrent neural network (RNN) called long short term memory (LSTM) to address these needs. Approach. We used electrocorticography (ECoG) data collected by Kubanek et al. The task consisted of individual finger flexions upon a visual cue. Our model combined a hierarchical feature extractor CNN and a RNN that was able to process sequential data and recognize temporal dynamics in the neural data. CNN was used as the feature extractor and LSTM was used as the regression algorithm to capture the temporal dynamics of the signal. Main results. We predicted the finger trajectory using ECoG signals and compared results for the least angle regression (LARS), CNN-LSTM, random forest, LSTM model (LSTM_HC, for using hard-coded features) and a decoding pipeline consisting of band-pass filtering, energy extraction, feature selection and linear regression. The results showed that the deep learning models performed better than the commonly used linear model. The deep learning models not only gave smoother and more realistic trajectories but also learned the transition between movement and rest state. Significance. This study demonstrated a decoding network for BMI that involved a convolutional and recurrent neural network model. It integrated the feature extraction pipeline into the convolution and pooling layer and used LSTM layer to capture the state transitions. The discussed network eliminated the need to separately train the model at each step in the decoding pipeline. The whole system can be jointly optimized using stochastic gradient descent and is capable of online learning.

Online adaptive neural control of a robotic lower limb prosthesis

NASA Astrophysics Data System (ADS)

Spanias, J. A.; Simon, A. M.; Finucane, S. B.; Perreault, E. J.; Hargrove, L. J.

2018-02-01

Objective. The purpose of this study was to develop and evaluate an adaptive intent recognition algorithm that continuously learns to incorporate a lower limb amputee’s neural information (acquired via electromyography (EMG)) as they ambulate with a robotic leg prosthesis. Approach. We present a powered lower limb prosthesis that was configured to acquire the user’s neural information and kinetic/kinematic information from embedded mechanical sensors, and identify and respond to the user’s intent. We conducted an experiment with eight transfemoral amputees over multiple days. EMG and mechanical sensor data were collected while subjects using a powered knee/ankle prosthesis completed various ambulation activities such as walking on level ground, stairs, and ramps. Our adaptive intent recognition algorithm automatically transitioned the prosthesis into the different locomotion modes and continuously updated the user’s model of neural data during ambulation. Main results. Our proposed algorithm accurately and consistently identified the user’s intent over multiple days, despite changing neural signals. The algorithm incorporated 96.31% [0.91%] (mean, [standard error]) of neural information across multiple experimental sessions, and outperformed non-adaptive versions of our algorithm—with a 6.66% [3.16%] relative decrease in error rate. Significance. This study demonstrates that our adaptive intent recognition algorithm enables incorporation of neural information over long periods of use, allowing assistive robotic devices to accurately respond to the user’s intent with low error rates.

A neural learning classifier system with self-adaptive constructivism for mobile robot control.

PubMed

Hurst, Jacob; Bull, Larry

2006-01-01

For artificial entities to achieve true autonomy and display complex lifelike behavior, they will need to exploit appropriate adaptable learning algorithms. In this context adaptability implies flexibility guided by the environment at any given time and an open-ended ability to learn appropriate behaviors. This article examines the use of constructivism-inspired mechanisms within a neural learning classifier system architecture that exploits parameter self-adaptation as an approach to realize such behavior. The system uses a rule structure in which each rule is represented by an artificial neural network. It is shown that appropriate internal rule complexity emerges during learning at a rate controlled by the learner and that the structure indicates underlying features of the task. Results are presented in simulated mazes before moving to a mobile robot platform.

Prototype-Incorporated Emotional Neural Network.

PubMed

Oyedotun, Oyebade K; Khashman, Adnan

2017-08-15

Artificial neural networks (ANNs) aim to simulate the biological neural activities. Interestingly, many ''engineering'' prospects in ANN have relied on motivations from cognition and psychology studies. So far, two important learning theories that have been subject of active research are the prototype and adaptive learning theories. The learning rules employed for ANNs can be related to adaptive learning theory, where several examples of the different classes in a task are supplied to the network for adjusting internal parameters. Conversely, the prototype-learning theory uses prototypes (representative examples); usually, one prototype per class of the different classes contained in the task. These prototypes are supplied for systematic matching with new examples so that class association can be achieved. In this paper, we propose and implement a novel neural network algorithm based on modifying the emotional neural network (EmNN) model to unify the prototype- and adaptive-learning theories. We refer to our new model as ``prototype-incorporated EmNN''. Furthermore, we apply the proposed model to two real-life challenging tasks, namely, static hand-gesture recognition and face recognition, and compare the result to those obtained using the popular back-propagation neural network (BPNN), emotional BPNN (EmNN), deep networks, an exemplar classification model, and k-nearest neighbor.

Indirect adaptive fuzzy wavelet neural network with self- recurrent consequent part for AC servo system.

PubMed

Hou, Runmin; Wang, Li; Gao, Qiang; Hou, Yuanglong; Wang, Chao

2017-09-01

This paper proposes a novel indirect adaptive fuzzy wavelet neural network (IAFWNN) to control the nonlinearity, wide variations in loads, time-variation and uncertain disturbance of the ac servo system. In the proposed approach, the self-recurrent wavelet neural network (SRWNN) is employed to construct an adaptive self-recurrent consequent part for each fuzzy rule of TSK fuzzy model. For the IAFWNN controller, the online learning algorithm is based on back propagation (BP) algorithm. Moreover, an improved particle swarm optimization (IPSO) is used to adapt the learning rate. The aid of an adaptive SRWNN identifier offers the real-time gradient information to the adaptive fuzzy wavelet neural controller to overcome the impact of parameter variations, load disturbances and other uncertainties effectively, and has a good dynamic. The asymptotical stability of the system is guaranteed by using the Lyapunov method. The result of the simulation and the prototype test prove that the proposed are effective and suitable. Copyright © 2017. Published by Elsevier Ltd.

Support vector machines

NASA Technical Reports Server (NTRS)

Garay, Michael J.; Mazzoni, Dominic; Davies, Roger; Wagstaff, Kiri

2004-01-01

Support Vector Machines (SVMs) are a type of supervised learning algorith,, other examples of which are Artificial Neural Networks (ANNs), Decision Trees, and Naive Bayesian Classifiers. Supervised learning algorithms are used to classify objects labled by a 'supervisor' - typically a human 'expert.'.

SAGRAD: A Program for Neural Network Training with Simulated Annealing and the Conjugate Gradient Method.

PubMed

Bernal, Javier; Torres-Jimenez, Jose

2015-01-01

SAGRAD (Simulated Annealing GRADient), a Fortran 77 program for computing neural networks for classification using batch learning, is discussed. Neural network training in SAGRAD is based on a combination of simulated annealing and Møller's scaled conjugate gradient algorithm, the latter a variation of the traditional conjugate gradient method, better suited for the nonquadratic nature of neural networks. Different aspects of the implementation of the training process in SAGRAD are discussed, such as the efficient computation of gradients and multiplication of vectors by Hessian matrices that are required by Møller's algorithm; the (re)initialization of weights with simulated annealing required to (re)start Møller's algorithm the first time and each time thereafter that it shows insufficient progress in reaching a possibly local minimum; and the use of simulated annealing when Møller's algorithm, after possibly making considerable progress, becomes stuck at a local minimum or flat area of weight space. Outlines of the scaled conjugate gradient algorithm, the simulated annealing procedure and the training process used in SAGRAD are presented together with results from running SAGRAD on two examples of training data.

The power of neural nets

NASA Technical Reports Server (NTRS)

Ryan, J. P.; Shah, B. H.

1987-01-01

Implementation of the Hopfield net which is used in the image processing type of applications where only partial information about the image may be available is discussed. The image classification type of algorithm of Hopfield and other learning algorithms, such as the Boltzmann machine and the back-propagation training algorithm, have many vital applications in space.

Automated isotope identification algorithm using artificial neural networks

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

Kamuda, Mark; Stinnett, Jacob; Sullivan, Clair

There is a need to develop an algorithm that can determine the relative activities of radio-isotopes in a large dataset of low-resolution gamma-ray spectra that contains a mixture of many radio-isotopes. Low-resolution gamma-ray spectra that contain mixtures of radio-isotopes often exhibit feature over-lap, requiring algorithms that can analyze these features when overlap occurs. While machine learning and pattern recognition algorithms have shown promise for the problem of radio-isotope identification, their ability to identify and quantify mixtures of radio-isotopes has not been studied. Because machine learning algorithms use abstract features of the spectrum, such as the shape of overlapping peaks andmore » Compton continuum, they are a natural choice for analyzing radio-isotope mixtures. An artificial neural network (ANN) has be trained to calculate the relative activities of 32 radio-isotopes in a spectrum. Furthermore, the ANN is trained with simulated gamma-ray spectra, allowing easy expansion of the library of target radio-isotopes. In this paper we present our initial algorithms based on an ANN and evaluate them against a series measured and simulated spectra.« less

Automated isotope identification algorithm using artificial neural networks

DOE PAGES

Kamuda, Mark; Stinnett, Jacob; Sullivan, Clair

2017-04-12

There is a need to develop an algorithm that can determine the relative activities of radio-isotopes in a large dataset of low-resolution gamma-ray spectra that contains a mixture of many radio-isotopes. Low-resolution gamma-ray spectra that contain mixtures of radio-isotopes often exhibit feature over-lap, requiring algorithms that can analyze these features when overlap occurs. While machine learning and pattern recognition algorithms have shown promise for the problem of radio-isotope identification, their ability to identify and quantify mixtures of radio-isotopes has not been studied. Because machine learning algorithms use abstract features of the spectrum, such as the shape of overlapping peaks andmore » Compton continuum, they are a natural choice for analyzing radio-isotope mixtures. An artificial neural network (ANN) has be trained to calculate the relative activities of 32 radio-isotopes in a spectrum. Furthermore, the ANN is trained with simulated gamma-ray spectra, allowing easy expansion of the library of target radio-isotopes. In this paper we present our initial algorithms based on an ANN and evaluate them against a series measured and simulated spectra.« less

Data-Driven Property Estimation for Protective Clothing

DTIC Science & Technology

2014-09-01

reliable predictions falls under the rubric “machine learning”. Inspired by the applications of machine learning in pharmaceutical drug design and...using genetic algorithms, for instance— descriptor selection can be automated as well. A well-known structured learning technique—Artificial Neural...descriptors automatically, by iteration, e.g., using a genetic algorithm [49]. 4.2.4 Avoiding Overfitting A peril of all regression—least squares as

Face recognition via Gabor and convolutional neural network

NASA Astrophysics Data System (ADS)

Lu, Tongwei; Wu, Menglu; Lu, Tao

2018-04-01

In recent years, the powerful feature learning and classification ability of convolutional neural network have attracted widely attention. Compared with the deep learning, the traditional machine learning algorithm has a good explanatory which deep learning does not have. Thus, In this paper, we propose a method to extract the feature of the traditional algorithm as the input of convolution neural network. In order to reduce the complexity of the network, the kernel function of Gabor wavelet is used to extract the feature from different position, frequency and direction of target image. It is sensitive to edge of image which can provide good direction and scale selection. The extraction of the image from eight directions on a scale are as the input of network that we proposed. The network have the advantage of weight sharing and local connection and texture feature of the input image can reduce the influence of facial expression, gesture and illumination. At the same time, we introduced a layer which combined the results of the pooling and convolution can extract deeper features. The training network used the open source caffe framework which is beneficial to feature extraction. The experiment results of the proposed method proved that the network structure effectively overcame the barrier of illumination and had a good robustness as well as more accurate and rapid than the traditional algorithm.

Predictive representations can link model-based reinforcement learning to model-free mechanisms.

PubMed

Russek, Evan M; Momennejad, Ida; Botvinick, Matthew M; Gershman, Samuel J; Daw, Nathaniel D

2017-09-01

Humans and animals are capable of evaluating actions by considering their long-run future rewards through a process described using model-based reinforcement learning (RL) algorithms. The mechanisms by which neural circuits perform the computations prescribed by model-based RL remain largely unknown; however, multiple lines of evidence suggest that neural circuits supporting model-based behavior are structurally homologous to and overlapping with those thought to carry out model-free temporal difference (TD) learning. Here, we lay out a family of approaches by which model-based computation may be built upon a core of TD learning. The foundation of this framework is the successor representation, a predictive state representation that, when combined with TD learning of value predictions, can produce a subset of the behaviors associated with model-based learning, while requiring less decision-time computation than dynamic programming. Using simulations, we delineate the precise behavioral capabilities enabled by evaluating actions using this approach, and compare them to those demonstrated by biological organisms. We then introduce two new algorithms that build upon the successor representation while progressively mitigating its limitations. Because this framework can account for the full range of observed putatively model-based behaviors while still utilizing a core TD framework, we suggest that it represents a neurally plausible family of mechanisms for model-based evaluation.

Predictive representations can link model-based reinforcement learning to model-free mechanisms

PubMed Central

Botvinick, Matthew M.

2017-01-01

Humans and animals are capable of evaluating actions by considering their long-run future rewards through a process described using model-based reinforcement learning (RL) algorithms. The mechanisms by which neural circuits perform the computations prescribed by model-based RL remain largely unknown; however, multiple lines of evidence suggest that neural circuits supporting model-based behavior are structurally homologous to and overlapping with those thought to carry out model-free temporal difference (TD) learning. Here, we lay out a family of approaches by which model-based computation may be built upon a core of TD learning. The foundation of this framework is the successor representation, a predictive state representation that, when combined with TD learning of value predictions, can produce a subset of the behaviors associated with model-based learning, while requiring less decision-time computation than dynamic programming. Using simulations, we delineate the precise behavioral capabilities enabled by evaluating actions using this approach, and compare them to those demonstrated by biological organisms. We then introduce two new algorithms that build upon the successor representation while progressively mitigating its limitations. Because this framework can account for the full range of observed putatively model-based behaviors while still utilizing a core TD framework, we suggest that it represents a neurally plausible family of mechanisms for model-based evaluation. PMID:28945743

Reinforced two-step-ahead weight adjustment technique for online training of recurrent neural networks.

PubMed

Chang, Li-Chiu; Chen, Pin-An; Chang, Fi-John

2012-08-01

A reliable forecast of future events possesses great value. The main purpose of this paper is to propose an innovative learning technique for reinforcing the accuracy of two-step-ahead (2SA) forecasts. The real-time recurrent learning (RTRL) algorithm for recurrent neural networks (RNNs) can effectively model the dynamics of complex processes and has been used successfully in one-step-ahead forecasts for various time series. A reinforced RTRL algorithm for 2SA forecasts using RNNs is proposed in this paper, and its performance is investigated by two famous benchmark time series and a streamflow during flood events in Taiwan. Results demonstrate that the proposed reinforced 2SA RTRL algorithm for RNNs can adequately forecast the benchmark (theoretical) time series, significantly improve the accuracy of flood forecasts, and effectively reduce time-lag effects.

An algorithm for generating modular hierarchical neural network classifiers: a step toward larger scale applications

NASA Astrophysics Data System (ADS)

Roverso, Davide

2003-08-01

Many-class learning is the problem of training a classifier to discriminate among a large number of target classes. Together with the problem of dealing with high-dimensional patterns (i.e. a high-dimensional input space), the many class problem (i.e. a high-dimensional output space) is a major obstacle to be faced when scaling-up classifier systems and algorithms from small pilot applications to large full-scale applications. The Autonomous Recursive Task Decomposition (ARTD) algorithm is here proposed as a solution to the problem of many-class learning. Example applications of ARTD to neural classifier training are also presented. In these examples, improvements in training time are shown to range from 4-fold to more than 30-fold in pattern classification tasks of both static and dynamic character.

Features Extraction of Flotation Froth Images and BP Neural Network Soft-Sensor Model of Concentrate Grade Optimized by Shuffled Cuckoo Searching Algorithm

PubMed Central

Wang, Jie-sheng; Han, Shuang; Shen, Na-na; Li, Shu-xia

2014-01-01

For meeting the forecasting target of key technology indicators in the flotation process, a BP neural network soft-sensor model based on features extraction of flotation froth images and optimized by shuffled cuckoo search algorithm is proposed. Based on the digital image processing technique, the color features in HSI color space, the visual features based on the gray level cooccurrence matrix, and the shape characteristics based on the geometric theory of flotation froth images are extracted, respectively, as the input variables of the proposed soft-sensor model. Then the isometric mapping method is used to reduce the input dimension, the network size, and learning time of BP neural network. Finally, a shuffled cuckoo search algorithm is adopted to optimize the BP neural network soft-sensor model. Simulation results show that the model has better generalization results and prediction accuracy. PMID:25133210

Prediction of Aerodynamic Coefficients for Wind Tunnel Data using a Genetic Algorithm Optimized Neural Network

NASA Technical Reports Server (NTRS)

Rajkumar, T.; Aragon, Cecilia; Bardina, Jorge; Britten, Roy

2002-01-01

A fast, reliable way of predicting aerodynamic coefficients is produced using a neural network optimized by a genetic algorithm. Basic aerodynamic coefficients (e.g. lift, drag, pitching moment) are modelled as functions of angle of attack and Mach number. The neural network is first trained on a relatively rich set of data from wind tunnel tests of numerical simulations to learn an overall model. Most of the aerodynamic parameters can be well-fitted using polynomial functions. A new set of data, which can be relatively sparse, is then supplied to the network to produce a new model consistent with the previous model and the new data. Because the new model interpolates realistically between the sparse test data points, it is suitable for use in piloted simulations. The genetic algorithm is used to choose a neural network architecture to give best results, avoiding over-and under-fitting of the test data.

Scalable learning method for feedforward neural networks using minimal-enclosing-ball approximation.

PubMed

Wang, Jun; Deng, Zhaohong; Luo, Xiaoqing; Jiang, Yizhang; Wang, Shitong

2016-06-01

Training feedforward neural networks (FNNs) is one of the most critical issues in FNNs studies. However, most FNNs training methods cannot be directly applied for very large datasets because they have high computational and space complexity. In order to tackle this problem, the CCMEB (Center-Constrained Minimum Enclosing Ball) problem in hidden feature space of FNN is discussed and a novel learning algorithm called HFSR-GCVM (hidden-feature-space regression using generalized core vector machine) is developed accordingly. In HFSR-GCVM, a novel learning criterion using L2-norm penalty-based ε-insensitive function is formulated and the parameters in the hidden nodes are generated randomly independent of the training sets. Moreover, the learning of parameters in its output layer is proved equivalent to a special CCMEB problem in FNN hidden feature space. As most CCMEB approximation based machine learning algorithms, the proposed HFSR-GCVM training algorithm has the following merits: The maximal training time of the HFSR-GCVM training is linear with the size of training datasets and the maximal space consumption is independent of the size of training datasets. The experiments on regression tasks confirm the above conclusions. Copyright © 2016 Elsevier Ltd. All rights reserved.

DCS-Neural-Network Program for Aircraft Control and Testing

NASA Technical Reports Server (NTRS)

Jorgensen, Charles C.

2006-01-01

A computer program implements a dynamic-cell-structure (DCS) artificial neural network that can perform such tasks as learning selected aerodynamic characteristics of an airplane from wind-tunnel test data and computing real-time stability and control derivatives of the airplane for use in feedback linearized control. A DCS neural network is one of several types of neural networks that can incorporate additional nodes in order to rapidly learn increasingly complex relationships between inputs and outputs. In the DCS neural network implemented by the present program, the insertion of nodes is based on accumulated error. A competitive Hebbian learning rule (a supervised-learning rule in which connection weights are adjusted to minimize differences between actual and desired outputs for training examples) is used. A Kohonen-style learning rule (derived from a relatively simple training algorithm, implements a Delaunay triangulation layout of neurons) is used to adjust node positions during training. Neighborhood topology determines which nodes are used to estimate new values. The network learns, starting with two nodes, and adds new nodes sequentially in locations chosen to maximize reductions in global error. At any given time during learning, the error becomes homogeneously distributed over all nodes.

Simple artificial neural networks that match probability and exploit and explore when confronting a multiarmed bandit.

PubMed

Dawson, Michael R W; Dupuis, Brian; Spetch, Marcia L; Kelly, Debbie M

2009-08-01

The matching law (Herrnstein 1961) states that response rates become proportional to reinforcement rates; this is related to the empirical phenomenon called probability matching (Vulkan 2000). Here, we show that a simple artificial neural network generates responses consistent with probability matching. This behavior was then used to create an operant procedure for network learning. We use the multiarmed bandit (Gittins 1989), a classic problem of choice behavior, to illustrate that operant training balances exploiting the bandit arm expected to pay off most frequently with exploring other arms. Perceptrons provide a medium for relating results from neural networks, genetic algorithms, animal learning, contingency theory, reinforcement learning, and theories of choice.

Predicting coronary artery disease using different artificial neural network models.

PubMed

Colak, M Cengiz; Colak, Cemil; Kocatürk, Hasan; Sağiroğlu, Seref; Barutçu, Irfan

2008-08-01

Eight different learning algorithms used for creating artificial neural network (ANN) models and the different ANN models in the prediction of coronary artery disease (CAD) are introduced. This work was carried out as a retrospective case-control study. Overall, 124 consecutive patients who had been diagnosed with CAD by coronary angiography (at least 1 coronary stenosis > 50% in major epicardial arteries) were enrolled in the work. Angiographically, the 113 people (group 2) with normal coronary arteries were taken as control subjects. Multi-layered perceptrons ANN architecture were applied. The ANN models trained with different learning algorithms were performed in 237 records, divided into training (n=171) and testing (n=66) data sets. The performance of prediction was evaluated by sensitivity, specificity and accuracy values based on standard definitions. The results have demonstrated that ANN models trained with eight different learning algorithms are promising because of high (greater than 71%) sensitivity, specificity and accuracy values in the prediction of CAD. Accuracy, sensitivity and specificity values varied between 83.63%-100%, 86.46%-100% and 74.67%-100% for training, respectively. For testing, the values were more than 71% for sensitivity, 76% for specificity and 81% for accuracy. It may be proposed that the use of different learning algorithms other than backpropagation and larger sample sizes can improve the performance of prediction. The proposed ANN models trained with these learning algorithms could be used a promising approach for predicting CAD without the need for invasive diagnostic methods and could help in the prognostic clinical decision.

A New Efficient Hybrid Intelligent Model for Biodegradation Process of DMP with Fuzzy Wavelet Neural Networks

NASA Astrophysics Data System (ADS)

Huang, Mingzhi; Zhang, Tao; Ruan, Jujun; Chen, Xiaohong

2017-01-01

A new efficient hybrid intelligent approach based on fuzzy wavelet neural network (FWNN) was proposed for effectively modeling and simulating biodegradation process of Dimethyl phthalate (DMP) in an anaerobic/anoxic/oxic (AAO) wastewater treatment process. With the self learning and memory abilities of neural networks (NN), handling uncertainty capacity of fuzzy logic (FL), analyzing local details superiority of wavelet transform (WT) and global search of genetic algorithm (GA), the proposed hybrid intelligent model can extract the dynamic behavior and complex interrelationships from various water quality variables. For finding the optimal values for parameters of the proposed FWNN, a hybrid learning algorithm integrating an improved genetic optimization and gradient descent algorithm is employed. The results show, compared with NN model (optimized by GA) and kinetic model, the proposed FWNN model have the quicker convergence speed, the higher prediction performance, and smaller RMSE (0.080), MSE (0.0064), MAPE (1.8158) and higher R2 (0.9851) values. which illustrates FWNN model simulates effluent DMP more accurately than the mechanism model.

A New Efficient Hybrid Intelligent Model for Biodegradation Process of DMP with Fuzzy Wavelet Neural Networks

PubMed Central

Huang, Mingzhi; Zhang, Tao; Ruan, Jujun; Chen, Xiaohong

2017-01-01

A new efficient hybrid intelligent approach based on fuzzy wavelet neural network (FWNN) was proposed for effectively modeling and simulating biodegradation process of Dimethyl phthalate (DMP) in an anaerobic/anoxic/oxic (AAO) wastewater treatment process. With the self learning and memory abilities of neural networks (NN), handling uncertainty capacity of fuzzy logic (FL), analyzing local details superiority of wavelet transform (WT) and global search of genetic algorithm (GA), the proposed hybrid intelligent model can extract the dynamic behavior and complex interrelationships from various water quality variables. For finding the optimal values for parameters of the proposed FWNN, a hybrid learning algorithm integrating an improved genetic optimization and gradient descent algorithm is employed. The results show, compared with NN model (optimized by GA) and kinetic model, the proposed FWNN model have the quicker convergence speed, the higher prediction performance, and smaller RMSE (0.080), MSE (0.0064), MAPE (1.8158) and higher R2 (0.9851) values. which illustrates FWNN model simulates effluent DMP more accurately than the mechanism model. PMID:28120889

Classification of multispectral image data by the Binary Diamond neural network and by nonparametric, pixel-by-pixel methods

NASA Technical Reports Server (NTRS)

Salu, Yehuda; Tilton, James

1993-01-01

The classification of multispectral image data obtained from satellites has become an important tool for generating ground cover maps. This study deals with the application of nonparametric pixel-by-pixel classification methods in the classification of pixels, based on their multispectral data. A new neural network, the Binary Diamond, is introduced, and its performance is compared with a nearest neighbor algorithm and a back-propagation network. The Binary Diamond is a multilayer, feed-forward neural network, which learns from examples in unsupervised, 'one-shot' mode. It recruits its neurons according to the actual training set, as it learns. The comparisons of the algorithms were done by using a realistic data base, consisting of approximately 90,000 Landsat 4 Thematic Mapper pixels. The Binary Diamond and the nearest neighbor performances were close, with some advantages to the Binary Diamond. The performance of the back-propagation network lagged behind. An efficient nearest neighbor algorithm, the binned nearest neighbor, is described. Ways for improving the performances, such as merging categories, and analyzing nonboundary pixels, are addressed and evaluated.

Toolkits and Libraries for Deep Learning.

PubMed

Erickson, Bradley J; Korfiatis, Panagiotis; Akkus, Zeynettin; Kline, Timothy; Philbrick, Kenneth

2017-08-01

Deep learning is an important new area of machine learning which encompasses a wide range of neural network architectures designed to complete various tasks. In the medical imaging domain, example tasks include organ segmentation, lesion detection, and tumor classification. The most popular network architecture for deep learning for images is the convolutional neural network (CNN). Whereas traditional machine learning requires determination and calculation of features from which the algorithm learns, deep learning approaches learn the important features as well as the proper weighting of those features to make predictions for new data. In this paper, we will describe some of the libraries and tools that are available to aid in the construction and efficient execution of deep learning as applied to medical images.

Thermodynamic efficiency of learning a rule in neural networks

NASA Astrophysics Data System (ADS)

Goldt, Sebastian; Seifert, Udo

2017-11-01

Biological systems have to build models from their sensory input data that allow them to efficiently process previously unseen inputs. Here, we study a neural network learning a binary classification rule for these inputs from examples provided by a teacher. We analyse the ability of the network to apply the rule to new inputs, that is to generalise from past experience. Using stochastic thermodynamics, we show that the thermodynamic costs of the learning process provide an upper bound on the amount of information that the network is able to learn from its teacher for both batch and online learning. This allows us to introduce a thermodynamic efficiency of learning. We analytically compute the dynamics and the efficiency of a noisy neural network performing online learning in the thermodynamic limit. In particular, we analyse three popular learning algorithms, namely Hebbian, Perceptron and AdaTron learning. Our work extends the methods of stochastic thermodynamics to a new type of learning problem and might form a suitable basis for investigating the thermodynamics of decision-making.

Discovery Learning in Autonomous Agents Using Genetic Algorithms

DTIC Science & Technology

1993-12-01

Meyer and Wilson (47). 65. Roitblat , H. L., et al. "Biomimetic Sonar Processing: Prom Dolphin Echoloc-Ation to Artificial Neural Networks." In Meyer and...34 In Meyer and Wilson (47). 65. Roitblat , H. L., et al. "Biomimetic Sonar Processing: From Dolphin Echolocation to Artificial Neural Networks." In

Two neural network algorithms for designing optimal terminal controllers with open final time

NASA Technical Reports Server (NTRS)

Plumer, Edward S.

1992-01-01

Multilayer neural networks, trained by the backpropagation through time algorithm (BPTT), have been used successfully as state-feedback controllers for nonlinear terminal control problems. Current BPTT techniques, however, are not able to deal systematically with open final-time situations such as minimum-time problems. Two approaches which extend BPTT to open final-time problems are presented. In the first, a neural network learns a mapping from initial-state to time-to-go. In the second, the optimal number of steps for each trial run is found using a line-search. Both methods are derived using Lagrange multiplier techniques. This theoretical framework is used to demonstrate that the derived algorithms are direct extensions of forward/backward sweep methods used in N-stage optimal control. The two algorithms are tested on a Zermelo problem and the resulting trajectories compare favorably to optimal control results.

Real-Time Adaptive Color Segmentation by Neural Networks

NASA Technical Reports Server (NTRS)

Duong, Tuan A.

2004-01-01

Artificial neural networks that would utilize the cascade error projection (CEP) algorithm have been proposed as means of autonomous, real-time, adaptive color segmentation of images that change with time. In the original intended application, such a neural network would be used to analyze digitized color video images of terrain on a remote planet as viewed from an uninhabited spacecraft approaching the planet. During descent toward the surface of the planet, information on the segmentation of the images into differently colored areas would be updated adaptively in real time to capture changes in contrast, brightness, and resolution, all in an effort to identify a safe and scientifically productive landing site and provide control feedback to steer the spacecraft toward that site. Potential terrestrial applications include monitoring images of crops to detect insect invasions and monitoring of buildings and other facilities to detect intruders. The CEP algorithm is reliable and is well suited to implementation in very-large-scale integrated (VLSI) circuitry. It was chosen over other neural-network learning algorithms because it is better suited to realtime learning: It provides a self-evolving neural-network structure, requires fewer iterations to converge and is more tolerant to low resolution (that is, fewer bits) in the quantization of neural-network synaptic weights. Consequently, a CEP neural network learns relatively quickly, and the circuitry needed to implement it is relatively simple. Like other neural networks, a CEP neural network includes an input layer, hidden units, and output units (see figure). As in other neural networks, a CEP network is presented with a succession of input training patterns, giving rise to a set of outputs that are compared with the desired outputs. Also as in other neural networks, the synaptic weights are updated iteratively in an effort to bring the outputs closer to target values. A distinctive feature of the CEP neural network and algorithm is that each update of synaptic weights takes place in conjunction with the addition of another hidden unit, which then remains in place as still other hidden units are added on subsequent iterations. For a given training pattern, the synaptic weight between (1) the inputs and the previously added hidden units and (2) the newly added hidden unit is updated by an amount proportional to the partial derivative of a quadratic error function with respect to the synaptic weight. The synaptic weight between the newly added hidden unit and each output unit is given by a more complex function that involves the errors between the outputs and their target values, the transfer functions (hyperbolic tangents) of the neural units, and the derivatives of the transfer functions.

ANNS An X Window Based Version of the AFIT Neural Network Simulator

DTIC Science & Technology

1993-06-01

programer or user can view the dy- namic behavior of an algorithm and its changes of learning state while the neural network paradigms or algorithms...an object as "something you can do things to. An object has state, behavior , and identity, the structure and behavior of similar objects are defined in...their common class. The terms instance and object are interchangeable" [5:516]. The behavior of an object is "characterized by the actions that it

A self-organizing neural network for the traveling salesman problem that is competitive with simulated annealing.

PubMed

Budinich, M

1996-02-15

Unsupervised learning applied to an unstructured neural network can give approximate solutions to the traveling salesman problem. For 50 cities in the plane this algorithm performs like the elastic net of Durbin and Willshaw (1987) and it improves when increasing the number of cities to get better than simulated annealing for problems with more than 500 cities. In all the tests this algorithm requires a fraction of the time taken by simulated annealing.

CrossSim

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

Plimpton, Steven J.; Agarwal, Sapan; Schiek, Richard

2016-09-02

CrossSim is a simulator for modeling neural-inspired machine learning algorithms on analog hardware, such as resistive memory crossbars. It includes noise models for reading and updating the resistances, which can be based on idealized equations or experimental data. It can also introduce noise and finite precision effects when converting values from digital to analog and vice versa. All of these effects can be turned on or off as an algorithm processes a data set and attempts to learn its salient attributes so that it can be categorized in the machine learning training/classification context. CrossSim thus allows the robustness, accuracy, andmore » energy usage of a machine learning algorithm to be tested on simulated hardware.« less

Improving Maritime Domain Awareness Using Neural Networks for Target of Interest Classification

DTIC Science & Technology

2015-03-01

spreading SCG scaled conjugate gradient xv THIS PAGE INTENTIONALLY LEFT BLANK xvi EXECUTIVE SUMMARY The research detailed in this thesis is a...algorithms were explored for training the neural networks: resilient backpropagation (RP) and scaled conjugate gradient backpropagation ( SCG ). The...results of the neural network training performance are presented using mean squared error convergence plots. In all implementations, the SCG learning

Competitive repetition suppression (CoRe) clustering: a biologically inspired learning model with application to robust clustering.

PubMed

Bacciu, Davide; Starita, Antonina

2008-11-01

Determining a compact neural coding for a set of input stimuli is an issue that encompasses several biological memory mechanisms as well as various artificial neural network models. In particular, establishing the optimal network structure is still an open problem when dealing with unsupervised learning models. In this paper, we introduce a novel learning algorithm, named competitive repetition-suppression (CoRe) learning, inspired by a cortical memory mechanism called repetition suppression (RS). We show how such a mechanism is used, at various levels of the cerebral cortex, to generate compact neural representations of the visual stimuli. From the general CoRe learning model, we derive a clustering algorithm, named CoRe clustering, that can automatically estimate the unknown cluster number from the data without using a priori information concerning the input distribution. We illustrate how CoRe clustering, besides its biological plausibility, posses strong theoretical properties in terms of robustness to noise and outliers, and we provide an error function describing CoRe learning dynamics. Such a description is used to analyze CoRe relationships with the state-of-the art clustering models and to highlight CoRe similitude with rival penalized competitive learning (RPCL), showing how CoRe extends such a model by strengthening the rival penalization estimation by means of loss functions from robust statistics.

Railway obstacle detection algorithm using neural network

NASA Astrophysics Data System (ADS)

Yu, Mingyang; Yang, Peng; Wei, Sen

2018-05-01

Aiming at the difficulty of detection of obstacle in outdoor railway scene, a data-oriented method based on neural network to obtain image objects is proposed. First, we mark objects of images(such as people, trains, animals) acquired on the Internet. and then use the residual learning units to build Fast R-CNN framework. Then, the neural network is trained to get the target image characteristics by using stochastic gradient descent algorithm. Finally, a well-trained model is used to identify an outdoor railway image. if it includes trains and other objects, it will issue an alert. Experiments show that the correct rate of warning reached 94.85%.

Multiprocessor and memory architecture of the neurocomputer SYNAPSE-1.

PubMed

Ramacher, U; Raab, W; Anlauf, J; Hachmann, U; Beichter, J; Brüls, N; Wesseling, M; Sicheneder, E; Männer, R; Glass, J

1993-12-01

A general purpose neurocomputer, SYNAPSE-1, which exhibits a multiprocessor and memory architecture is presented. It offers wide flexibility with respect to neural algorithms and a speed-up factor of several orders of magnitude--including learning. The computational power is provided by a 2-dimensional systolic array of neural signal processors. Since the weights are stored outside these NSPs, memory size and processing power can be adapted individually to the application needs. A neural algorithms programming language, embedded in C(+2) has been defined for the user to cope with the neurocomputer. In a benchmark test, the prototype of SYNAPSE-1 was 8000 times as fast as a standard workstation.

Diagnosing Parkinson's Diseases Using Fuzzy Neural System

PubMed Central

Abiyev, Rahib H.; Abizade, Sanan

2016-01-01

This study presents the design of the recognition system that will discriminate between healthy people and people with Parkinson's disease. A diagnosing of Parkinson's diseases is performed using fusion of the fuzzy system and neural networks. The structure and learning algorithms of the proposed fuzzy neural system (FNS) are presented. The approach described in this paper allows enhancing the capability of the designed system and efficiently distinguishing healthy individuals. It was proved through simulation of the system that has been performed using data obtained from UCI machine learning repository. A comparative study was carried out and the simulation results demonstrated that the proposed fuzzy neural system improves the recognition rate of the designed system. PMID:26881009

System impairment compensation in coherent optical communications by using a bio-inspired detector based on artificial neural network and genetic algorithm

NASA Astrophysics Data System (ADS)

Wang, Danshi; Zhang, Min; Li, Ze; Song, Chuang; Fu, Meixia; Li, Jin; Chen, Xue

2017-09-01

A bio-inspired detector based on the artificial neural network (ANN) and genetic algorithm is proposed in the context of a coherent optical transmission system. The ANN is designed to mitigate 16-quadrature amplitude modulation system impairments, including linear impairment: Gaussian white noise, laser phase noise, in-phase/quadrature component imbalance, and nonlinear impairment: nonlinear phase. Without prior information or heuristic assumptions, the ANN, functioning as a machine learning algorithm, can learn and capture the characteristics of impairments from observed data. Numerical simulations were performed, and dispersion-shifted, dispersion-managed, and dispersion-unmanaged fiber links were investigated. The launch power dynamic range and maximum transmission distance for the bio-inspired method were 2.7 dBm and 240 km greater, respectively, than those of the maximum likelihood estimation algorithm. Moreover, the linewidth tolerance of the bio-inspired technique was 170 kHz greater than that of the k-means method, demonstrating its usability for digital signal processing in coherent systems.

Dreaming of Atmospheres

NASA Astrophysics Data System (ADS)

Waldmann, I. P.

2016-04-01

Here, we introduce the RobERt (Robotic Exoplanet Recognition) algorithm for the classification of exoplanetary emission spectra. Spectral retrieval of exoplanetary atmospheres frequently requires the preselection of molecular/atomic opacities to be defined by the user. In the era of open-source, automated, and self-sufficient retrieval algorithms, manual input should be avoided. User dependent input could, in worst-case scenarios, lead to incomplete models and biases in the retrieval. The RobERt algorithm is based on deep-belief neural (DBN) networks trained to accurately recognize molecular signatures for a wide range of planets, atmospheric thermal profiles, and compositions. Reconstructions of the learned features, also referred to as the “dreams” of the network, indicate good convergence and an accurate representation of molecular features in the DBN. Using these deep neural networks, we work toward retrieval algorithms that themselves understand the nature of the observed spectra, are able to learn from current and past data, and make sensible qualitative preselections of atmospheric opacities to be used for the quantitative stage of the retrieval process.

Neurale Netwerken en Radarsystemen (Neural Networks and Radar Systems)

DTIC Science & Technology

1989-08-01

general issues in cognitive science", Parallel distributed processing, Vol 1: Foundations, Rumelhart et al. 1986 pp 110-146 THO rapport Pagina 151 36 D.E...34Neural networks (part 2)",Expert Focus, IEEE Expert, Spring 1988. 61 J.A. Anderson, " Cognitive and Psychological Computations with Neural Models", IEEE...Pagina 154 69 David H. Ackley, Geoffrey E. Hinton and Terrence J. Sejnowski, "A Learning Algorithm for Boltzmann machines", cognitive science 9, 147-169

Detection of Interictal Discharges With Convolutional Neural Networks Using Discrete Ordered Multichannel Intracranial EEG.

PubMed

Antoniades, Andreas; Spyrou, Loukianos; Martin-Lopez, David; Valentin, Antonio; Alarcon, Gonzalo; Sanei, Saeid; Cheong Took, Clive

2017-12-01

Detection algorithms for electroencephalography (EEG) data, especially in the field of interictal epileptiform discharge (IED) detection, have traditionally employed handcrafted features, which utilized specific characteristics of neural responses. Although these algorithms achieve high accuracy, mere detection of an IED holds little clinical significance. In this paper, we consider deep learning for epileptic subjects to accommodate automatic feature generation from intracranial EEG data, while also providing clinical insight. Convolutional neural networks are trained in a subject independent fashion to demonstrate how meaningful features are automatically learned in a hierarchical process. We illustrate how the convolved filters in the deepest layers provide insight toward the different types of IEDs within the group, as confirmed by our expert clinicians. The morphology of the IEDs found in filters can help evaluate the treatment of a patient. To improve the learning of the deep model, moderately different score classes are utilized as opposed to binary IED and non-IED labels. The resulting model achieves state-of-the-art classification performance and is also invariant to time differences between the IEDs. This paper suggests that deep learning is suitable for automatic feature generation from intracranial EEG data, while also providing insight into the data.

Neural network method for lossless two-conductor transmission line equations based on the IELM algorithm

NASA Astrophysics Data System (ADS)

Yang, Yunlei; Hou, Muzhou; Luo, Jianshu; Liu, Taohua

2018-06-01

With the increasing demands for vast amounts of data and high-speed signal transmission, the use of multi-conductor transmission lines is becoming more common. The impact of transmission lines on signal transmission is thus a key issue affecting the performance of high-speed digital systems. To solve the problem of lossless two-conductor transmission line equations (LTTLEs), a neural network model and algorithm are explored in this paper. By selecting the product of two triangular basis functions as the activation function of hidden layer neurons, we can guarantee the separation of time, space, and phase orthogonality. By adding the initial condition to the neural network, an improved extreme learning machine (IELM) algorithm for solving the network weight is obtained. This is different to the traditional method for converting the initial condition into the iterative constraint condition. Calculation software for solving the LTTLEs based on the IELM algorithm is developed. Numerical experiments show that the results are consistent with those of the traditional method. The proposed neural network algorithm can find the terminal voltage of the transmission line and also the voltage of any observation point. It is possible to calculate the value at any given point by using the neural network model to solve the transmission line equation.

The Livermore Brain: Massive Deep Learning Networks Enabled by High Performance Computing

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

Chen, Barry Y.

The proliferation of inexpensive sensor technologies like the ubiquitous digital image sensors has resulted in the collection and sharing of vast amounts of unsorted and unexploited raw data. Companies and governments who are able to collect and make sense of large datasets to help them make better decisions more rapidly will have a competitive advantage in the information era. Machine Learning technologies play a critical role for automating the data understanding process; however, to be maximally effective, useful intermediate representations of the data are required. These representations or “features” are transformations of the raw data into a form where patternsmore » are more easily recognized. Recent breakthroughs in Deep Learning have made it possible to learn these features from large amounts of labeled data. The focus of this project is to develop and extend Deep Learning algorithms for learning features from vast amounts of unlabeled data and to develop the HPC neural network training platform to support the training of massive network models. This LDRD project succeeded in developing new unsupervised feature learning algorithms for images and video and created a scalable neural network training toolkit for HPC. Additionally, this LDRD helped create the world’s largest freely-available image and video dataset supporting open multimedia research and used this dataset for training our deep neural networks. This research helped LLNL capture several work-for-others (WFO) projects, attract new talent, and establish collaborations with leading academic and commercial partners. Finally, this project demonstrated the successful training of the largest unsupervised image neural network using HPC resources and helped establish LLNL leadership at the intersection of Machine Learning and HPC research.« less

Quantum neural networks: Current status and prospects for development

NASA Astrophysics Data System (ADS)

Altaisky, M. V.; Kaputkina, N. E.; Krylov, V. A.

2014-11-01

The idea of quantum artificial neural networks, first formulated in [34], unites the artificial neural network concept with the quantum computation paradigm. Quantum artificial neural networks were first systematically considered in the PhD thesis by T. Menneer (1998). Based on the works of Menneer and Narayanan [42, 43], Kouda, Matsui, and Nishimura [35, 36], Altaisky [2, 68], Zhou [67], and others, quantum-inspired learning algorithms for neural networks were developed, and are now used in various training programs and computer games [29, 30]. The first practically realizable scaled hardware-implemented model of the quantum artificial neural network is obtained by D-Wave Systems, Inc. [33]. It is a quantum Hopfield network implemented on the basis of superconducting quantum interference devices (SQUIDs). In this work we analyze possibilities and underlying principles of an alternative way to implement quantum neural networks on the basis of quantum dots. A possibility of using quantum neural network algorithms in automated control systems, associative memory devices, and in modeling biological and social networks is examined.

Implementations of back propagation algorithm in ecosystems applications

NASA Astrophysics Data System (ADS)

Ali, Khalda F.; Sulaiman, Riza; Elamir, Amir Mohamed

2015-05-01

Artificial Neural Networks (ANNs) have been applied to an increasing number of real world problems of considerable complexity. Their most important advantage is in solving problems which are too complex for conventional technologies, that do not have an algorithmic solutions or their algorithmic Solutions is too complex to be found. In general, because of their abstraction from the biological brain, ANNs are developed from concept that evolved in the late twentieth century neuro-physiological experiments on the cells of the human brain to overcome the perceived inadequacies with conventional ecological data analysis methods. ANNs have gained increasing attention in ecosystems applications, because of ANN's capacity to detect patterns in data through non-linear relationships, this characteristic confers them a superior predictive ability. In this research, ANNs is applied in an ecological system analysis. The neural networks use the well known Back Propagation (BP) Algorithm with the Delta Rule for adaptation of the system. The Back Propagation (BP) training Algorithm is an effective analytical method for adaptation of the ecosystems applications, the main reason because of their capacity to detect patterns in data through non-linear relationships. This characteristic confers them a superior predicting ability. The BP algorithm uses supervised learning, which means that we provide the algorithm with examples of the inputs and outputs we want the network to compute, and then the error is calculated. The idea of the back propagation algorithm is to reduce this error, until the ANNs learns the training data. The training begins with random weights, and the goal is to adjust them so that the error will be minimal. This research evaluated the use of artificial neural networks (ANNs) techniques in an ecological system analysis and modeling. The experimental results from this research demonstrate that an artificial neural network system can be trained to act as an expert ecosystem analyzer for many applications in ecological fields. The pilot ecosystem analyzer shows promising ability for generalization and requires further tuning and refinement of the basis neural network system for optimal performance.

A spiking neural network model of model-free reinforcement learning with high-dimensional sensory input and perceptual ambiguity.

PubMed

Nakano, Takashi; Otsuka, Makoto; Yoshimoto, Junichiro; Doya, Kenji

2015-01-01

A theoretical framework of reinforcement learning plays an important role in understanding action selection in animals. Spiking neural networks provide a theoretically grounded means to test computational hypotheses on neurally plausible algorithms of reinforcement learning through numerical simulation. However, most of these models cannot handle observations which are noisy, or occurred in the past, even though these are inevitable and constraining features of learning in real environments. This class of problem is formally known as partially observable reinforcement learning (PORL) problems. It provides a generalization of reinforcement learning to partially observable domains. In addition, observations in the real world tend to be rich and high-dimensional. In this work, we use a spiking neural network model to approximate the free energy of a restricted Boltzmann machine and apply it to the solution of PORL problems with high-dimensional observations. Our spiking network model solves maze tasks with perceptually ambiguous high-dimensional observations without knowledge of the true environment. An extended model with working memory also solves history-dependent tasks. The way spiking neural networks handle PORL problems may provide a glimpse into the underlying laws of neural information processing which can only be discovered through such a top-down approach.

A Spiking Neural Network Model of Model-Free Reinforcement Learning with High-Dimensional Sensory Input and Perceptual Ambiguity

PubMed Central

Nakano, Takashi; Otsuka, Makoto; Yoshimoto, Junichiro; Doya, Kenji

2015-01-01

A theoretical framework of reinforcement learning plays an important role in understanding action selection in animals. Spiking neural networks provide a theoretically grounded means to test computational hypotheses on neurally plausible algorithms of reinforcement learning through numerical simulation. However, most of these models cannot handle observations which are noisy, or occurred in the past, even though these are inevitable and constraining features of learning in real environments. This class of problem is formally known as partially observable reinforcement learning (PORL) problems. It provides a generalization of reinforcement learning to partially observable domains. In addition, observations in the real world tend to be rich and high-dimensional. In this work, we use a spiking neural network model to approximate the free energy of a restricted Boltzmann machine and apply it to the solution of PORL problems with high-dimensional observations. Our spiking network model solves maze tasks with perceptually ambiguous high-dimensional observations without knowledge of the true environment. An extended model with working memory also solves history-dependent tasks. The way spiking neural networks handle PORL problems may provide a glimpse into the underlying laws of neural information processing which can only be discovered through such a top-down approach. PMID:25734662

Neural Modularity Helps Organisms Evolve to Learn New Skills without Forgetting Old Skills

PubMed Central

Ellefsen, Kai Olav; Mouret, Jean-Baptiste; Clune, Jeff

2015-01-01

A long-standing goal in artificial intelligence is creating agents that can learn a variety of different skills for different problems. In the artificial intelligence subfield of neural networks, a barrier to that goal is that when agents learn a new skill they typically do so by losing previously acquired skills, a problem called catastrophic forgetting. That occurs because, to learn the new task, neural learning algorithms change connections that encode previously acquired skills. How networks are organized critically affects their learning dynamics. In this paper, we test whether catastrophic forgetting can be reduced by evolving modular neural networks. Modularity intuitively should reduce learning interference between tasks by separating functionality into physically distinct modules in which learning can be selectively turned on or off. Modularity can further improve learning by having a reinforcement learning module separate from sensory processing modules, allowing learning to happen only in response to a positive or negative reward. In this paper, learning takes place via neuromodulation, which allows agents to selectively change the rate of learning for each neural connection based on environmental stimuli (e.g. to alter learning in specific locations based on the task at hand). To produce modularity, we evolve neural networks with a cost for neural connections. We show that this connection cost technique causes modularity, confirming a previous result, and that such sparsely connected, modular networks have higher overall performance because they learn new skills faster while retaining old skills more and because they have a separate reinforcement learning module. Our results suggest (1) that encouraging modularity in neural networks may help us overcome the long-standing barrier of networks that cannot learn new skills without forgetting old ones, and (2) that one benefit of the modularity ubiquitous in the brains of natural animals might be to alleviate the problem of catastrophic forgetting. PMID:25837826

Neural modularity helps organisms evolve to learn new skills without forgetting old skills.

PubMed

Ellefsen, Kai Olav; Mouret, Jean-Baptiste; Clune, Jeff

2015-04-01

A long-standing goal in artificial intelligence is creating agents that can learn a variety of different skills for different problems. In the artificial intelligence subfield of neural networks, a barrier to that goal is that when agents learn a new skill they typically do so by losing previously acquired skills, a problem called catastrophic forgetting. That occurs because, to learn the new task, neural learning algorithms change connections that encode previously acquired skills. How networks are organized critically affects their learning dynamics. In this paper, we test whether catastrophic forgetting can be reduced by evolving modular neural networks. Modularity intuitively should reduce learning interference between tasks by separating functionality into physically distinct modules in which learning can be selectively turned on or off. Modularity can further improve learning by having a reinforcement learning module separate from sensory processing modules, allowing learning to happen only in response to a positive or negative reward. In this paper, learning takes place via neuromodulation, which allows agents to selectively change the rate of learning for each neural connection based on environmental stimuli (e.g. to alter learning in specific locations based on the task at hand). To produce modularity, we evolve neural networks with a cost for neural connections. We show that this connection cost technique causes modularity, confirming a previous result, and that such sparsely connected, modular networks have higher overall performance because they learn new skills faster while retaining old skills more and because they have a separate reinforcement learning module. Our results suggest (1) that encouraging modularity in neural networks may help us overcome the long-standing barrier of networks that cannot learn new skills without forgetting old ones, and (2) that one benefit of the modularity ubiquitous in the brains of natural animals might be to alleviate the problem of catastrophic forgetting.

Comparison of rule induction, decision trees and formal concept analysis approaches for classification

NASA Astrophysics Data System (ADS)

Kotelnikov, E. V.; Milov, V. R.

2018-05-01

Rule-based learning algorithms have higher transparency and easiness to interpret in comparison with neural networks and deep learning algorithms. These properties make it possible to effectively use such algorithms to solve descriptive tasks of data mining. The choice of an algorithm depends also on its ability to solve predictive tasks. The article compares the quality of the solution of the problems with binary and multiclass classification based on the experiments with six datasets from the UCI Machine Learning Repository. The authors investigate three algorithms: Ripper (rule induction), C4.5 (decision trees), In-Close (formal concept analysis). The results of the experiments show that In-Close demonstrates the best quality of classification in comparison with Ripper and C4.5, however the latter two generate more compact rule sets.

Beam-column joint shear prediction using hybridized deep learning neural network with genetic algorithm

NASA Astrophysics Data System (ADS)

Mundher Yaseen, Zaher; Abdulmohsin Afan, Haitham; Tran, Minh-Tung

2018-04-01

Scientifically evidenced that beam-column joints are a critical point in the reinforced concrete (RC) structure under the fluctuation loads effects. In this novel hybrid data-intelligence model developed to predict the joint shear behavior of exterior beam-column structure frame. The hybrid data-intelligence model is called genetic algorithm integrated with deep learning neural network model (GA-DLNN). The genetic algorithm is used as prior modelling phase for the input approximation whereas the DLNN predictive model is used for the prediction phase. To demonstrate this structural problem, experimental data is collected from the literature that defined the dimensional and specimens’ properties. The attained findings evidenced the efficitveness of the hybrid GA-DLNN in modelling beam-column joint shear problem. In addition, the accurate prediction achived with less input variables owing to the feasibility of the evolutionary phase.

Maximum entropy methods for extracting the learned features of deep neural networks.

PubMed

Finnegan, Alex; Song, Jun S

2017-10-01

New architectures of multilayer artificial neural networks and new methods for training them are rapidly revolutionizing the application of machine learning in diverse fields, including business, social science, physical sciences, and biology. Interpreting deep neural networks, however, currently remains elusive, and a critical challenge lies in understanding which meaningful features a network is actually learning. We present a general method for interpreting deep neural networks and extracting network-learned features from input data. We describe our algorithm in the context of biological sequence analysis. Our approach, based on ideas from statistical physics, samples from the maximum entropy distribution over possible sequences, anchored at an input sequence and subject to constraints implied by the empirical function learned by a network. Using our framework, we demonstrate that local transcription factor binding motifs can be identified from a network trained on ChIP-seq data and that nucleosome positioning signals are indeed learned by a network trained on chemical cleavage nucleosome maps. Imposing a further constraint on the maximum entropy distribution also allows us to probe whether a network is learning global sequence features, such as the high GC content in nucleosome-rich regions. This work thus provides valuable mathematical tools for interpreting and extracting learned features from feed-forward neural networks.

A Grey Wolf Optimizer for Modular Granular Neural Networks for Human Recognition

PubMed Central

Sánchez, Daniela; Melin, Patricia

2017-01-01

A grey wolf optimizer for modular neural network (MNN) with a granular approach is proposed. The proposed method performs optimal granulation of data and design of modular neural networks architectures to perform human recognition, and to prove its effectiveness benchmark databases of ear, iris, and face biometric measures are used to perform tests and comparisons against other works. The design of a modular granular neural network (MGNN) consists in finding optimal parameters of its architecture; these parameters are the number of subgranules, percentage of data for the training phase, learning algorithm, goal error, number of hidden layers, and their number of neurons. Nowadays, there is a great variety of approaches and new techniques within the evolutionary computing area, and these approaches and techniques have emerged to help find optimal solutions to problems or models and bioinspired algorithms are part of this area. In this work a grey wolf optimizer is proposed for the design of modular granular neural networks, and the results are compared against a genetic algorithm and a firefly algorithm in order to know which of these techniques provides better results when applied to human recognition. PMID:28894461

A Grey Wolf Optimizer for Modular Granular Neural Networks for Human Recognition.

PubMed

Sánchez, Daniela; Melin, Patricia; Castillo, Oscar

2017-01-01

A grey wolf optimizer for modular neural network (MNN) with a granular approach is proposed. The proposed method performs optimal granulation of data and design of modular neural networks architectures to perform human recognition, and to prove its effectiveness benchmark databases of ear, iris, and face biometric measures are used to perform tests and comparisons against other works. The design of a modular granular neural network (MGNN) consists in finding optimal parameters of its architecture; these parameters are the number of subgranules, percentage of data for the training phase, learning algorithm, goal error, number of hidden layers, and their number of neurons. Nowadays, there is a great variety of approaches and new techniques within the evolutionary computing area, and these approaches and techniques have emerged to help find optimal solutions to problems or models and bioinspired algorithms are part of this area. In this work a grey wolf optimizer is proposed for the design of modular granular neural networks, and the results are compared against a genetic algorithm and a firefly algorithm in order to know which of these techniques provides better results when applied to human recognition.

An incremental approach to genetic-algorithms-based classification.

PubMed

Guan, Sheng-Uei; Zhu, Fangming

2005-04-01

Incremental learning has been widely addressed in the machine learning literature to cope with learning tasks where the learning environment is ever changing or training samples become available over time. However, most research work explores incremental learning with statistical algorithms or neural networks, rather than evolutionary algorithms. The work in this paper employs genetic algorithms (GAs) as basic learning algorithms for incremental learning within one or more classifier agents in a multiagent environment. Four new approaches with different initialization schemes are proposed. They keep the old solutions and use an "integration" operation to integrate them with new elements to accommodate new attributes, while biased mutation and crossover operations are adopted to further evolve a reinforced solution. The simulation results on benchmark classification data sets show that the proposed approaches can deal with the arrival of new input attributes and integrate them with the original input space. It is also shown that the proposed approaches can be successfully used for incremental learning and improve classification rates as compared to the retraining GA. Possible applications for continuous incremental training and feature selection are also discussed.

A comparative study of breast cancer diagnosis based on neural network ensemble via improved training algorithms.

PubMed

Azami, Hamed; Escudero, Javier

2015-08-01

Breast cancer is one of the most common types of cancer in women all over the world. Early diagnosis of this kind of cancer can significantly increase the chances of long-term survival. Since diagnosis of breast cancer is a complex problem, neural network (NN) approaches have been used as a promising solution. Considering the low speed of the back-propagation (BP) algorithm to train a feed-forward NN, we consider a number of improved NN trainings for the Wisconsin breast cancer dataset: BP with momentum, BP with adaptive learning rate, BP with adaptive learning rate and momentum, Polak-Ribikre conjugate gradient algorithm (CGA), Fletcher-Reeves CGA, Powell-Beale CGA, scaled CGA, resilient BP (RBP), one-step secant and quasi-Newton methods. An NN ensemble, which is a learning paradigm to combine a number of NN outputs, is used to improve the accuracy of the classification task. Results demonstrate that NN ensemble-based classification methods have better performance than NN-based algorithms. The highest overall average accuracy is 97.68% obtained by NN ensemble trained by RBP for 50%-50% training-test evaluation method.

Super-resolution reconstruction of MR image with a novel residual learning network algorithm

NASA Astrophysics Data System (ADS)

Shi, Jun; Liu, Qingping; Wang, Chaofeng; Zhang, Qi; Ying, Shihui; Xu, Haoyu

2018-04-01

Spatial resolution is one of the key parameters of magnetic resonance imaging (MRI). The image super-resolution (SR) technique offers an alternative approach to improve the spatial resolution of MRI due to its simplicity. Convolutional neural networks (CNN)-based SR algorithms have achieved state-of-the-art performance, in which the global residual learning (GRL) strategy is now commonly used due to its effectiveness for learning image details for SR. However, the partial loss of image details usually happens in a very deep network due to the degradation problem. In this work, we propose a novel residual learning-based SR algorithm for MRI, which combines both multi-scale GRL and shallow network block-based local residual learning (LRL). The proposed LRL module works effectively in capturing high-frequency details by learning local residuals. One simulated MRI dataset and two real MRI datasets have been used to evaluate our algorithm. The experimental results show that the proposed SR algorithm achieves superior performance to all of the other compared CNN-based SR algorithms in this work.

Learning in Artificial Neural Systems

NASA Technical Reports Server (NTRS)

Matheus, Christopher J.; Hohensee, William E.

1987-01-01

This paper presents an overview and analysis of learning in Artificial Neural Systems (ANS's). It begins with a general introduction to neural networks and connectionist approaches to information processing. The basis for learning in ANS's is then described, and compared with classical Machine learning. While similar in some ways, ANS learning deviates from tradition in its dependence on the modification of individual weights to bring about changes in a knowledge representation distributed across connections in a network. This unique form of learning is analyzed from two aspects: the selection of an appropriate network architecture for representing the problem, and the choice of a suitable learning rule capable of reproducing the desired function within the given network. The various network architectures are classified, and then identified with explicit restrictions on the types of functions they are capable of representing. The learning rules, i.e., algorithms that specify how the network weights are modified, are similarly taxonomized, and where possible, the limitations inherent to specific classes of rules are outlined.

Growing a hypercubical output space in a self-organizing feature map.

PubMed

Bauer, H U; Villmann, T

1997-01-01

Neural maps project data from an input space onto a neuron position in a (often lower dimensional) output space grid in a neighborhood preserving way, with neighboring neurons in the output space responding to neighboring data points in the input space. A map-learning algorithm can achieve an optimal neighborhood preservation only, if the output space topology roughly matches the effective structure of the data in the input space. We here present a growth algorithm, called the GSOM or growing self-organizing map, which enhances a widespread map self-organization process, Kohonen's self-organizing feature map (SOFM), by an adaptation of the output space grid during learning. The GSOM restricts the output space structure to the shape of a general hypercubical shape, with the overall dimensionality of the grid and its extensions along the different directions being subject of the adaptation. This constraint meets the demands of many larger information processing systems, of which the neural map can be a part. We apply our GSOM-algorithm to three examples, two of which involve real world data. Using recently developed methods for measuring the degree of neighborhood preservation in neural maps, we find the GSOM-algorithm to produce maps which preserve neighborhoods in a nearly optimal fashion.

Proceedings of the Second Joint Technology Workshop on Neural Networks and Fuzzy Logic, volume 2

NASA Technical Reports Server (NTRS)

Lea, Robert N. (Editor); Villarreal, James A. (Editor)

1991-01-01

Documented here are papers presented at the Neural Networks and Fuzzy Logic Workshop sponsored by NASA and the University of Texas, Houston. Topics addressed included adaptive systems, learning algorithms, network architectures, vision, robotics, neurobiological connections, speech recognition and synthesis, fuzzy set theory and application, control and dynamics processing, space applications, fuzzy logic and neural network computers, approximate reasoning, and multiobject decision making.

Temporal neural networks and transient analysis of complex engineering systems

NASA Astrophysics Data System (ADS)

Uluyol, Onder

A theory is introduced for a multi-layered Local Output Gamma Feedback (LOGF) neural network within the paradigm of Locally-Recurrent Globally-Feedforward neural networks. It is developed for the identification, prediction, and control tasks of spatio-temporal systems and allows for the presentation of different time scales through incorporation of a gamma memory. It is initially applied to the tasks of sunspot and Mackey-Glass series prediction as benchmarks, then it is extended to the task of power level control of a nuclear reactor at different fuel cycle conditions. The developed LOGF neuron model can also be viewed as a Transformed Input and State (TIS) Gamma memory for neural network architectures for temporal processing. The novel LOGF neuron model extends the static neuron model by incorporating into it a short-term memory structure in the form of a digital gamma filter. A feedforward neural network made up of LOGF neurons can thus be used to model dynamic systems. A learning algorithm based upon the Backpropagation-Through-Time (BTT) approach is derived. It is applicable for training a general L-layer LOGF neural network. The spatial and temporal weights and parameters of the network are iteratively optimized for a given problem using the derived learning algorithm.

SAGRAD: A Program for Neural Network Training with Simulated Annealing and the Conjugate Gradient Method

PubMed Central

Bernal, Javier; Torres-Jimenez, Jose

2015-01-01

SAGRAD (Simulated Annealing GRADient), a Fortran 77 program for computing neural networks for classification using batch learning, is discussed. Neural network training in SAGRAD is based on a combination of simulated annealing and Møller’s scaled conjugate gradient algorithm, the latter a variation of the traditional conjugate gradient method, better suited for the nonquadratic nature of neural networks. Different aspects of the implementation of the training process in SAGRAD are discussed, such as the efficient computation of gradients and multiplication of vectors by Hessian matrices that are required by Møller’s algorithm; the (re)initialization of weights with simulated annealing required to (re)start Møller’s algorithm the first time and each time thereafter that it shows insufficient progress in reaching a possibly local minimum; and the use of simulated annealing when Møller’s algorithm, after possibly making considerable progress, becomes stuck at a local minimum or flat area of weight space. Outlines of the scaled conjugate gradient algorithm, the simulated annealing procedure and the training process used in SAGRAD are presented together with results from running SAGRAD on two examples of training data. PMID:26958442

Application of neural nets in structural optimization

NASA Technical Reports Server (NTRS)

Berke, Laszlo; Hajela, Prabhat

1993-01-01

The biological motivation for Artificial Neural Net developments is briefly discussed, and the most popular paradigm, the feedforward supervised learning net with error back propagation training algorithm, is introduced. Possible approaches for utilization in structural optimization is illustrated through simple examples. Other currently ongoing developments for application in structural mechanics are also mentioned.

Visual NNet: An Educational ANN's Simulation Environment Reusing Matlab Neural Networks Toolbox

ERIC Educational Resources Information Center

Garcia-Roselló, Emilio; González-Dacosta, Jacinto; Lado, Maria J.; Méndez, Arturo J.; Garcia Pérez-Schofield, Baltasar; Ferrer, Fátima

2011-01-01

Artificial Neural Networks (ANN's) are nowadays a common subject in different curricula of graduate and postgraduate studies. Due to the complex algorithms involved and the dynamic nature of ANN's, simulation software has been commonly used to teach this subject. This software has usually been developed specifically for learning purposes, because…

Boosted ARTMAP: modifications to fuzzy ARTMAP motivated by boosting theory.

PubMed

Verzi, Stephen J; Heileman, Gregory L; Georgiopoulos, Michael

2006-05-01

In this paper, several modifications to the Fuzzy ARTMAP neural network architecture are proposed for conducting classification in complex, possibly noisy, environments. The goal of these modifications is to improve upon the generalization performance of Fuzzy ART-based neural networks, such as Fuzzy ARTMAP, in these situations. One of the major difficulties of employing Fuzzy ARTMAP on such learning problems involves over-fitting of the training data. Structural risk minimization is a machine-learning framework that addresses the issue of over-fitting by providing a backbone for analysis as well as an impetus for the design of better learning algorithms. The theory of structural risk minimization reveals a trade-off between training error and classifier complexity in reducing generalization error, which will be exploited in the learning algorithms proposed in this paper. Boosted ART extends Fuzzy ART by allowing the spatial extent of each cluster formed to be adjusted independently. Boosted ARTMAP generalizes upon Fuzzy ARTMAP by allowing non-zero training error in an effort to reduce the hypothesis complexity and hence improve overall generalization performance. Although Boosted ARTMAP is strictly speaking not a boosting algorithm, the changes it encompasses were motivated by the goals that one strives to achieve when employing boosting. Boosted ARTMAP is an on-line learner, it does not require excessive parameter tuning to operate, and it reduces precisely to Fuzzy ARTMAP for particular parameter values. Another architecture described in this paper is Structural Boosted ARTMAP, which uses both Boosted ART and Boosted ARTMAP to perform structural risk minimization learning. Structural Boosted ARTMAP will allow comparison of the capabilities of off-line versus on-line learning as well as empirical risk minimization versus structural risk minimization using Fuzzy ARTMAP-based neural network architectures. Both empirical and theoretical results are presented to enhance the understanding of these architectures.

Investigating the performance of neural network backpropagation algorithms for TEC estimations using South African GPS data

NASA Astrophysics Data System (ADS)

Habarulema, J. B.; McKinnell, L.-A.

2012-05-01

In this work, results obtained by investigating the application of different neural network backpropagation training algorithms are presented. This was done to assess the performance accuracy of each training algorithm in total electron content (TEC) estimations using identical datasets in models development and verification processes. Investigated training algorithms are standard backpropagation (SBP), backpropagation with weight delay (BPWD), backpropagation with momentum (BPM) term, backpropagation with chunkwise weight update (BPC) and backpropagation for batch (BPB) training. These five algorithms are inbuilt functions within the Stuttgart Neural Network Simulator (SNNS) and the main objective was to find out the training algorithm that generates the minimum error between the TEC derived from Global Positioning System (GPS) observations and the modelled TEC data. Another investigated algorithm is the MatLab based Levenberg-Marquardt backpropagation (L-MBP), which achieves convergence after the least number of iterations during training. In this paper, neural network (NN) models were developed using hourly TEC data (for 8 years: 2000-2007) derived from GPS observations over a receiver station located at Sutherland (SUTH) (32.38° S, 20.81° E), South Africa. Verification of the NN models for all algorithms considered was performed on both "seen" and "unseen" data. Hourly TEC values over SUTH for 2003 formed the "seen" dataset. The "unseen" dataset consisted of hourly TEC data for 2002 and 2008 over Cape Town (CPTN) (33.95° S, 18.47° E) and SUTH, respectively. The models' verification showed that all algorithms investigated provide comparable results statistically, but differ significantly in terms of time required to achieve convergence during input-output data training/learning. This paper therefore provides a guide to neural network users for choosing appropriate algorithms based on the availability of computation capabilities used for research.

An application of artificial neural networks to experimental data approximation

NASA Technical Reports Server (NTRS)

Meade, Andrew J., Jr.

1993-01-01

As an initial step in the evaluation of networks, a feedforward architecture is trained to approximate experimental data by the backpropagation algorithm. Several drawbacks were detected and an alternative learning algorithm was then developed to partially address the drawbacks. This noniterative algorithm has a number of advantages over the backpropagation method and is easily implemented on existing hardware.

Learning a trajectory using adjoint functions and teacher forcing

NASA Technical Reports Server (NTRS)

Toomarian, Nikzad B.; Barhen, Jacob

1992-01-01

A new methodology for faster supervised temporal learning in nonlinear neural networks is presented which builds upon the concept of adjoint operators to allow fast computation of the gradients of an error functional with respect to all parameters of the neural architecture, and exploits the concept of teacher forcing to incorporate information on the desired output into the activation dynamics. The importance of the initial or final time conditions for the adjoint equations is discussed. A new algorithm is presented in which the adjoint equations are solved simultaneously (i.e., forward in time) with the activation dynamics of the neural network. We also indicate how teacher forcing can be modulated in time as learning proceeds. The results obtained show that the learning time is reduced by one to two orders of magnitude with respect to previously published results, while trajectory tracking is significantly improved. The proposed methodology makes hardware implementation of temporal learning attractive for real-time applications.

Comparison of machine-learning algorithms to build a predictive model for detecting undiagnosed diabetes - ELSA-Brasil: accuracy study.

PubMed

Olivera, André Rodrigues; Roesler, Valter; Iochpe, Cirano; Schmidt, Maria Inês; Vigo, Álvaro; Barreto, Sandhi Maria; Duncan, Bruce Bartholow

2017-01-01

Type 2 diabetes is a chronic disease associated with a wide range of serious health complications that have a major impact on overall health. The aims here were to develop and validate predictive models for detecting undiagnosed diabetes using data from the Longitudinal Study of Adult Health (ELSA-Brasil) and to compare the performance of different machine-learning algorithms in this task. Comparison of machine-learning algorithms to develop predictive models using data from ELSA-Brasil. After selecting a subset of 27 candidate variables from the literature, models were built and validated in four sequential steps: (i) parameter tuning with tenfold cross-validation, repeated three times; (ii) automatic variable selection using forward selection, a wrapper strategy with four different machine-learning algorithms and tenfold cross-validation (repeated three times), to evaluate each subset of variables; (iii) error estimation of model parameters with tenfold cross-validation, repeated ten times; and (iv) generalization testing on an independent dataset. The models were created with the following machine-learning algorithms: logistic regression, artificial neural network, naïve Bayes, K-nearest neighbor and random forest. The best models were created using artificial neural networks and logistic regression. -These achieved mean areas under the curve of, respectively, 75.24% and 74.98% in the error estimation step and 74.17% and 74.41% in the generalization testing step. Most of the predictive models produced similar results, and demonstrated the feasibility of identifying individuals with highest probability of having undiagnosed diabetes, through easily-obtained clinical data.

An efficient self-organizing map designed by genetic algorithms for the traveling salesman problem.

PubMed

Jin, Hui-Dong; Leung, Kwong-Sak; Wong, Man-Leung; Xu, Z B

2003-01-01

As a typical combinatorial optimization problem, the traveling salesman problem (TSP) has attracted extensive research interest. In this paper, we develop a self-organizing map (SOM) with a novel learning rule. It is called the integrated SOM (ISOM) since its learning rule integrates the three learning mechanisms in the SOM literature. Within a single learning step, the excited neuron is first dragged toward the input city, then pushed to the convex hull of the TSP, and finally drawn toward the middle point of its two neighboring neurons. A genetic algorithm is successfully specified to determine the elaborate coordination among the three learning mechanisms as well as the suitable parameter setting. The evolved ISOM (eISOM) is examined on three sets of TSP to demonstrate its power and efficiency. The computation complexity of the eISOM is quadratic, which is comparable to other SOM-like neural networks. Moreover, the eISOM can generate more accurate solutions than several typical approaches for TSP including the SOM developed by Budinich, the expanding SOM, the convex elastic net, and the FLEXMAP algorithm. Though its solution accuracy is not yet comparable to some sophisticated heuristics, the eISOM is one of the most accurate neural networks for the TSP.

The HEP.TrkX Project: deep neural networks for HL-LHC online and offline tracking

DOE PAGES

Farrell, Steven; Anderson, Dustin; Calafiura, Paolo; ...

2017-08-08

Particle track reconstruction in dense environments such as the detectors of the High Luminosity Large Hadron Collider (HL-LHC) is a challenging pattern recognition problem. Traditional tracking algorithms such as the combinatorial Kalman Filter have been used with great success in LHC experiments for years. However, these state-of-the-art techniques are inherently sequential and scale poorly with the expected increases in detector occupancy in the HL-LHC conditions. The HEP.TrkX project is a pilot project with the aim to identify and develop cross-experiment solutions based on machine learning algorithms for track reconstruction. Machine learning algorithms bring a lot of potential to this problemmore » thanks to their capability to model complex non-linear data dependencies, to learn effective representations of high-dimensional data through training, and to parallelize easily on high-throughput architectures such as GPUs. This contribution will describe our initial explorations into this relatively unexplored idea space. Furthermore, we will discuss the use of recurrent (LSTM) and convolutional neural networks to find and fit tracks in toy detector data.« less

The HEP.TrkX Project: deep neural networks for HL-LHC online and offline tracking

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

Farrell, Steven; Anderson, Dustin; Calafiura, Paolo

Particle track reconstruction in dense environments such as the detectors of the High Luminosity Large Hadron Collider (HL-LHC) is a challenging pattern recognition problem. Traditional tracking algorithms such as the combinatorial Kalman Filter have been used with great success in LHC experiments for years. However, these state-of-the-art techniques are inherently sequential and scale poorly with the expected increases in detector occupancy in the HL-LHC conditions. The HEP.TrkX project is a pilot project with the aim to identify and develop cross-experiment solutions based on machine learning algorithms for track reconstruction. Machine learning algorithms bring a lot of potential to this problemmore » thanks to their capability to model complex non-linear data dependencies, to learn effective representations of high-dimensional data through training, and to parallelize easily on high-throughput architectures such as GPUs. This contribution will describe our initial explorations into this relatively unexplored idea space. Furthermore, we will discuss the use of recurrent (LSTM) and convolutional neural networks to find and fit tracks in toy detector data.« less

The HEP.TrkX Project: deep neural networks for HL-LHC online and offline tracking

NASA Astrophysics Data System (ADS)

Farrell, Steven; Anderson, Dustin; Calafiura, Paolo; Cerati, Giuseppe; Gray, Lindsey; Kowalkowski, Jim; Mudigonda, Mayur; Prabhat; Spentzouris, Panagiotis; Spiropoulou, Maria; Tsaris, Aristeidis; Vlimant, Jean-Roch; Zheng, Stephan

2017-08-01

Particle track reconstruction in dense environments such as the detectors of the High Luminosity Large Hadron Collider (HL-LHC) is a challenging pattern recognition problem. Traditional tracking algorithms such as the combinatorial Kalman Filter have been used with great success in LHC experiments for years. However, these state-of-the-art techniques are inherently sequential and scale poorly with the expected increases in detector occupancy in the HL-LHC conditions. The HEP.TrkX project is a pilot project with the aim to identify and develop cross-experiment solutions based on machine learning algorithms for track reconstruction. Machine learning algorithms bring a lot of potential to this problem thanks to their capability to model complex non-linear data dependencies, to learn effective representations of high-dimensional data through training, and to parallelize easily on high-throughput architectures such as GPUs. This contribution will describe our initial explorations into this relatively unexplored idea space. We will discuss the use of recurrent (LSTM) and convolutional neural networks to find and fit tracks in toy detector data.

Oscillatory neural network for pattern recognition: trajectory based classification and supervised learning.

PubMed

Miller, Vonda H; Jansen, Ben H

2008-12-01

Computer algorithms that match human performance in recognizing written text or spoken conversation remain elusive. The reasons why the human brain far exceeds any existing recognition scheme to date in the ability to generalize and to extract invariant characteristics relevant to category matching are not clear. However, it has been postulated that the dynamic distribution of brain activity (spatiotemporal activation patterns) is the mechanism by which stimuli are encoded and matched to categories. This research focuses on supervised learning using a trajectory based distance metric for category discrimination in an oscillatory neural network model. Classification is accomplished using a trajectory based distance metric. Since the distance metric is differentiable, a supervised learning algorithm based on gradient descent is demonstrated. Classification of spatiotemporal frequency transitions and their relation to a priori assessed categories is shown along with the improved classification results after supervised training. The results indicate that this spatiotemporal representation of stimuli and the associated distance metric is useful for simple pattern recognition tasks and that supervised learning improves classification results.

A new simple /spl infin/OH neuron model as a biologically plausible principal component analyzer.

PubMed

Jankovic, M V

2003-01-01

A new approach to unsupervised learning in a single-layer neural network is discussed. An algorithm for unsupervised learning based upon the Hebbian learning rule is presented. A simple neuron model is analyzed. A dynamic neural model, which contains both feed-forward and feedback connections between the input and the output, has been adopted. The, proposed learning algorithm could be more correctly named self-supervised rather than unsupervised. The solution proposed here is a modified Hebbian rule, in which the modification of the synaptic strength is proportional not to pre- and postsynaptic activity, but instead to the presynaptic and averaged value of postsynaptic activity. It is shown that the model neuron tends to extract the principal component from a stationary input vector sequence. Usually accepted additional decaying terms for the stabilization of the original Hebbian rule are avoided. Implementation of the basic Hebbian scheme would not lead to unrealistic growth of the synaptic strengths, thanks to the adopted network structure.

Recurrent Neural Networks With Auxiliary Memory Units.

PubMed

Wang, Jianyong; Zhang, Lei; Guo, Quan; Yi, Zhang

2018-05-01

Memory is one of the most important mechanisms in recurrent neural networks (RNNs) learning. It plays a crucial role in practical applications, such as sequence learning. With a good memory mechanism, long term history can be fused with current information, and can thus improve RNNs learning. Developing a suitable memory mechanism is always desirable in the field of RNNs. This paper proposes a novel memory mechanism for RNNs. The main contributions of this paper are: 1) an auxiliary memory unit (AMU) is proposed, which results in a new special RNN model (AMU-RNN), separating the memory and output explicitly and 2) an efficient learning algorithm is developed by employing the technique of error flow truncation. The proposed AMU-RNN model, together with the developed learning algorithm, can learn and maintain stable memory over a long time range. This method overcomes both the learning conflict problem and gradient vanishing problem. Unlike the traditional method, which mixes the memory and output with a single neuron in a recurrent unit, the AMU provides an auxiliary memory neuron to maintain memory in particular. By separating the memory and output in a recurrent unit, the problem of learning conflicts can be eliminated easily. Moreover, by using the technique of error flow truncation, each auxiliary memory neuron ensures constant error flow during the learning process. The experiments demonstrate good performance of the proposed AMU-RNNs and the developed learning algorithm. The method exhibits quite efficient learning performance with stable convergence in the AMU-RNN learning and outperforms the state-of-the-art RNN models in sequence generation and sequence classification tasks.

Conduction Delay Learning Model for Unsupervised and Supervised Classification of Spatio-Temporal Spike Patterns

PubMed Central

Matsubara, Takashi

2017-01-01

Precise spike timing is considered to play a fundamental role in communications and signal processing in biological neural networks. Understanding the mechanism of spike timing adjustment would deepen our understanding of biological systems and enable advanced engineering applications such as efficient computational architectures. However, the biological mechanisms that adjust and maintain spike timing remain unclear. Existing algorithms adopt a supervised approach, which adjusts the axonal conduction delay and synaptic efficacy until the spike timings approximate the desired timings. This study proposes a spike timing-dependent learning model that adjusts the axonal conduction delay and synaptic efficacy in both unsupervised and supervised manners. The proposed learning algorithm approximates the Expectation-Maximization algorithm, and classifies the input data encoded into spatio-temporal spike patterns. Even in the supervised classification, the algorithm requires no external spikes indicating the desired spike timings unlike existing algorithms. Furthermore, because the algorithm is consistent with biological models and hypotheses found in existing biological studies, it could capture the mechanism underlying biological delay learning. PMID:29209191

Conduction Delay Learning Model for Unsupervised and Supervised Classification of Spatio-Temporal Spike Patterns.

PubMed

Matsubara, Takashi

2017-01-01

Precise spike timing is considered to play a fundamental role in communications and signal processing in biological neural networks. Understanding the mechanism of spike timing adjustment would deepen our understanding of biological systems and enable advanced engineering applications such as efficient computational architectures. However, the biological mechanisms that adjust and maintain spike timing remain unclear. Existing algorithms adopt a supervised approach, which adjusts the axonal conduction delay and synaptic efficacy until the spike timings approximate the desired timings. This study proposes a spike timing-dependent learning model that adjusts the axonal conduction delay and synaptic efficacy in both unsupervised and supervised manners. The proposed learning algorithm approximates the Expectation-Maximization algorithm, and classifies the input data encoded into spatio-temporal spike patterns. Even in the supervised classification, the algorithm requires no external spikes indicating the desired spike timings unlike existing algorithms. Furthermore, because the algorithm is consistent with biological models and hypotheses found in existing biological studies, it could capture the mechanism underlying biological delay learning.

Neural networks for continuous online learning and control.

PubMed

Choy, Min Chee; Srinivasan, Dipti; Cheu, Ruey Long

2006-11-01

This paper proposes a new hybrid neural network (NN) model that employs a multistage online learning process to solve the distributed control problem with an infinite horizon. Various techniques such as reinforcement learning and evolutionary algorithm are used to design the multistage online learning process. For this paper, the infinite horizon distributed control problem is implemented in the form of real-time distributed traffic signal control for intersections in a large-scale traffic network. The hybrid neural network model is used to design each of the local traffic signal controllers at the respective intersections. As the state of the traffic network changes due to random fluctuation of traffic volumes, the NN-based local controllers will need to adapt to the changing dynamics in order to provide effective traffic signal control and to prevent the traffic network from becoming overcongested. Such a problem is especially challenging if the local controllers are used for an infinite horizon problem where online learning has to take place continuously once the controllers are implemented into the traffic network. A comprehensive simulation model of a section of the Central Business District (CBD) of Singapore has been developed using PARAMICS microscopic simulation program. As the complexity of the simulation increases, results show that the hybrid NN model provides significant improvement in traffic conditions when evaluated against an existing traffic signal control algorithm as well as a new, continuously updated simultaneous perturbation stochastic approximation-based neural network (SPSA-NN). Using the hybrid NN model, the total mean delay of each vehicle has been reduced by 78% and the total mean stoppage time of each vehicle has been reduced by 84% compared to the existing traffic signal control algorithm. This shows the efficacy of the hybrid NN model in solving large-scale traffic signal control problem in a distributed manner. Also, it indicates the possibility of using the hybrid NN model for other applications that are similar in nature as the infinite horizon distributed control problem.

A regularization approach to continuous learning with an application to financial derivatives pricing.

PubMed

Ormoneit, D

1999-12-01

We consider the training of neural networks in cases where the nonlinear relationship of interest gradually changes over time. One possibility to deal with this problem is by regularization where a variation penalty is added to the usual mean squared error criterion. To learn the regularized network weights we suggest the Iterative Extended Kalman Filter (IEKF) as a learning rule, which may be derived from a Bayesian perspective on the regularization problem. A primary application of our algorithm is in financial derivatives pricing, where neural networks may be used to model the dependency of the derivatives' price on one or several underlying assets. After giving a brief introduction to the problem of derivatives pricing we present experiments with German stock index options data showing that a regularized neural network trained with the IEKF outperforms several benchmark models and alternative learning procedures. In particular, the performance may be greatly improved using a newly designed neural network architecture that accounts for no-arbitrage pricing restrictions.

From modulated Hebbian plasticity to simple behavior learning through noise and weight saturation.

PubMed

Soltoggio, Andrea; Stanley, Kenneth O

2012-10-01

Synaptic plasticity is a major mechanism for adaptation, learning, and memory. Yet current models struggle to link local synaptic changes to the acquisition of behaviors. The aim of this paper is to demonstrate a computational relationship between local Hebbian plasticity and behavior learning by exploiting two traditionally unwanted features: neural noise and synaptic weight saturation. A modulation signal is employed to arbitrate the sign of plasticity: when the modulation is positive, the synaptic weights saturate to express exploitative behavior; when it is negative, the weights converge to average values, and neural noise reconfigures the network's functionality. This process is demonstrated through simulating neural dynamics in the autonomous emergence of fearful and aggressive navigating behaviors and in the solution to reward-based problems. The neural model learns, memorizes, and modifies different behaviors that lead to positive modulation in a variety of settings. The algorithm establishes a simple relationship between local plasticity and behavior learning by demonstrating the utility of noise and weight saturation. Moreover, it provides a new tool to simulate adaptive behavior, and contributes to bridging the gap between synaptic changes and behavior in neural computation. Copyright © 2012 Elsevier Ltd. All rights reserved.

A Spatiotemporal Prediction Framework for Air Pollution Based on Deep RNN

NASA Astrophysics Data System (ADS)

Fan, J.; Li, Q.; Hou, J.; Feng, X.; Karimian, H.; Lin, S.

2017-10-01

Time series data in practical applications always contain missing values due to sensor malfunction, network failure, outliers etc. In order to handle missing values in time series, as well as the lack of considering temporal properties in machine learning models, we propose a spatiotemporal prediction framework based on missing value processing algorithms and deep recurrent neural network (DRNN). By using missing tag and missing interval to represent time series patterns, we implement three different missing value fixing algorithms, which are further incorporated into deep neural network that consists of LSTM (Long Short-term Memory) layers and fully connected layers. Real-world air quality and meteorological datasets (Jingjinji area, China) are used for model training and testing. Deep feed forward neural networks (DFNN) and gradient boosting decision trees (GBDT) are trained as baseline models against the proposed DRNN. Performances of three missing value fixing algorithms, as well as different machine learning models are evaluated and analysed. Experiments show that the proposed DRNN framework outperforms both DFNN and GBDT, therefore validating the capacity of the proposed framework. Our results also provides useful insights for better understanding of different strategies that handle missing values.

Cascaded VLSI neural network architecture for on-line learning

NASA Technical Reports Server (NTRS)

Thakoor, Anilkumar P. (Inventor); Duong, Tuan A. (Inventor); Daud, Taher (Inventor)

1992-01-01

High-speed, analog, fully-parallel, and asynchronous building blocks are cascaded for larger sizes and enhanced resolution. A hardware compatible algorithm permits hardware-in-the-loop learning despite limited weight resolution. A computation intensive feature classification application was demonstrated with this flexible hardware and new algorithm at high speed. This result indicates that these building block chips can be embedded as an application specific coprocessor for solving real world problems at extremely high data rates.

Cascaded VLSI neural network architecture for on-line learning

NASA Technical Reports Server (NTRS)

Duong, Tuan A. (Inventor); Daud, Taher (Inventor); Thakoor, Anilkumar P. (Inventor)

1995-01-01

High-speed, analog, fully-parallel and asynchronous building blocks are cascaded for larger sizes and enhanced resolution. A hardware-compatible algorithm permits hardware-in-the-loop learning despite limited weight resolution. A comparison-intensive feature classification application has been demonstrated with this flexible hardware and new algorithm at high speed. This result indicates that these building block chips can be embedded as application-specific-coprocessors for solving real-world problems at extremely high data rates.

Application of machine learning methods for traffic signs recognition

NASA Astrophysics Data System (ADS)

Filatov, D. V.; Ignatev, K. V.; Deviatkin, A. V.; Serykh, E. V.

2018-02-01

This paper focuses on solving a relevant and pressing safety issue on intercity roads. Two approaches were considered for solving the problem of traffic signs recognition; the approaches involved neural networks to analyze images obtained from a camera in the real-time mode. The first approach is based on a sequential image processing. At the initial stage, with the help of color filters and morphological operations (dilatation and erosion), the area containing the traffic sign is located on the image, then the selected and scaled fragment of the image is analyzed using a feedforward neural network to determine the meaning of the found traffic sign. Learning of the neural network in this approach is carried out using a backpropagation method. The second approach involves convolution neural networks at both stages, i.e. when searching and selecting the area of the image containing the traffic sign, and when determining its meaning. Learning of the neural network in the second approach is carried out using the intersection over union function and a loss function. For neural networks to learn and the proposed algorithms to be tested, a series of videos from a dash cam were used that were shot under various weather and illumination conditions. As a result, the proposed approaches for traffic signs recognition were analyzed and compared by key indicators such as recognition rate percentage and the complexity of neural networks’ learning process.

On the Relationships between Generative Encodings, Regularity, and Learning Abilities when Evolving Plastic Artificial Neural Networks

PubMed Central

Tonelli, Paul; Mouret, Jean-Baptiste

2013-01-01

A major goal of bio-inspired artificial intelligence is to design artificial neural networks with abilities that resemble those of animal nervous systems. It is commonly believed that two keys for evolving nature-like artificial neural networks are (1) the developmental process that links genes to nervous systems, which enables the evolution of large, regular neural networks, and (2) synaptic plasticity, which allows neural networks to change during their lifetime. So far, these two topics have been mainly studied separately. The present paper shows that they are actually deeply connected. Using a simple operant conditioning task and a classic evolutionary algorithm, we compare three ways to encode plastic neural networks: a direct encoding, a developmental encoding inspired by computational neuroscience models, and a developmental encoding inspired by morphogen gradients (similar to HyperNEAT). Our results suggest that using a developmental encoding could improve the learning abilities of evolved, plastic neural networks. Complementary experiments reveal that this result is likely the consequence of the bias of developmental encodings towards regular structures: (1) in our experimental setup, encodings that tend to produce more regular networks yield networks with better general learning abilities; (2) whatever the encoding is, networks that are the more regular are statistically those that have the best learning abilities. PMID:24236099

A Neural Network Model of Retrieval-Induced Forgetting

ERIC Educational Resources Information Center

Norman, Kenneth A.; Newman, Ehren L.; Detre, Greg

2007-01-01

Retrieval-induced forgetting (RIF) refers to the finding that retrieving a memory can impair subsequent recall of related memories. Here, the authors present a new model of how the brain gives rise to RIF in both semantic and episodic memory. The core of the model is a recently developed neural network learning algorithm that leverages regular…

Neural Network Classifier Architectures for Phoneme Recognition. CRC Technical Note No. CRC-TN-92-001.

ERIC Educational Resources Information Center

Treurniet, William

A study applied artificial neural networks, trained with the back-propagation learning algorithm, to modelling phonemes extracted from the DARPA TIMIT multi-speaker, continuous speech data base. A number of proposed network architectures were applied to the phoneme classification task, ranging from the simple feedforward multilayer network to more…

On the use of harmony search algorithm in the training of wavelet neural networks

NASA Astrophysics Data System (ADS)

Lai, Kee Huong; Zainuddin, Zarita; Ong, Pauline

2015-10-01

Wavelet neural networks (WNNs) are a class of feedforward neural networks that have been used in a wide range of industrial and engineering applications to model the complex relationships between the given inputs and outputs. The training of WNNs involves the configuration of the weight values between neurons. The backpropagation training algorithm, which is a gradient-descent method, can be used for this training purpose. Nonetheless, the solutions found by this algorithm often get trapped at local minima. In this paper, a harmony search-based algorithm is proposed for the training of WNNs. The training of WNNs, thus can be formulated as a continuous optimization problem, where the objective is to maximize the overall classification accuracy. Each candidate solution proposed by the harmony search algorithm represents a specific WNN architecture. In order to speed up the training process, the solution space is divided into disjoint partitions during the random initialization step of harmony search algorithm. The proposed training algorithm is tested onthree benchmark problems from the UCI machine learning repository, as well as one real life application, namely, the classification of electroencephalography signals in the task of epileptic seizure detection. The results obtained show that the proposed algorithm outperforms the traditional harmony search algorithm in terms of overall classification accuracy.

Two algorithms for neural-network design and training with application to channel equalization.

PubMed

Sweatman, C Z; Mulgrew, B; Gibson, G J

1998-01-01

We describe two algorithms for designing and training neural-network classifiers. The first, the linear programming slab algorithm (LPSA), is motivated by the problem of reconstructing digital signals corrupted by passage through a dispersive channel and by additive noise. It constructs a multilayer perceptron (MLP) to separate two disjoint sets by using linear programming methods to identify network parameters. The second, the perceptron learning slab algorithm (PLSA), avoids the computational costs of linear programming by using an error-correction approach to identify parameters. Both algorithms operate in highly constrained parameter spaces and are able to exploit symmetry in the classification problem. Using these algorithms, we develop a number of procedures for the adaptive equalization of a complex linear 4-quadrature amplitude modulation (QAM) channel, and compare their performance in a simulation study. Results are given for both stationary and time-varying channels, the latter based on the COST 207 GSM propagation model.

Equilibrium point control of a monkey arm simulator by a fast learning tree structured artificial neural network.

PubMed

Dornay, M; Sanger, T D

1993-01-01

A planar 17 muscle model of the monkey's arm based on realistic biomechanical measurements was simulated on a Symbolics Lisp Machine. The simulator implements the equilibrium point hypothesis for the control of arm movements. Given initial and final desired positions, it generates a minimum-jerk desired trajectory of the hand and uses the backdriving algorithm to determine an appropriate sequence of motor commands to the muscles (Flash 1987; Mussa-Ivaldi et al. 1991; Dornay 1991b). These motor commands specify a temporal sequence of stable (attractive) equilibrium positions which lead to the desired hand movement. A strong disadvantage of the simulator is that it has no memory of previous computations. Determining the desired trajectory using the minimum-jerk model is instantaneous, but the laborious backdriving algorithm is slow, and can take up to one hour for some trajectories. The complexity of the required computations makes it a poor model for biological motor control. We propose a computationally simpler and more biologically plausible method for control which achieves the benefits of the backdriving algorithm. A fast learning, tree-structured network (Sanger 1991c) was trained to remember the knowledge obtained by the backdriving algorithm. The neural network learned the nonlinear mapping from a 2-dimensional cartesian planar hand position (x,y) to a 17-dimensional motor command space (u1, . . ., u17). Learning 20 training trajectories, each composed of 26 sample points [[x,y], [u1, . . ., u17] took only 20 min on a Sun-4 Sparc workstation. After the learning stage, new, untrained test trajectories as well as the original trajectories of the hand were given to the neural network as input. The network calculated the required motor commands for these movements. The resulting movements were close to the desired ones for both the training and test cases.

Melanoma segmentation based on deep learning.

PubMed

Zhang, Xiaoqing

2017-12-01

Malignant melanoma is one of the most deadly forms of skin cancer, which is one of the world's fastest-growing cancers. Early diagnosis and treatment is critical. In this study, a neural network structure is utilized to construct a broad and accurate basis for the diagnosis of skin cancer, thereby reducing screening errors. The technique is able to improve the efficacy for identification of normally indistinguishable lesions (such as pigment spots) versus clinically unknown lesions, and to ultimately improve the diagnostic accuracy. In the field of medical imaging, in general, using neural networks for image segmentation is relatively rare. The existing traditional machine-learning neural network algorithms still cannot completely solve the problem of information loss, nor detect the precise division of the boundary area. We use an improved neural network framework, described herein, to achieve efficacious feature learning, and satisfactory segmentation of melanoma images. The architecture of the network includes multiple convolution layers, dropout layers, softmax layers, multiple filters, and activation functions. The number of data sets can be increased via rotation of the training set. A non-linear activation function (such as ReLU and ELU) is employed to alleviate the problem of gradient disappearance, and RMSprop/Adam are incorporated to optimize the loss algorithm. A batch normalization layer is added between the convolution layer and the activation layer to solve the problem of gradient disappearance and explosion. Experiments, described herein, show that our improved neural network architecture achieves higher accuracy for segmentation of melanoma images as compared with existing processes.

Construction cost estimation of spherical storage tanks: artificial neural networks and hybrid regression—GA algorithms

NASA Astrophysics Data System (ADS)

Arabzadeh, Vida; Niaki, S. T. A.; Arabzadeh, Vahid

2017-10-01

One of the most important processes in the early stages of construction projects is to estimate the cost involved. This process involves a wide range of uncertainties, which make it a challenging task. Because of unknown issues, using the experience of the experts or looking for similar cases are the conventional methods to deal with cost estimation. The current study presents data-driven methods for cost estimation based on the application of artificial neural network (ANN) and regression models. The learning algorithms of the ANN are the Levenberg-Marquardt and the Bayesian regulated. Moreover, regression models are hybridized with a genetic algorithm to obtain better estimates of the coefficients. The methods are applied in a real case, where the input parameters of the models are assigned based on the key issues involved in a spherical tank construction. The results reveal that while a high correlation between the estimated cost and the real cost exists; both ANNs could perform better than the hybridized regression models. In addition, the ANN with the Levenberg-Marquardt learning algorithm (LMNN) obtains a better estimation than the ANN with the Bayesian-regulated learning algorithm (BRNN). The correlation between real data and estimated values is over 90%, while the mean square error is achieved around 0.4. The proposed LMNN model can be effective to reduce uncertainty and complexity in the early stages of the construction project.

Learning a single-hidden layer feedforward neural network using a rank correlation-based strategy with application to high dimensional gene expression and proteomic spectra datasets in cancer detection.

PubMed

Belciug, Smaranda; Gorunescu, Florin

2018-06-08

Methods based on microarrays (MA), mass spectrometry (MS), and machine learning (ML) algorithms have evolved rapidly in recent years, allowing for early detection of several types of cancer. A pitfall of these approaches, however, is the overfitting of data due to large number of attributes and small number of instances -- a phenomenon known as the 'curse of dimensionality'. A potentially fruitful idea to avoid this drawback is to develop algorithms that combine fast computation with a filtering module for the attributes. The goal of this paper is to propose a statistical strategy to initiate the hidden nodes of a single-hidden layer feedforward neural network (SLFN) by using both the knowledge embedded in data and a filtering mechanism for attribute relevance. In order to attest its feasibility, the proposed model has been tested on five publicly available high-dimensional datasets: breast, lung, colon, and ovarian cancer regarding gene expression and proteomic spectra provided by cDNA arrays, DNA microarray, and MS. The novel algorithm, called adaptive SLFN (aSLFN), has been compared with four major classification algorithms: traditional ELM, radial basis function network (RBF), single-hidden layer feedforward neural network trained by backpropagation algorithm (BP-SLFN), and support vector-machine (SVM). Experimental results showed that the classification performance of aSLFN is competitive with the comparison models. Copyright © 2018. Published by Elsevier Inc.

DREAMING OF ATMOSPHERES

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

Waldmann, I. P., E-mail: ingo@star.ucl.ac.uk

Here, we introduce the RobERt (Robotic Exoplanet Recognition) algorithm for the classification of exoplanetary emission spectra. Spectral retrieval of exoplanetary atmospheres frequently requires the preselection of molecular/atomic opacities to be defined by the user. In the era of open-source, automated, and self-sufficient retrieval algorithms, manual input should be avoided. User dependent input could, in worst-case scenarios, lead to incomplete models and biases in the retrieval. The RobERt algorithm is based on deep-belief neural (DBN) networks trained to accurately recognize molecular signatures for a wide range of planets, atmospheric thermal profiles, and compositions. Reconstructions of the learned features, also referred to as themore » “dreams” of the network, indicate good convergence and an accurate representation of molecular features in the DBN. Using these deep neural networks, we work toward retrieval algorithms that themselves understand the nature of the observed spectra, are able to learn from current and past data, and make sensible qualitative preselections of atmospheric opacities to be used for the quantitative stage of the retrieval process.« less

Designing Artificial Neural Networks Using Particle Swarm Optimization Algorithms

PubMed Central

Vázquez, Roberto A.

2015-01-01

Artificial Neural Network (ANN) design is a complex task because its performance depends on the architecture, the selected transfer function, and the learning algorithm used to train the set of synaptic weights. In this paper we present a methodology that automatically designs an ANN using particle swarm optimization algorithms such as Basic Particle Swarm Optimization (PSO), Second Generation of Particle Swarm Optimization (SGPSO), and a New Model of PSO called NMPSO. The aim of these algorithms is to evolve, at the same time, the three principal components of an ANN: the set of synaptic weights, the connections or architecture, and the transfer functions for each neuron. Eight different fitness functions were proposed to evaluate the fitness of each solution and find the best design. These functions are based on the mean square error (MSE) and the classification error (CER) and implement a strategy to avoid overtraining and to reduce the number of connections in the ANN. In addition, the ANN designed with the proposed methodology is compared with those designed manually using the well-known Back-Propagation and Levenberg-Marquardt Learning Algorithms. Finally, the accuracy of the method is tested with different nonlinear pattern classification problems. PMID:26221132

Study on application of adaptive fuzzy control and neural network in the automatic leveling system

NASA Astrophysics Data System (ADS)

Xu, Xiping; Zhao, Zizhao; Lan, Weiyong; Sha, Lei; Qian, Cheng

2015-04-01

This paper discusses the adaptive fuzzy control and neural network BP algorithm in large flat automatic leveling control system application. The purpose is to develop a measurement system with a flat quick leveling, Make the installation on the leveling system of measurement with tablet, to be able to achieve a level in precision measurement work quickly, improve the efficiency of the precision measurement. This paper focuses on the automatic leveling system analysis based on fuzzy controller, Use of the method of combining fuzzy controller and BP neural network, using BP algorithm improve the experience rules .Construct an adaptive fuzzy control system. Meanwhile the learning rate of the BP algorithm has also been run-rate adjusted to accelerate convergence. The simulation results show that the proposed control method can effectively improve the leveling precision of automatic leveling system and shorten the time of leveling.

A New Stochastic Technique for Painlevé Equation-I Using Neural Network Optimized with Swarm Intelligence

PubMed Central

Raja, Muhammad Asif Zahoor; Khan, Junaid Ali; Ahmad, Siraj-ul-Islam; Qureshi, Ijaz Mansoor

2012-01-01

A methodology for solution of Painlevé equation-I is presented using computational intelligence technique based on neural networks and particle swarm optimization hybridized with active set algorithm. The mathematical model of the equation is developed with the help of linear combination of feed-forward artificial neural networks that define the unsupervised error of the model. This error is minimized subject to the availability of appropriate weights of the networks. The learning of the weights is carried out using particle swarm optimization algorithm used as a tool for viable global search method, hybridized with active set algorithm for rapid local convergence. The accuracy, convergence rate, and computational complexity of the scheme are analyzed based on large number of independents runs and their comprehensive statistical analysis. The comparative studies of the results obtained are made with MATHEMATICA solutions, as well as, with variational iteration method and homotopy perturbation method. PMID:22919371

Implementation study of an analog spiking neural network for assisting cardiac delay prediction in a cardiac resynchronization therapy device.

PubMed

Sun, Qing; Schwartz, François; Michel, Jacques; Herve, Yannick; Dalmolin, Renzo

2011-06-01

In this paper, we aim at developing an analog spiking neural network (SNN) for reinforcing the performance of conventional cardiac resynchronization therapy (CRT) devices (also called biventricular pacemakers). Targeting an alternative analog solution in 0.13- μm CMOS technology, this paper proposes an approach to improve cardiac delay predictions in every cardiac period in order to assist the CRT device to provide real-time optimal heartbeats. The primary analog SNN architecture is proposed and its implementation is studied to fulfill the requirement of very low energy consumption. By using the Hebbian learning and reinforcement learning algorithms, the intended adaptive CRT device works with different functional modes. The simulations of both learning algorithms have been carried out, and they were shown to demonstrate the global functionalities. To improve the realism of the system, we introduce various heart behavior models (with constant/variable heart rates) that allow pathologic simulations with/without noise on the signals of the input sensors. The simulations of the global system (pacemaker models coupled with heart models) have been investigated and used to validate the analog spiking neural network implementation.

Learning in stochastic neural networks for constraint satisfaction problems

NASA Technical Reports Server (NTRS)

Johnston, Mark D.; Adorf, Hans-Martin

1989-01-01

Researchers describe a newly-developed artificial neural network algorithm for solving constraint satisfaction problems (CSPs) which includes a learning component that can significantly improve the performance of the network from run to run. The network, referred to as the Guarded Discrete Stochastic (GDS) network, is based on the discrete Hopfield network but differs from it primarily in that auxiliary networks (guards) are asymmetrically coupled to the main network to enforce certain types of constraints. Although the presence of asymmetric connections implies that the network may not converge, it was found that, for certain classes of problems, the network often quickly converges to find satisfactory solutions when they exist. The network can run efficiently on serial machines and can find solutions to very large problems (e.g., N-queens for N as large as 1024). One advantage of the network architecture is that network connection strengths need not be instantiated when the network is established: they are needed only when a participating neural element transitions from off to on. They have exploited this feature to devise a learning algorithm, based on consistency techniques for discrete CSPs, that updates the network biases and connection strengths and thus improves the network performance.

Spatial-Temporal Reasoning Applications of Computational Intelligence in the Game of Go and Computer Networks

DTIC Science & Technology

2012-01-01

dimensionality, Tesauro used a backpropagation- based , three-layer neural network and implemented the outcome from a self-play game as the reinforcement signal...a school of fish, flock of birds, and colony of ants. Our literature review reveals that no one has used PSO to train the neural network ...trained with a variant of PSO called cellular PSO (CPSO). CSRN is a supervised learning neural network (SLNN). The proposed algorithm for the

A Multiobjective Sparse Feature Learning Model for Deep Neural Networks.

PubMed

Gong, Maoguo; Liu, Jia; Li, Hao; Cai, Qing; Su, Linzhi

2015-12-01

Hierarchical deep neural networks are currently popular learning models for imitating the hierarchical architecture of human brain. Single-layer feature extractors are the bricks to build deep networks. Sparse feature learning models are popular models that can learn useful representations. But most of those models need a user-defined constant to control the sparsity of representations. In this paper, we propose a multiobjective sparse feature learning model based on the autoencoder. The parameters of the model are learnt by optimizing two objectives, reconstruction error and the sparsity of hidden units simultaneously to find a reasonable compromise between them automatically. We design a multiobjective induced learning procedure for this model based on a multiobjective evolutionary algorithm. In the experiments, we demonstrate that the learning procedure is effective, and the proposed multiobjective model can learn useful sparse features.

Neural network for image compression

NASA Astrophysics Data System (ADS)

Panchanathan, Sethuraman; Yeap, Tet H.; Pilache, B.

1992-09-01

In this paper, we propose a new scheme for image compression using neural networks. Image data compression deals with minimization of the amount of data required to represent an image while maintaining an acceptable quality. Several image compression techniques have been developed in recent years. We note that the coding performance of these techniques may be improved by employing adaptivity. Over the last few years neural network has emerged as an effective tool for solving a wide range of problems involving adaptivity and learning. A multilayer feed-forward neural network trained using the backward error propagation algorithm is used in many applications. However, this model is not suitable for image compression because of its poor coding performance. Recently, a self-organizing feature map (SOFM) algorithm has been proposed which yields a good coding performance. However, this algorithm requires a long training time because the network starts with random initial weights. In this paper we have used the backward error propagation algorithm (BEP) to quickly obtain the initial weights which are then used to speedup the training time required by the SOFM algorithm. The proposed approach (BEP-SOFM) combines the advantages of the two techniques and, hence, achieves a good coding performance in a shorter training time. Our simulation results demonstrate the potential gains using the proposed technique.

Deep learning in the small sample size setting: cascaded feed forward neural networks for medical image segmentation

NASA Astrophysics Data System (ADS)

Gaonkar, Bilwaj; Hovda, David; Martin, Neil; Macyszyn, Luke

2016-03-01

Deep Learning, refers to large set of neural network based algorithms, have emerged as promising machine- learning tools in the general imaging and computer vision domains. Convolutional neural networks (CNNs), a specific class of deep learning algorithms, have been extremely effective in object recognition and localization in natural images. A characteristic feature of CNNs, is the use of a locally connected multi layer topology that is inspired by the animal visual cortex (the most powerful vision system in existence). While CNNs, perform admirably in object identification and localization tasks, typically require training on extremely large datasets. Unfortunately, in medical image analysis, large datasets are either unavailable or are extremely expensive to obtain. Further, the primary tasks in medical imaging are organ identification and segmentation from 3D scans, which are different from the standard computer vision tasks of object recognition. Thus, in order to translate the advantages of deep learning to medical image analysis, there is a need to develop deep network topologies and training methodologies, that are geared towards medical imaging related tasks and can work in a setting where dataset sizes are relatively small. In this paper, we present a technique for stacked supervised training of deep feed forward neural networks for segmenting organs from medical scans. Each `neural network layer' in the stack is trained to identify a sub region of the original image, that contains the organ of interest. By layering several such stacks together a very deep neural network is constructed. Such a network can be used to identify extremely small regions of interest in extremely large images, inspite of a lack of clear contrast in the signal or easily identifiable shape characteristics. What is even more intriguing is that the network stack achieves accurate segmentation even when it is trained on a single image with manually labelled ground truth. We validate this approach,using a publicly available head and neck CT dataset. We also show that a deep neural network of similar depth, if trained directly using backpropagation, cannot acheive the tasks achieved using our layer wise training paradigm.

Application of furniture images selection based on neural network

NASA Astrophysics Data System (ADS)

Wang, Yong; Gao, Wenwen; Wang, Ying

2018-05-01

In the construction of 2 million furniture image databases, aiming at the problem of low quality of database, a combination of CNN and Metric learning algorithm is proposed, which makes it possible to quickly and accurately remove duplicate and irrelevant samples in the furniture image database. Solve problems that images screening method is complex, the accuracy is not high, time-consuming is long. Deep learning algorithm achieve excellent image matching ability in actual furniture retrieval applications after improving data quality.

Magnified gradient function with deterministic weight modification in adaptive learning.

PubMed

Ng, Sin-Chun; Cheung, Chi-Chung; Leung, Shu-Hung

2004-11-01

This paper presents two novel approaches, backpropagation (BP) with magnified gradient function (MGFPROP) and deterministic weight modification (DWM), to speed up the convergence rate and improve the global convergence capability of the standard BP learning algorithm. The purpose of MGFPROP is to increase the convergence rate by magnifying the gradient function of the activation function, while the main objective of DWM is to reduce the system error by changing the weights of a multilayered feedforward neural network in a deterministic way. Simulation results show that the performance of the above two approaches is better than BP and other modified BP algorithms for a number of learning problems. Moreover, the integration of the above two approaches forming a new algorithm called MDPROP, can further improve the performance of MGFPROP and DWM. From our simulation results, the MDPROP algorithm always outperforms BP and other modified BP algorithms in terms of convergence rate and global convergence capability.

A rapid learning and dynamic stepwise updating algorithm for flat neural networks and the application to time-series prediction.

PubMed

Chen, C P; Wan, J Z

1999-01-01

A fast learning algorithm is proposed to find an optimal weights of the flat neural networks (especially, the functional-link network). Although the flat networks are used for nonlinear function approximation, they can be formulated as linear systems. Thus, the weights of the networks can be solved easily using a linear least-square method. This formulation makes it easier to update the weights instantly for both a new added pattern and a new added enhancement node. A dynamic stepwise updating algorithm is proposed to update the weights of the system on-the-fly. The model is tested on several time-series data including an infrared laser data set, a chaotic time-series, a monthly flour price data set, and a nonlinear system identification problem. The simulation results are compared to existing models in which more complex architectures and more costly training are needed. The results indicate that the proposed model is very attractive to real-time processes.

Online Learning Flight Control for Intelligent Flight Control Systems (IFCS)

NASA Technical Reports Server (NTRS)

Niewoehner, Kevin R.; Carter, John (Technical Monitor)

2001-01-01

The research accomplishments for the cooperative agreement 'Online Learning Flight Control for Intelligent Flight Control Systems (IFCS)' include the following: (1) previous IFC program data collection and analysis; (2) IFC program support site (configured IFC systems support network, configured Tornado/VxWorks OS development system, made Configuration and Documentation Management Systems Internet accessible); (3) Airborne Research Test Systems (ARTS) II Hardware (developed hardware requirements specification, developing environmental testing requirements, hardware design, and hardware design development); (4) ARTS II software development laboratory unit (procurement of lab style hardware, configured lab style hardware, and designed interface module equivalent to ARTS II faceplate); (5) program support documentation (developed software development plan, configuration management plan, and software verification and validation plan); (6) LWR algorithm analysis (performed timing and profiling on algorithm); (7) pre-trained neural network analysis; (8) Dynamic Cell Structures (DCS) Neural Network Analysis (performing timing and profiling on algorithm); and (9) conducted technical interchange and quarterly meetings to define IFC research goals.

Wireless Sensor Network Congestion Control Based on Standard Particle Swarm Optimization and Single Neuron PID

PubMed Central

Yang, Xiaoping; Chen, Xueying; Xia, Riting; Qian, Zhihong

2018-01-01

Aiming at the problem of network congestion caused by the large number of data transmissions in wireless routing nodes of wireless sensor network (WSN), this paper puts forward an algorithm based on standard particle swarm–neural PID congestion control (PNPID). Firstly, PID control theory was applied to the queue management of wireless sensor nodes. Then, the self-learning and self-organizing ability of neurons was used to achieve online adjustment of weights to adjust the proportion, integral and differential parameters of the PID controller. Finally, the standard particle swarm optimization to neural PID (NPID) algorithm of initial values of proportion, integral and differential parameters and neuron learning rates were used for online optimization. This paper describes experiments and simulations which show that the PNPID algorithm effectively stabilized queue length near the expected value. At the same time, network performance, such as throughput and packet loss rate, was greatly improved, which alleviated network congestion and improved network QoS. PMID:29671822

Wireless Sensor Network Congestion Control Based on Standard Particle Swarm Optimization and Single Neuron PID.

PubMed

Yang, Xiaoping; Chen, Xueying; Xia, Riting; Qian, Zhihong

2018-04-19

Aiming at the problem of network congestion caused by the large number of data transmissions in wireless routing nodes of wireless sensor network (WSN), this paper puts forward an algorithm based on standard particle swarm⁻neural PID congestion control (PNPID). Firstly, PID control theory was applied to the queue management of wireless sensor nodes. Then, the self-learning and self-organizing ability of neurons was used to achieve online adjustment of weights to adjust the proportion, integral and differential parameters of the PID controller. Finally, the standard particle swarm optimization to neural PID (NPID) algorithm of initial values of proportion, integral and differential parameters and neuron learning rates were used for online optimization. This paper describes experiments and simulations which show that the PNPID algorithm effectively stabilized queue length near the expected value. At the same time, network performance, such as throughput and packet loss rate, was greatly improved, which alleviated network congestion and improved network QoS.

Why don’t you use Evolutionary Algorithms in Big Data?

NASA Astrophysics Data System (ADS)

Stanovov, Vladimir; Brester, Christina; Kolehmainen, Mikko; Semenkina, Olga

2017-02-01

In this paper we raise the question of using evolutionary algorithms in the area of Big Data processing. We show that evolutionary algorithms provide evident advantages due to their high scalability and flexibility, their ability to solve global optimization problems and optimize several criteria at the same time for feature selection, instance selection and other data reduction problems. In particular, we consider the usage of evolutionary algorithms with all kinds of machine learning tools, such as neural networks and fuzzy systems. All our examples prove that Evolutionary Machine Learning is becoming more and more important in data analysis and we expect to see the further development of this field especially in respect to Big Data.

Maximization of Learning Speed Due to Neuronal Redundancy in Reinforcement Learning

NASA Astrophysics Data System (ADS)

Takiyama, Ken

2016-11-01

Adaptable neural activity contributes to the flexibility of human behavior, which is optimized in situations such as motor learning and decision making. Although learning signals in motor learning and decision making are low-dimensional, neural activity, which is very high dimensional, must be modified to achieve optimal performance based on the low-dimensional signal, resulting in a severe credit-assignment problem. Despite this problem, the human brain contains a vast number of neurons, leaving an open question: what is the functional significance of the huge number of neurons? Here, I address this question by analyzing a redundant neural network with a reinforcement-learning algorithm in which the numbers of neurons and output units are N and M, respectively. Because many combinations of neural activity can generate the same output under the condition of N ≫ M, I refer to the index N - M as neuronal redundancy. Although greater neuronal redundancy makes the credit-assignment problem more severe, I demonstrate that a greater degree of neuronal redundancy facilitates learning speed. Thus, in an apparent contradiction of the credit-assignment problem, I propose the hypothesis that a functional role of a huge number of neurons or a huge degree of neuronal redundancy is to facilitate learning speed.

Neuromorphic learning of continuous-valued mappings from noise-corrupted data

NASA Technical Reports Server (NTRS)

Troudet, T.; Merrill, W.

1991-01-01

The effect of noise on the learning performance of the backpropagation algorithm is analyzed. A selective sampling of the training set is proposed to maximize the learning of control laws by backpropagation, when the data have been corrupted by noise. The training scheme is applied to the nonlinear control of a cart-pole system in the presence of noise. The neural computation provides the neurocontroller with good noise-filtering properties. In the presence of plant noise, the neurocontroller is found to be more stable than the teacher. A novel perspective on the application of neural network technology to control engineering is presented.

Model-Free Optimal Tracking Control via Critic-Only Q-Learning.

PubMed

Luo, Biao; Liu, Derong; Huang, Tingwen; Wang, Ding

2016-10-01

Model-free control is an important and promising topic in control fields, which has attracted extensive attention in the past few years. In this paper, we aim to solve the model-free optimal tracking control problem of nonaffine nonlinear discrete-time systems. A critic-only Q-learning (CoQL) method is developed, which learns the optimal tracking control from real system data, and thus avoids solving the tracking Hamilton-Jacobi-Bellman equation. First, the Q-learning algorithm is proposed based on the augmented system, and its convergence is established. Using only one neural network for approximating the Q-function, the CoQL method is developed to implement the Q-learning algorithm. Furthermore, the convergence of the CoQL method is proved with the consideration of neural network approximation error. With the convergent Q-function obtained from the CoQL method, the adaptive optimal tracking control is designed based on the gradient descent scheme. Finally, the effectiveness of the developed CoQL method is demonstrated through simulation studies. The developed CoQL method learns with off-policy data and implements with a critic-only structure, thus it is easy to realize and overcome the inadequate exploration problem.

Hexacopter trajectory control using a neural network

NASA Astrophysics Data System (ADS)

Artale, V.; Collotta, M.; Pau, G.; Ricciardello, A.

2013-10-01

The modern flight control systems are complex due to their non-linear nature. In fact, modern aerospace vehicles are expected to have non-conventional flight envelopes and, then, they must guarantee a high level of robustness and adaptability in order to operate in uncertain environments. Neural Networks (NN), with real-time learning capability, for flight control can be used in applications with manned or unmanned aerial vehicles. Indeed, using proven lower level control algorithms with adaptive elements that exhibit long term learning could help in achieving better adaptation performance while performing aggressive maneuvers. In this paper we show a mathematical modeling and a Neural Network for a hexacopter dynamics in order to develop proper methods for stabilization and trajectory control.

Intelligent control of robotic arm/hand systems for the NASA EVA retriever using neural networks

NASA Technical Reports Server (NTRS)

Mclauchlan, Robert A.

1989-01-01

Adaptive/general learning algorithms using varying neural network models are considered for the intelligent control of robotic arm plus dextrous hand/manipulator systems. Results are summarized and discussed for the use of the Barto/Sutton/Anderson neuronlike, unsupervised learning controller as applied to the stabilization of an inverted pendulum on a cart system. Recommendations are made for the application of the controller and a kinematic analysis for trajectory planning to simple object retrieval (chase/approach and capture/grasp) scenarios in two dimensions.

Improving Pattern Recognition and Neural Network Algorithms with Applications to Solar Panel Energy Optimization

NASA Astrophysics Data System (ADS)

Zamora Ramos, Ernesto

Artificial Intelligence is a big part of automation and with today's technological advances, artificial intelligence has taken great strides towards positioning itself as the technology of the future to control, enhance and perfect automation. Computer vision includes pattern recognition and classification and machine learning. Computer vision is at the core of decision making and it is a vast and fruitful branch of artificial intelligence. In this work, we expose novel algorithms and techniques built upon existing technologies to improve pattern recognition and neural network training, initially motivated by a multidisciplinary effort to build a robot that helps maintain and optimize solar panel energy production. Our contributions detail an improved non-linear pre-processing technique to enhance poorly illuminated images based on modifications to the standard histogram equalization for an image. While the original motivation was to improve nocturnal navigation, the results have applications in surveillance, search and rescue, medical imaging enhancing, and many others. We created a vision system for precise camera distance positioning motivated to correctly locate the robot for capture of solar panel images for classification. The classification algorithm marks solar panels as clean or dirty for later processing. Our algorithm extends past image classification and, based on historical and experimental data, it identifies the optimal moment in which to perform maintenance on marked solar panels as to minimize the energy and profit loss. In order to improve upon the classification algorithm, we delved into feedforward neural networks because of their recent advancements, proven universal approximation and classification capabilities, and excellent recognition rates. We explore state-of-the-art neural network training techniques offering pointers and insights, culminating on the implementation of a complete library with support for modern deep learning architectures, multilayer percepterons and convolutional neural networks. Our research with neural networks has encountered a great deal of difficulties regarding hyperparameter estimation for good training convergence rate and accuracy. Most hyperparameters, including architecture, learning rate, regularization, trainable parameters (or weights) initialization, and so on, are chosen via a trial and error process with some educated guesses. However, we developed the first quantitative method to compare weight initialization strategies, a critical hyperparameter choice during training, to estimate among a group of candidate strategies which would make the network converge to the highest classification accuracy faster with high probability. Our method provides a quick, objective measure to compare initialization strategies to select the best possible among them beforehand without having to complete multiple training sessions for each candidate strategy to compare final results.

Neural correlates of forward planning in a spatial decision task in humans

PubMed Central

Simon, Dylan Alexander; Daw, Nathaniel D.

2011-01-01

Although reinforcement learning (RL) theories have been influential in characterizing the brain’s mechanisms for reward-guided choice, the predominant temporal difference (TD) algorithm cannot explain many flexible or goal-directed actions that have been demonstrated behaviorally. We investigate such actions by contrasting an RL algorithm that is model-based, in that it relies on learning a map or model of the task and planning within it, to traditional model-free TD learning. To distinguish these approaches in humans, we used fMRI in a continuous spatial navigation task, in which frequent changes to the layout of the maze forced subjects continually to relearn their favored routes, thereby exposing the RL mechanisms employed. We sought evidence for the neural substrates of such mechanisms by comparing choice behavior and BOLD signals to decision variables extracted from simulations of either algorithm. Both choices and value-related BOLD signals in striatum, though most often associated with TD learning, were better explained by the model-based theory. Further, predecessor quantities for the model-based value computation were correlated with BOLD signals in the medial temporal lobe and frontal cortex. These results point to a significant extension of both the computational and anatomical substrates for RL in the brain. PMID:21471389

Fast temporal neural learning using teacher forcing

NASA Technical Reports Server (NTRS)

Toomarian, Nikzad (Inventor); Bahren, Jacob (Inventor)

1992-01-01

A neural network is trained to output a time dependent target vector defined over a predetermined time interval in response to a time dependent input vector defined over the same time interval by applying corresponding elements of the error vector, or difference between the target vector and the actual neuron output vector, to the inputs of corresponding output neurons of the network as corrective feedback. This feedback decreases the error and quickens the learning process, so that a much smaller number of training cycles are required to complete the learning process. A conventional gradient descent algorithm is employed to update the neural network parameters at the end of the predetermined time interval. The foregoing process is repeated in repetitive cycles until the actual output vector corresponds to the target vector. In the preferred embodiment, as the overall error of the neural network output decreasing during successive training cycles, the portion of the error fed back to the output neurons is decreased accordingly, allowing the network to learn with greater freedom from teacher forcing as the network parameters converge to their optimum values. The invention may also be used to train a neural network with stationary training and target vectors.

Fast temporal neural learning using teacher forcing

NASA Technical Reports Server (NTRS)

Toomarian, Nikzad (Inventor); Bahren, Jacob (Inventor)

1995-01-01

A neural network is trained to output a time dependent target vector defined over a predetermined time interval in response to a time dependent input vector defined over the same time interval by applying corresponding elements of the error vector, or difference between the target vector and the actual neuron output vector, to the inputs of corresponding output neurons of the network as corrective feedback. This feedback decreases the error and quickens the learning process, so that a much smaller number of training cycles are required to complete the learning process. A conventional gradient descent algorithm is employed to update the neural network parameters at the end of the predetermined time interval. The foregoing process is repeated in repetitive cycles until the actual output vector corresponds to the target vector. In the preferred embodiment, as the overall error of the neural network output decreasing during successive training cycles, the portion of the error fed back to the output neurons is decreased accordingly, allowing the network to learn with greater freedom from teacher forcing as the network parameters converge to their optimum values. The invention may also be used to train a neural network with stationary training and target vectors.

Learning in fully recurrent neural networks by approaching tangent planes to constraint surfaces.

PubMed

May, P; Zhou, E; Lee, C W

2012-10-01

In this paper we present a new variant of the online real time recurrent learning algorithm proposed by Williams and Zipser (1989). Whilst the original algorithm utilises gradient information to guide the search towards the minimum training error, it is very slow in most applications and often gets stuck in local minima of the search space. It is also sensitive to the choice of learning rate and requires careful tuning. The new variant adjusts weights by moving to the tangent planes to constraint surfaces. It is simple to implement and requires no parameters to be set manually. Experimental results show that this new algorithm gives significantly faster convergence whilst avoiding problems like local minima. Copyright © 2012 Elsevier Ltd. All rights reserved.

Training Spiking Neural Models Using Artificial Bee Colony

PubMed Central

Vazquez, Roberto A.; Garro, Beatriz A.

2015-01-01

Spiking neurons are models designed to simulate, in a realistic manner, the behavior of biological neurons. Recently, it has been proven that this type of neurons can be applied to solve pattern recognition problems with great efficiency. However, the lack of learning strategies for training these models do not allow to use them in several pattern recognition problems. On the other hand, several bioinspired algorithms have been proposed in the last years for solving a broad range of optimization problems, including those related to the field of artificial neural networks (ANNs). Artificial bee colony (ABC) is a novel algorithm based on the behavior of bees in the task of exploring their environment to find a food source. In this paper, we describe how the ABC algorithm can be used as a learning strategy to train a spiking neuron aiming to solve pattern recognition problems. Finally, the proposed approach is tested on several pattern recognition problems. It is important to remark that to realize the powerfulness of this type of model only one neuron will be used. In addition, we analyze how the performance of these models is improved using this kind of learning strategy. PMID:25709644

Markov Chain Monte Carlo Bayesian Learning for Neural Networks

NASA Technical Reports Server (NTRS)

Goodrich, Michael S.

2011-01-01

Conventional training methods for neural networks involve starting al a random location in the solution space of the network weights, navigating an error hyper surface to reach a minimum, and sometime stochastic based techniques (e.g., genetic algorithms) to avoid entrapment in a local minimum. It is further typically necessary to preprocess the data (e.g., normalization) to keep the training algorithm on course. Conversely, Bayesian based learning is an epistemological approach concerned with formally updating the plausibility of competing candidate hypotheses thereby obtaining a posterior distribution for the network weights conditioned on the available data and a prior distribution. In this paper, we developed a powerful methodology for estimating the full residual uncertainty in network weights and therefore network predictions by using a modified Jeffery's prior combined with a Metropolis Markov Chain Monte Carlo method.

Proceedings of the Second Joint Technology Workshop on Neural Networks and Fuzzy Logic, volume 1

NASA Technical Reports Server (NTRS)

Lea, Robert N. (Editor); Villarreal, James (Editor)

1991-01-01

Documented here are papers presented at the Neural Networks and Fuzzy Logic Workshop sponsored by NASA and the University of Houston, Clear Lake. The workshop was held April 11 to 13 at the Johnson Space Flight Center. Technical topics addressed included adaptive systems, learning algorithms, network architectures, vision, robotics, neurobiological connections, speech recognition and synthesis, fuzzy set theory and application, control and dynamics processing, space applications, fuzzy logic and neural network computers, approximate reasoning, and multiobject decision making.

Neurofeedback Training for BCI Control

NASA Astrophysics Data System (ADS)

Neuper, Christa; Pfurtscheller, Gert

Brain-computer interface (BCI) systems detect changes in brain signals that reflect human intention, then translate these signals to control monitors or external devices (for a comprehensive review, see [1]). BCIs typically measure electrical signals resulting from neural firing (i.e. neuronal action potentials, Electroencephalogram (ECoG), or Electroencephalogram (EEG)). Sophisticated pattern recognition and classification algorithms convert neural activity into the required control signals. BCI research has focused heavily on developing powerful signal processing and machine learning techniques to accurately classify neural activity [2-4].

Robust Blind Learning Algorithm for Nonlinear Equalization Using Input Decision Information.

PubMed

Xu, Lu; Huang, Defeng David; Guo, Yingjie Jay

2015-12-01

In this paper, we propose a new blind learning algorithm, namely, the Benveniste-Goursat input-output decision (BG-IOD), to enhance the convergence performance of neural network-based equalizers for nonlinear channel equalization. In contrast to conventional blind learning algorithms, where only the output of the equalizer is employed for updating system parameters, the BG-IOD exploits a new type of extra information, the input decision information obtained from the input of the equalizer, to mitigate the influence of the nonlinear equalizer structure on parameters learning, thereby leading to improved convergence performance. We prove that, with the input decision information, a desirable convergence capability that the output symbol error rate (SER) is always less than the input SER if the input SER is below a threshold, can be achieved. Then, the BG soft-switching technique is employed to combine the merits of both input and output decision information, where the former is used to guarantee SER convergence and the latter is to improve SER performance. Simulation results show that the proposed algorithm outperforms conventional blind learning algorithms, such as stochastic quadratic distance and dual mode constant modulus algorithm, in terms of both convergence performance and SER performance, for nonlinear equalization.

Evaluation of a parallel implementation of the learning portion of the backward error propagation neural network: experiments in artifact identification.

PubMed Central

Sittig, D. F.; Orr, J. A.

1991-01-01

Various methods have been proposed in an attempt to solve problems in artifact and/or alarm identification including expert systems, statistical signal processing techniques, and artificial neural networks (ANN). ANNs consist of a large number of simple processing units connected by weighted links. To develop truly robust ANNs, investigators are required to train their networks on huge training data sets, requiring enormous computing power. We implemented a parallel version of the backward error propagation neural network training algorithm in the widely portable parallel programming language C-Linda. A maximum speedup of 4.06 was obtained with six processors. This speedup represents a reduction in total run-time from approximately 6.4 hours to 1.5 hours. We conclude that use of the master-worker model of parallel computation is an excellent method for obtaining speedups in the backward error propagation neural network training algorithm. PMID:1807607

Next Day Building Load Predictions based on Limited Input Features Using an On-Line Laterally Primed Adaptive Resonance Theory Artificial Neural Network.

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

Jones, Christian Birk; Robinson, Matt; Yasaei, Yasser

Optimal integration of thermal energy storage within commercial building applications requires accurate load predictions. Several methods exist that provide an estimate of a buildings future needs. Methods include component-based models and data-driven algorithms. This work implemented a previously untested algorithm for this application that is called a Laterally Primed Adaptive Resonance Theory (LAPART) artificial neural network (ANN). The LAPART algorithm provided accurate results over a two month period where minimal historical data and a small amount of input types were available. These results are significant, because common practice has often overlooked the implementation of an ANN. ANN have often beenmore » perceived to be too complex and require large amounts of data to provide accurate results. The LAPART neural network was implemented in an on-line learning manner. On-line learning refers to the continuous updating of training data as time occurs. For this experiment, training began with a singe day and grew to two months of data. This approach provides a platform for immediate implementation that requires minimal time and effort. The results from the LAPART algorithm were compared with statistical regression and a component-based model. The comparison was based on the predictions linear relationship with the measured data, mean squared error, mean bias error, and cost savings achieved by the respective prediction techniques. The results show that the LAPART algorithm provided a reliable and cost effective means to predict the building load for the next day.« less

Invariant-feature-based adaptive automatic target recognition in obscured 3D point clouds

NASA Astrophysics Data System (ADS)

Khuon, Timothy; Kershner, Charles; Mattei, Enrico; Alverio, Arnel; Rand, Robert

2014-06-01

Target recognition and classification in a 3D point cloud is a non-trivial process due to the nature of the data collected from a sensor system. The signal can be corrupted by noise from the environment, electronic system, A/D converter, etc. Therefore, an adaptive system with a desired tolerance is required to perform classification and recognition optimally. The feature-based pattern recognition algorithm architecture as described below is particularly devised for solving a single-sensor classification non-parametrically. Feature set is extracted from an input point cloud, normalized, and classifier a neural network classifier. For instance, automatic target recognition in an urban area would require different feature sets from one in a dense foliage area. The figure above (see manuscript) illustrates the architecture of the feature based adaptive signature extraction of 3D point cloud including LIDAR, RADAR, and electro-optical data. This network takes a 3D cluster and classifies it into a specific class. The algorithm is a supervised and adaptive classifier with two modes: the training mode and the performing mode. For the training mode, a number of novel patterns are selected from actual or artificial data. A particular 3D cluster is input to the network as shown above for the decision class output. The network consists of three sequential functional modules. The first module is for feature extraction that extracts the input cluster into a set of singular value features or feature vector. Then the feature vector is input into the feature normalization module to normalize and balance it before being fed to the neural net classifier for the classification. The neural net can be trained by actual or artificial novel data until each trained output reaches the declared output within the defined tolerance. In case new novel data is added after the neural net has been learned, the training is then resumed until the neural net has incrementally learned with the new novel data. The associative memory capability of the neural net enables the incremental learning. The back propagation algorithm or support vector machine can be utilized for the classification and recognition.

Evaluation of machine learning algorithms for improved risk assessment for Down's syndrome.

PubMed

Koivu, Aki; Korpimäki, Teemu; Kivelä, Petri; Pahikkala, Tapio; Sairanen, Mikko

2018-05-04

Prenatal screening generates a great amount of data that is used for predicting risk of various disorders. Prenatal risk assessment is based on multiple clinical variables and overall performance is defined by how well the risk algorithm is optimized for the population in question. This article evaluates machine learning algorithms to improve performance of first trimester screening of Down syndrome. Machine learning algorithms pose an adaptive alternative to develop better risk assessment models using the existing clinical variables. Two real-world data sets were used to experiment with multiple classification algorithms. Implemented models were tested with a third, real-world, data set and performance was compared to a predicate method, a commercial risk assessment software. Best performing deep neural network model gave an area under the curve of 0.96 and detection rate of 78% with 1% false positive rate with the test data. Support vector machine model gave area under the curve of 0.95 and detection rate of 61% with 1% false positive rate with the same test data. When compared with the predicate method, the best support vector machine model was slightly inferior, but an optimized deep neural network model was able to give higher detection rates with same false positive rate or similar detection rate but with markedly lower false positive rate. This finding could further improve the first trimester screening for Down syndrome, by using existing clinical variables and a large training data derived from a specific population. Copyright © 2018 Elsevier Ltd. All rights reserved.

Evolution, learning, and cognition

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

Lee, Y.C.

1988-01-01

The book comprises more than fifteen articles in the areas of neural networks and connectionist systems, classifier systems, adaptive network systems, genetic algorithm, cellular automata, artificial immune systems, evolutionary genetics, cognitive science, optical computing, combinatorial optimization, and cybernetics.

Neural manufacturing: a novel concept for processing modeling, monitoring, and control

NASA Astrophysics Data System (ADS)

Fu, Chi Y.; Petrich, Loren; Law, Benjamin

1995-09-01

Semiconductor fabrication lines have become extremely costly, and achieving a good return from such a high capital investment requires efficient utilization of these expensive facilities. It is highly desirable to shorten processing development time, increase fabrication yield, enhance flexibility, improve quality, and minimize downtime. We propose that these ends can be achieved by applying recent advances in the areas of artificial neural networks, fuzzy logic, machine learning, and genetic algorithms. We use the term neural manufacturing to describe such applications. This paper describes our use of artificial neural networks to improve the monitoring and control of semiconductor process.

Learning Instance-Specific Predictive Models

PubMed Central

Visweswaran, Shyam; Cooper, Gregory F.

2013-01-01

This paper introduces a Bayesian algorithm for constructing predictive models from data that are optimized to predict a target variable well for a particular instance. This algorithm learns Markov blanket models, carries out Bayesian model averaging over a set of models to predict a target variable of the instance at hand, and employs an instance-specific heuristic to locate a set of suitable models to average over. We call this method the instance-specific Markov blanket (ISMB) algorithm. The ISMB algorithm was evaluated on 21 UCI data sets using five different performance measures and its performance was compared to that of several commonly used predictive algorithms, including nave Bayes, C4.5 decision tree, logistic regression, neural networks, k-Nearest Neighbor, Lazy Bayesian Rules, and AdaBoost. Over all the data sets, the ISMB algorithm performed better on average on all performance measures against all the comparison algorithms. PMID:25045325

Smoothing of cost function leads to faster convergence of neural network learning

NASA Astrophysics Data System (ADS)

Xu, Li-Qun; Hall, Trevor J.

1994-03-01

One of the major problems in supervised learning of neural networks is the inevitable local minima inherent in the cost function f(W,D). This often makes classic gradient-descent-based learning algorithms that calculate the weight updates for each iteration according to (Delta) W(t) equals -(eta) (DOT)$DELwf(W,D) powerless. In this paper we describe a new strategy to solve this problem, which, adaptively, changes the learning rate and manipulates the gradient estimator simultaneously. The idea is to implicitly convert the local- minima-laden cost function f((DOT)) into a sequence of its smoothed versions {f(beta t)}Ttequals1, which, subject to the parameter (beta) t, bears less details at time t equals 1 and gradually more later on, the learning is actually performed on this sequence of functionals. The corresponding smoothed global minima obtained in this way, {Wt}Ttequals1, thus progressively approximate W-the desired global minimum. Experimental results on a nonconvex function minimization problem and a typical neural network learning task are given, analyses and discussions of some important issues are provided.

Recursive Bayesian recurrent neural networks for time-series modeling.

PubMed

Mirikitani, Derrick T; Nikolaev, Nikolay

2010-02-01

This paper develops a probabilistic approach to recursive second-order training of recurrent neural networks (RNNs) for improved time-series modeling. A general recursive Bayesian Levenberg-Marquardt algorithm is derived to sequentially update the weights and the covariance (Hessian) matrix. The main strengths of the approach are a principled handling of the regularization hyperparameters that leads to better generalization, and stable numerical performance. The framework involves the adaptation of a noise hyperparameter and local weight prior hyperparameters, which represent the noise in the data and the uncertainties in the model parameters. Experimental investigations using artificial and real-world data sets show that RNNs equipped with the proposed approach outperform standard real-time recurrent learning and extended Kalman training algorithms for recurrent networks, as well as other contemporary nonlinear neural models, on time-series modeling.

Noise reduction and image enhancement using a hardware implementation of artificial neural networks

NASA Astrophysics Data System (ADS)

David, Robert; Williams, Erin; de Tremiolles, Ghislain; Tannhof, Pascal

1999-03-01

In this paper, we present a neural based solution developed for noise reduction and image enhancement using the ZISC, an IBM hardware processor which implements the Restricted Coulomb Energy algorithm and the K-Nearest Neighbor algorithm. Artificial neural networks present the advantages of processing time reduction in comparison with classical models, adaptability, and the weighted property of pattern learning. The goal of the developed application is image enhancement in order to restore old movies (noise reduction, focus correction, etc.), to improve digital television images, or to treat images which require adaptive processing (medical images, spatial images, special effects, etc.). Image results show a quantitative improvement over the noisy image as well as the efficiency of this system. Further enhancements are being examined to improve the output of the system.

Solving graph data issues using a layered architecture approach with applications to web spam detection.

PubMed

Scarselli, Franco; Tsoi, Ah Chung; Hagenbuchner, Markus; Noi, Lucia Di

2013-12-01

This paper proposes the combination of two state-of-the-art algorithms for processing graph input data, viz., the probabilistic mapping graph self organizing map, an unsupervised learning approach, and the graph neural network, a supervised learning approach. We organize these two algorithms in a cascade architecture containing a probabilistic mapping graph self organizing map, and a graph neural network. We show that this combined approach helps us to limit the long-term dependency problem that exists when training the graph neural network resulting in an overall improvement in performance. This is demonstrated in an application to a benchmark problem requiring the detection of spam in a relatively large set of web sites. It is found that the proposed method produces results which reach the state of the art when compared with some of the best results obtained by others using quite different approaches. A particular strength of our method is its applicability towards any input domain which can be represented as a graph. Copyright © 2013 Elsevier Ltd. All rights reserved.

Algorithm for Training a Recurrent Multilayer Perceptron

NASA Technical Reports Server (NTRS)

Parlos, Alexander G.; Rais, Omar T.; Menon, Sunil K.; Atiya, Amir F.

2004-01-01

An improved algorithm has been devised for training a recurrent multilayer perceptron (RMLP) for optimal performance in predicting the behavior of a complex, dynamic, and noisy system multiple time steps into the future. [An RMLP is a computational neural network with self-feedback and cross-talk (both delayed by one time step) among neurons in hidden layers]. Like other neural-network-training algorithms, this algorithm adjusts network biases and synaptic-connection weights according to a gradient-descent rule. The distinguishing feature of this algorithm is a combination of global feedback (the use of predictions as well as the current output value in computing the gradient at each time step) and recursiveness. The recursive aspect of the algorithm lies in the inclusion of the gradient of predictions at each time step with respect to the predictions at the preceding time step; this recursion enables the RMLP to learn the dynamics. It has been conjectured that carrying the recursion to even earlier time steps would enable the RMLP to represent a noisier, more complex system.

Intelligent neural network and fuzzy logic control of industrial and power systems

NASA Astrophysics Data System (ADS)

Kuljaca, Ognjen

The main role played by neural network and fuzzy logic intelligent control algorithms today is to identify and compensate unknown nonlinear system dynamics. There are a number of methods developed, but often the stability analysis of neural network and fuzzy control systems was not provided. This work will meet those problems for the several algorithms. Some more complicated control algorithms included backstepping and adaptive critics will be designed. Nonlinear fuzzy control with nonadaptive fuzzy controllers is also analyzed. An experimental method for determining describing function of SISO fuzzy controller is given. The adaptive neural network tracking controller for an autonomous underwater vehicle is analyzed. A novel stability proof is provided. The implementation of the backstepping neural network controller for the coupled motor drives is described. Analysis and synthesis of adaptive critic neural network control is also provided in the work. Novel tuning laws for the system with action generating neural network and adaptive fuzzy critic are given. Stability proofs are derived for all those control methods. It is shown how these control algorithms and approaches can be used in practical engineering control. Stability proofs are given. Adaptive fuzzy logic control is analyzed. Simulation study is conducted to analyze the behavior of the adaptive fuzzy system on the different environment changes. A novel stability proof for adaptive fuzzy logic systems is given. Also, adaptive elastic fuzzy logic control architecture is described and analyzed. A novel membership function is used for elastic fuzzy logic system. The stability proof is proffered. Adaptive elastic fuzzy logic control is compared with the adaptive nonelastic fuzzy logic control. The work described in this dissertation serves as foundation on which analysis of particular representative industrial systems will be conducted. Also, it gives a good starting point for analysis of learning abilities of adaptive and neural network control systems, as well as for the analysis of the different algorithms such as elastic fuzzy systems.

Deterministic convergence of chaos injection-based gradient method for training feedforward neural networks.

PubMed

Zhang, Huisheng; Zhang, Ying; Xu, Dongpo; Liu, Xiaodong

2015-06-01

It has been shown that, by adding a chaotic sequence to the weight update during the training of neural networks, the chaos injection-based gradient method (CIBGM) is superior to the standard backpropagation algorithm. This paper presents the theoretical convergence analysis of CIBGM for training feedforward neural networks. We consider both the case of batch learning as well as the case of online learning. Under mild conditions, we prove the weak convergence, i.e., the training error tends to a constant and the gradient of the error function tends to zero. Moreover, the strong convergence of CIBGM is also obtained with the help of an extra condition. The theoretical results are substantiated by a simulation example.

Human-level control through deep reinforcement learning.

PubMed

Mnih, Volodymyr; Kavukcuoglu, Koray; Silver, David; Rusu, Andrei A; Veness, Joel; Bellemare, Marc G; Graves, Alex; Riedmiller, Martin; Fidjeland, Andreas K; Ostrovski, Georg; Petersen, Stig; Beattie, Charles; Sadik, Amir; Antonoglou, Ioannis; King, Helen; Kumaran, Dharshan; Wierstra, Daan; Legg, Shane; Hassabis, Demis

2015-02-26

The theory of reinforcement learning provides a normative account, deeply rooted in psychological and neuroscientific perspectives on animal behaviour, of how agents may optimize their control of an environment. To use reinforcement learning successfully in situations approaching real-world complexity, however, agents are confronted with a difficult task: they must derive efficient representations of the environment from high-dimensional sensory inputs, and use these to generalize past experience to new situations. Remarkably, humans and other animals seem to solve this problem through a harmonious combination of reinforcement learning and hierarchical sensory processing systems, the former evidenced by a wealth of neural data revealing notable parallels between the phasic signals emitted by dopaminergic neurons and temporal difference reinforcement learning algorithms. While reinforcement learning agents have achieved some successes in a variety of domains, their applicability has previously been limited to domains in which useful features can be handcrafted, or to domains with fully observed, low-dimensional state spaces. Here we use recent advances in training deep neural networks to develop a novel artificial agent, termed a deep Q-network, that can learn successful policies directly from high-dimensional sensory inputs using end-to-end reinforcement learning. We tested this agent on the challenging domain of classic Atari 2600 games. We demonstrate that the deep Q-network agent, receiving only the pixels and the game score as inputs, was able to surpass the performance of all previous algorithms and achieve a level comparable to that of a professional human games tester across a set of 49 games, using the same algorithm, network architecture and hyperparameters. This work bridges the divide between high-dimensional sensory inputs and actions, resulting in the first artificial agent that is capable of learning to excel at a diverse array of challenging tasks.

Human-level control through deep reinforcement learning

NASA Astrophysics Data System (ADS)

Mnih, Volodymyr; Kavukcuoglu, Koray; Silver, David; Rusu, Andrei A.; Veness, Joel; Bellemare, Marc G.; Graves, Alex; Riedmiller, Martin; Fidjeland, Andreas K.; Ostrovski, Georg; Petersen, Stig; Beattie, Charles; Sadik, Amir; Antonoglou, Ioannis; King, Helen; Kumaran, Dharshan; Wierstra, Daan; Legg, Shane; Hassabis, Demis

2015-02-01

The theory of reinforcement learning provides a normative account, deeply rooted in psychological and neuroscientific perspectives on animal behaviour, of how agents may optimize their control of an environment. To use reinforcement learning successfully in situations approaching real-world complexity, however, agents are confronted with a difficult task: they must derive efficient representations of the environment from high-dimensional sensory inputs, and use these to generalize past experience to new situations. Remarkably, humans and other animals seem to solve this problem through a harmonious combination of reinforcement learning and hierarchical sensory processing systems, the former evidenced by a wealth of neural data revealing notable parallels between the phasic signals emitted by dopaminergic neurons and temporal difference reinforcement learning algorithms. While reinforcement learning agents have achieved some successes in a variety of domains, their applicability has previously been limited to domains in which useful features can be handcrafted, or to domains with fully observed, low-dimensional state spaces. Here we use recent advances in training deep neural networks to develop a novel artificial agent, termed a deep Q-network, that can learn successful policies directly from high-dimensional sensory inputs using end-to-end reinforcement learning. We tested this agent on the challenging domain of classic Atari 2600 games. We demonstrate that the deep Q-network agent, receiving only the pixels and the game score as inputs, was able to surpass the performance of all previous algorithms and achieve a level comparable to that of a professional human games tester across a set of 49 games, using the same algorithm, network architecture and hyperparameters. This work bridges the divide between high-dimensional sensory inputs and actions, resulting in the first artificial agent that is capable of learning to excel at a diverse array of challenging tasks.

The computational neurobiology of learning and reward.

PubMed

Daw, Nathaniel D; Doya, Kenji

2006-04-01

Following the suggestion that midbrain dopaminergic neurons encode a signal, known as a 'reward prediction error', used by artificial intelligence algorithms for learning to choose advantageous actions, the study of the neural substrates for reward-based learning has been strongly influenced by computational theories. In recent work, such theories have been increasingly integrated into experimental design and analysis. Such hybrid approaches have offered detailed new insights into the function of a number of brain areas, especially the cortex and basal ganglia. In part this is because these approaches enable the study of neural correlates of subjective factors (such as a participant's beliefs about the reward to be received for performing some action) that the computational theories purport to quantify.

Fast, Simple and Accurate Handwritten Digit Classification by Training Shallow Neural Network Classifiers with the ‘Extreme Learning Machine’ Algorithm

PubMed Central

McDonnell, Mark D.; Tissera, Migel D.; Vladusich, Tony; van Schaik, André; Tapson, Jonathan

2015-01-01

Recent advances in training deep (multi-layer) architectures have inspired a renaissance in neural network use. For example, deep convolutional networks are becoming the default option for difficult tasks on large datasets, such as image and speech recognition. However, here we show that error rates below 1% on the MNIST handwritten digit benchmark can be replicated with shallow non-convolutional neural networks. This is achieved by training such networks using the ‘Extreme Learning Machine’ (ELM) approach, which also enables a very rapid training time (∼ 10 minutes). Adding distortions, as is common practise for MNIST, reduces error rates even further. Our methods are also shown to be capable of achieving less than 5.5% error rates on the NORB image database. To achieve these results, we introduce several enhancements to the standard ELM algorithm, which individually and in combination can significantly improve performance. The main innovation is to ensure each hidden-unit operates only on a randomly sized and positioned patch of each image. This form of random ‘receptive field’ sampling of the input ensures the input weight matrix is sparse, with about 90% of weights equal to zero. Furthermore, combining our methods with a small number of iterations of a single-batch backpropagation method can significantly reduce the number of hidden-units required to achieve a particular performance. Our close to state-of-the-art results for MNIST and NORB suggest that the ease of use and accuracy of the ELM algorithm for designing a single-hidden-layer neural network classifier should cause it to be given greater consideration either as a standalone method for simpler problems, or as the final classification stage in deep neural networks applied to more difficult problems. PMID:26262687

Neural correlates of strategic reasoning during competitive games.

PubMed

Seo, Hyojung; Cai, Xinying; Donahue, Christopher H; Lee, Daeyeol

2014-10-17

Although human and animal behaviors are largely shaped by reinforcement and punishment, choices in social settings are also influenced by information about the knowledge and experience of other decision-makers. During competitive games, monkeys increased their payoffs by systematically deviating from a simple heuristic learning algorithm and thereby countering the predictable exploitation by their computer opponent. Neurons in the dorsomedial prefrontal cortex (dmPFC) signaled the animal's recent choice and reward history that reflected the computer's exploitative strategy. The strength of switching signals in the dmPFC also correlated with the animal's tendency to deviate from the heuristic learning algorithm. Therefore, the dmPFC might provide control signals for overriding simple heuristic learning algorithms based on the inferred strategies of the opponent. Copyright © 2014, American Association for the Advancement of Science.

Learning algorithm in restricted Boltzmann machines using Kullback-Leibler importance estimation procedure

NASA Astrophysics Data System (ADS)

Yasuda, Muneki; Sakurai, Tetsuharu; Tanaka, Kazuyuki

Restricted Boltzmann machines (RBMs) are bipartite structured statistical neural networks and consist of two layers. One of them is a layer of visible units and the other one is a layer of hidden units. In each layer, any units do not connect to each other. RBMs have high flexibility and rich structure and have been expected to applied to various applications, for example, image and pattern recognitions, face detections and so on. However, most of computational models in RBMs are intractable and often belong to the class of NP-hard problem. In this paper, in order to construct a practical learning algorithm for them, we employ the Kullback-Leibler Importance Estimation Procedure (KLIEP) to RBMs, and give a new scheme of practical approximate learning algorithm for RBMs based on the KLIEP.

Learning to forget: continual prediction with LSTM.

PubMed

Gers, F A; Schmidhuber, J; Cummins, F

2000-10-01

Long short-term memory (LSTM; Hochreiter & Schmidhuber, 1997) can solve numerous tasks not solvable by previous learning algorithms for recurrent neural networks (RNNs). We identify a weakness of LSTM networks processing continual input streams that are not a priori segmented into subsequences with explicitly marked ends at which the network's internal state could be reset. Without resets, the state may grow indefinitely and eventually cause the network to break down. Our remedy is a novel, adaptive "forget gate" that enables an LSTM cell to learn to reset itself at appropriate times, thus releasing internal resources. We review illustrative benchmark problems on which standard LSTM outperforms other RNN algorithms. All algorithms (including LSTM) fail to solve continual versions of these problems. LSTM with forget gates, however, easily solves them, and in an elegant way.

Neuromorphic Learning From Noisy Data

NASA Technical Reports Server (NTRS)

Merrill, Walter C.; Troudet, Terry

1993-01-01

Two reports present numerical study of performance of feedforward neural network trained by back-propagation algorithm in learning continuous-valued mappings from data corrupted by noise. Two types of noise considered: plant noise which affects dynamics of controlled process and data-processing noise, which occurs during analog processing and digital sampling of signals. Study performed with view toward use of neural networks as neurocontrollers to substitute for, or enhance, performances of human experts in controlling mechanical devices in presence of sensor and actuator noise and to enhance performances of more-conventional digital feedback electronic process controllers in noisy environments.

A deep learning method for lincRNA detection using auto-encoder algorithm.

PubMed

Yu, Ning; Yu, Zeng; Pan, Yi

2017-12-06

RNA sequencing technique (RNA-seq) enables scientists to develop novel data-driven methods for discovering more unidentified lincRNAs. Meantime, knowledge-based technologies are experiencing a potential revolution ignited by the new deep learning methods. By scanning the newly found data set from RNA-seq, scientists have found that: (1) the expression of lincRNAs appears to be regulated, that is, the relevance exists along the DNA sequences; (2) lincRNAs contain some conversed patterns/motifs tethered together by non-conserved regions. The two evidences give the reasoning for adopting knowledge-based deep learning methods in lincRNA detection. Similar to coding region transcription, non-coding regions are split at transcriptional sites. However, regulatory RNAs rather than message RNAs are generated. That is, the transcribed RNAs participate the biological process as regulatory units instead of generating proteins. Identifying these transcriptional regions from non-coding regions is the first step towards lincRNA recognition. The auto-encoder method achieves 100% and 92.4% prediction accuracy on transcription sites over the putative data sets. The experimental results also show the excellent performance of predictive deep neural network on the lincRNA data sets compared with support vector machine and traditional neural network. In addition, it is validated through the newly discovered lincRNA data set and one unreported transcription site is found by feeding the whole annotated sequences through the deep learning machine, which indicates that deep learning method has the extensive ability for lincRNA prediction. The transcriptional sequences of lincRNAs are collected from the annotated human DNA genome data. Subsequently, a two-layer deep neural network is developed for the lincRNA detection, which adopts the auto-encoder algorithm and utilizes different encoding schemes to obtain the best performance over intergenic DNA sequence data. Driven by those newly annotated lincRNA data, deep learning methods based on auto-encoder algorithm can exert their capability in knowledge learning in order to capture the useful features and the information correlation along DNA genome sequences for lincRNA detection. As our knowledge, this is the first application to adopt the deep learning techniques for identifying lincRNA transcription sequences.

Segmented-memory recurrent neural networks.

PubMed

Chen, Jinmiao; Chaudhari, Narendra S

2009-08-01

Conventional recurrent neural networks (RNNs) have difficulties in learning long-term dependencies. To tackle this problem, we propose an architecture called segmented-memory recurrent neural network (SMRNN). A symbolic sequence is broken into segments and then presented as inputs to the SMRNN one symbol per cycle. The SMRNN uses separate internal states to store symbol-level context, as well as segment-level context. The symbol-level context is updated for each symbol presented for input. The segment-level context is updated after each segment. The SMRNN is trained using an extended real-time recurrent learning algorithm. We test the performance of SMRNN on the information latching problem, the "two-sequence problem" and the problem of protein secondary structure (PSS) prediction. Our implementation results indicate that SMRNN performs better on long-term dependency problems than conventional RNNs. Besides, we also theoretically analyze how the segmented memory of SMRNN helps learning long-term temporal dependencies and study the impact of the segment length.

Control chart pattern recognition using RBF neural network with new training algorithm and practical features.

PubMed

Addeh, Abdoljalil; Khormali, Aminollah; Golilarz, Noorbakhsh Amiri

2018-05-04

The control chart patterns are the most commonly used statistical process control (SPC) tools to monitor process changes. When a control chart produces an out-of-control signal, this means that the process has been changed. In this study, a new method based on optimized radial basis function neural network (RBFNN) is proposed for control chart patterns (CCPs) recognition. The proposed method consists of four main modules: feature extraction, feature selection, classification and learning algorithm. In the feature extraction module, shape and statistical features are used. Recently, various shape and statistical features have been presented for the CCPs recognition. In the feature selection module, the association rules (AR) method has been employed to select the best set of the shape and statistical features. In the classifier section, RBFNN is used and finally, in RBFNN, learning algorithm has a high impact on the network performance. Therefore, a new learning algorithm based on the bees algorithm has been used in the learning module. Most studies have considered only six patterns: Normal, Cyclic, Increasing Trend, Decreasing Trend, Upward Shift and Downward Shift. Since three patterns namely Normal, Stratification, and Systematic are very similar to each other and distinguishing them is very difficult, in most studies Stratification and Systematic have not been considered. Regarding to the continuous monitoring and control over the production process and the exact type detection of the problem encountered during the production process, eight patterns have been investigated in this study. The proposed method is tested on a dataset containing 1600 samples (200 samples from each pattern) and the results showed that the proposed method has a very good performance. Copyright © 2018 ISA. Published by Elsevier Ltd. All rights reserved.

Video data compression using artificial neural network differential vector quantization

NASA Technical Reports Server (NTRS)

Krishnamurthy, Ashok K.; Bibyk, Steven B.; Ahalt, Stanley C.

1991-01-01

An artificial neural network vector quantizer is developed for use in data compression applications such as Digital Video. Differential Vector Quantization is used to preserve edge features, and a new adaptive algorithm, known as Frequency-Sensitive Competitive Learning, is used to develop the vector quantizer codebook. To develop real time performance, a custom Very Large Scale Integration Application Specific Integrated Circuit (VLSI ASIC) is being developed to realize the associative memory functions needed in the vector quantization algorithm. By using vector quantization, the need for Huffman coding can be eliminated, resulting in superior performance against channel bit errors than methods that use variable length codes.

Convolutional neural networks with balanced batches for facial expressions recognition

NASA Astrophysics Data System (ADS)

Battini Sönmez, Elena; Cangelosi, Angelo

2017-03-01

This paper considers the issue of fully automatic emotion classification on 2D faces. In spite of the great effort done in recent years, traditional machine learning approaches based on hand-crafted feature extraction followed by the classification stage failed to develop a real-time automatic facial expression recognition system. The proposed architecture uses Convolutional Neural Networks (CNN), which are built as a collection of interconnected processing elements to simulate the brain of human beings. The basic idea of CNNs is to learn a hierarchical representation of the input data, which results in a better classification performance. In this work we present a block-based CNN algorithm, which uses noise, as data augmentation technique, and builds batches with a balanced number of samples per class. The proposed architecture is a very simple yet powerful CNN, which can yield state-of-the-art accuracy on the very competitive benchmark algorithm of the Extended Cohn Kanade database.

Mutual information, neural networks and the renormalization group

NASA Astrophysics Data System (ADS)

Koch-Janusz, Maciej; Ringel, Zohar

2018-06-01

Physical systems differing in their microscopic details often display strikingly similar behaviour when probed at macroscopic scales. Those universal properties, largely determining their physical characteristics, are revealed by the powerful renormalization group (RG) procedure, which systematically retains `slow' degrees of freedom and integrates out the rest. However, the important degrees of freedom may be difficult to identify. Here we demonstrate a machine-learning algorithm capable of identifying the relevant degrees of freedom and executing RG steps iteratively without any prior knowledge about the system. We introduce an artificial neural network based on a model-independent, information-theoretic characterization of a real-space RG procedure, which performs this task. We apply the algorithm to classical statistical physics problems in one and two dimensions. We demonstrate RG flow and extract the Ising critical exponent. Our results demonstrate that machine-learning techniques can extract abstract physical concepts and consequently become an integral part of theory- and model-building.

Condition monitoring of 3G cellular networks through competitive neural models.

PubMed

Barreto, Guilherme A; Mota, João C M; Souza, Luis G M; Frota, Rewbenio A; Aguayo, Leonardo

2005-09-01

We develop an unsupervised approach to condition monitoring of cellular networks using competitive neural algorithms. Training is carried out with state vectors representing the normal functioning of a simulated CDMA2000 network. Once training is completed, global and local normality profiles (NPs) are built from the distribution of quantization errors of the training state vectors and their components, respectively. The global NP is used to evaluate the overall condition of the cellular system. If abnormal behavior is detected, local NPs are used in a component-wise fashion to find abnormal state variables. Anomaly detection tests are performed via percentile-based confidence intervals computed over the global and local NPs. We compared the performance of four competitive algorithms [winner-take-all (WTA), frequency-sensitive competitive learning (FSCL), self-organizing map (SOM), and neural-gas algorithm (NGA)] and the results suggest that the joint use of global and local NPs is more efficient and more robust than current single-threshold methods.

An Efficient Implementation of Deep Convolutional Neural Networks for MRI Segmentation.

PubMed

Hoseini, Farnaz; Shahbahrami, Asadollah; Bayat, Peyman

2018-02-27

Image segmentation is one of the most common steps in digital image processing, classifying a digital image into different segments. The main goal of this paper is to segment brain tumors in magnetic resonance images (MRI) using deep learning. Tumors having different shapes, sizes, brightness and textures can appear anywhere in the brain. These complexities are the reasons to choose a high-capacity Deep Convolutional Neural Network (DCNN) containing more than one layer. The proposed DCNN contains two parts: architecture and learning algorithms. The architecture and the learning algorithms are used to design a network model and to optimize parameters for the network training phase, respectively. The architecture contains five convolutional layers, all using 3 × 3 kernels, and one fully connected layer. Due to the advantage of using small kernels with fold, it allows making the effect of larger kernels with smaller number of parameters and fewer computations. Using the Dice Similarity Coefficient metric, we report accuracy results on the BRATS 2016, brain tumor segmentation challenge dataset, for the complete, core, and enhancing regions as 0.90, 0.85, and 0.84 respectively. The learning algorithm includes the task-level parallelism. All the pixels of an MR image are classified using a patch-based approach for segmentation. We attain a good performance and the experimental results show that the proposed DCNN increases the segmentation accuracy compared to previous techniques.

A meta-learning system based on genetic algorithms

NASA Astrophysics Data System (ADS)

Pellerin, Eric; Pigeon, Luc; Delisle, Sylvain

2004-04-01

The design of an efficient machine learning process through self-adaptation is a great challenge. The goal of meta-learning is to build a self-adaptive learning system that is constantly adapting to its specific (and dynamic) environment. To that end, the meta-learning mechanism must improve its bias dynamically by updating the current learning strategy in accordance with its available experiences or meta-knowledge. We suggest using genetic algorithms as the basis of an adaptive system. In this work, we propose a meta-learning system based on a combination of the a priori and a posteriori concepts. A priori refers to input information and knowledge available at the beginning in order to built and evolve one or more sets of parameters by exploiting the context of the system"s information. The self-learning component is based on genetic algorithms and neural Darwinism. A posteriori refers to the implicit knowledge discovered by estimation of the future states of parameters and is also applied to the finding of optimal parameters values. The in-progress research presented here suggests a framework for the discovery of knowledge that can support human experts in their intelligence information assessment tasks. The conclusion presents avenues for further research in genetic algorithms and their capability to learn to learn.

SpikeTemp: An Enhanced Rank-Order-Based Learning Approach for Spiking Neural Networks With Adaptive Structure.

PubMed

Wang, Jinling; Belatreche, Ammar; Maguire, Liam P; McGinnity, Thomas Martin

2017-01-01

This paper presents an enhanced rank-order-based learning algorithm, called SpikeTemp, for spiking neural networks (SNNs) with a dynamically adaptive structure. The trained feed-forward SNN consists of two layers of spiking neurons: 1) an encoding layer which temporally encodes real-valued features into spatio-temporal spike patterns and 2) an output layer of dynamically grown neurons which perform spatio-temporal classification. Both Gaussian receptive fields and square cosine population encoding schemes are employed to encode real-valued features into spatio-temporal spike patterns. Unlike the rank-order-based learning approach, SpikeTemp uses the precise times of the incoming spikes for adjusting the synaptic weights such that early spikes result in a large weight change and late spikes lead to a smaller weight change. This removes the need to rank all the incoming spikes and, thus, reduces the computational cost of SpikeTemp. The proposed SpikeTemp algorithm is demonstrated on several benchmark data sets and on an image recognition task. The results show that SpikeTemp can achieve better classification performance and is much faster than the existing rank-order-based learning approach. In addition, the number of output neurons is much smaller when the square cosine encoding scheme is employed. Furthermore, SpikeTemp is benchmarked against a selection of existing machine learning algorithms, and the results demonstrate the ability of SpikeTemp to classify different data sets after just one presentation of the training samples with comparable classification performance.

Component Pin Recognition Using Algorithms Based on Machine Learning

NASA Astrophysics Data System (ADS)

Xiao, Yang; Hu, Hong; Liu, Ze; Xu, Jiangchang

2018-04-01

The purpose of machine vision for a plug-in machine is to improve the machine’s stability and accuracy, and recognition of the component pin is an important part of the vision. This paper focuses on component pin recognition using three different techniques. The first technique involves traditional image processing using the core algorithm for binary large object (BLOB) analysis. The second technique uses the histogram of oriented gradients (HOG), to experimentally compare the effect of the support vector machine (SVM) and the adaptive boosting machine (AdaBoost) learning meta-algorithm classifiers. The third technique is the use of an in-depth learning method known as convolution neural network (CNN), which involves identifying the pin by comparing a sample to its training. The main purpose of the research presented in this paper is to increase the knowledge of learning methods used in the plug-in machine industry in order to achieve better results.

Phylogenetic convolutional neural networks in metagenomics.

PubMed

Fioravanti, Diego; Giarratano, Ylenia; Maggio, Valerio; Agostinelli, Claudio; Chierici, Marco; Jurman, Giuseppe; Furlanello, Cesare

2018-03-08

Convolutional Neural Networks can be effectively used only when data are endowed with an intrinsic concept of neighbourhood in the input space, as is the case of pixels in images. We introduce here Ph-CNN, a novel deep learning architecture for the classification of metagenomics data based on the Convolutional Neural Networks, with the patristic distance defined on the phylogenetic tree being used as the proximity measure. The patristic distance between variables is used together with a sparsified version of MultiDimensional Scaling to embed the phylogenetic tree in a Euclidean space. Ph-CNN is tested with a domain adaptation approach on synthetic data and on a metagenomics collection of gut microbiota of 38 healthy subjects and 222 Inflammatory Bowel Disease patients, divided in 6 subclasses. Classification performance is promising when compared to classical algorithms like Support Vector Machines and Random Forest and a baseline fully connected neural network, e.g. the Multi-Layer Perceptron. Ph-CNN represents a novel deep learning approach for the classification of metagenomics data. Operatively, the algorithm has been implemented as a custom Keras layer taking care of passing to the following convolutional layer not only the data but also the ranked list of neighbourhood of each sample, thus mimicking the case of image data, transparently to the user.

Improved neural network based scene-adaptive nonuniformity correction method for infrared focal plane arrays.

PubMed

Lai, Rui; Yang, Yin-tang; Zhou, Duan; Li, Yue-jin

2008-08-20

An improved scene-adaptive nonuniformity correction (NUC) algorithm for infrared focal plane arrays (IRFPAs) is proposed. This method simultaneously estimates the infrared detectors' parameters and eliminates the nonuniformity causing fixed pattern noise (FPN) by using a neural network (NN) approach. In the learning process of neuron parameter estimation, the traditional LMS algorithm is substituted with the newly presented variable step size (VSS) normalized least-mean square (NLMS) based adaptive filtering algorithm, which yields faster convergence, smaller misadjustment, and lower computational cost. In addition, a new NN structure is designed to estimate the desired target value, which promotes the calibration precision considerably. The proposed NUC method reaches high correction performance, which is validated by the experimental results quantitatively tested with a simulative testing sequence and a real infrared image sequence.

Random Bits Forest: a Strong Classifier/Regressor for Big Data

NASA Astrophysics Data System (ADS)

Wang, Yi; Li, Yi; Pu, Weilin; Wen, Kathryn; Shugart, Yin Yao; Xiong, Momiao; Jin, Li

2016-07-01

Efficiency, memory consumption, and robustness are common problems with many popular methods for data analysis. As a solution, we present Random Bits Forest (RBF), a classification and regression algorithm that integrates neural networks (for depth), boosting (for width), and random forests (for prediction accuracy). Through a gradient boosting scheme, it first generates and selects ~10,000 small, 3-layer random neural networks. These networks are then fed into a modified random forest algorithm to obtain predictions. Testing with datasets from the UCI (University of California, Irvine) Machine Learning Repository shows that RBF outperforms other popular methods in both accuracy and robustness, especially with large datasets (N > 1000). The algorithm also performed highly in testing with an independent data set, a real psoriasis genome-wide association study (GWAS).

A comparison of neural network and fuzzy clustering techniques in segmenting magnetic resonance images of the brain

NASA Technical Reports Server (NTRS)

Hall, Lawrence O.; Bensaid, Amine M.; Clarke, Laurence P.; Velthuizen, Robert P.; Silbiger, Martin S.; Bezdek, James C.

1992-01-01

Magnetic resonance (MR) brain section images are segmented and then synthetically colored to give visual representations of the original data with three approaches: the literal and approximate fuzzy c-means unsupervised clustering algorithms and a supervised computational neural network, a dynamic multilayered perception trained with the cascade correlation learning algorithm. Initial clinical results are presented on both normal volunteers and selected patients with brain tumors surrounded by edema. Supervised and unsupervised segmentation techniques provide broadly similar results. Unsupervised fuzzy algorithms were visually observed to show better segmentation when compared with raw image data for volunteer studies. However, for a more complex segmentation problem with tumor/edema or cerebrospinal fluid boundary, where the tissues have similar MR relaxation behavior, inconsistency in rating among experts was observed.

Using the ADAP Learning Algorithm to Forecast the Onset of Diabetes Mellitus

PubMed Central

Smith, Jack W.; Everhart, J.E.; Dickson, W.C.; Knowler, W.C.; Johannes, R.S.

1988-01-01

Neural networks or connectionist models for parallel processing are not new. However, a resurgence of interest in the past half decade has occurred. In part, this is related to a better understanding of what are now referred to as hidden nodes. These algorithms are considered to be of marked value in pattern recognition problems. Because of that, we tested the ability of an early neural network model, ADAP, to forecast the onset of diabetes mellitus in a high risk population of Pima Indians. The algorithm's performance was analyzed using standard measures for clinical tests: sensitivity, specificity, and a receiver operating characteristic curve. The crossover point for sensitivity and specificity is 0.76. We are currently further examining these methods by comparing the ADAP results with those obtained from logistic regression and linear perceptron models using precisely the same training and forecasting sets. A description of the algorithm is included.

Back-Propagation Operation for Analog Neural Network Hardware with Synapse Components Having Hysteresis Characteristics

PubMed Central

Ueda, Michihito; Nishitani, Yu; Kaneko, Yukihiro; Omote, Atsushi

2014-01-01

To realize an analog artificial neural network hardware, the circuit element for synapse function is important because the number of synapse elements is much larger than that of neuron elements. One of the candidates for this synapse element is a ferroelectric memristor. This device functions as a voltage controllable variable resistor, which can be applied to a synapse weight. However, its conductance shows hysteresis characteristics and dispersion to the input voltage. Therefore, the conductance values vary according to the history of the height and the width of the applied pulse voltage. Due to the difficulty of controlling the accurate conductance, it is not easy to apply the back-propagation learning algorithm to the neural network hardware having memristor synapses. To solve this problem, we proposed and simulated a learning operation procedure as follows. Employing a weight perturbation technique, we derived the error change. When the error reduced, the next pulse voltage was updated according to the back-propagation learning algorithm. If the error increased the amplitude of the next voltage pulse was set in such way as to cause similar memristor conductance but in the opposite voltage scanning direction. By this operation, we could eliminate the hysteresis and confirmed that the simulation of the learning operation converged. We also adopted conductance dispersion numerically in the simulation. We examined the probability that the error decreased to a designated value within a predetermined loop number. The ferroelectric has the characteristics that the magnitude of polarization does not become smaller when voltages having the same polarity are applied. These characteristics greatly improved the probability even if the learning rate was small, if the magnitude of the dispersion is adequate. Because the dispersion of analog circuit elements is inevitable, this learning operation procedure is useful for analog neural network hardware. PMID:25393715

Deep Learning and Its Applications in Biomedicine.

PubMed

Cao, Chensi; Liu, Feng; Tan, Hai; Song, Deshou; Shu, Wenjie; Li, Weizhong; Zhou, Yiming; Bo, Xiaochen; Xie, Zhi

2018-02-01

Advances in biological and medical technologies have been providing us explosive volumes of biological and physiological data, such as medical images, electroencephalography, genomic and protein sequences. Learning from these data facilitates the understanding of human health and disease. Developed from artificial neural networks, deep learning-based algorithms show great promise in extracting features and learning patterns from complex data. The aim of this paper is to provide an overview of deep learning techniques and some of the state-of-the-art applications in the biomedical field. We first introduce the development of artificial neural network and deep learning. We then describe two main components of deep learning, i.e., deep learning architectures and model optimization. Subsequently, some examples are demonstrated for deep learning applications, including medical image classification, genomic sequence analysis, as well as protein structure classification and prediction. Finally, we offer our perspectives for the future directions in the field of deep learning. Copyright © 2018. Production and hosting by Elsevier B.V.

DeepSite: protein-binding site predictor using 3D-convolutional neural networks.

PubMed

Jiménez, J; Doerr, S; Martínez-Rosell, G; Rose, A S; De Fabritiis, G

2017-10-01

An important step in structure-based drug design consists in the prediction of druggable binding sites. Several algorithms for detecting binding cavities, those likely to bind to a small drug compound, have been developed over the years by clever exploitation of geometric, chemical and evolutionary features of the protein. Here we present a novel knowledge-based approach that uses state-of-the-art convolutional neural networks, where the algorithm is learned by examples. In total, 7622 proteins from the scPDB database of binding sites have been evaluated using both a distance and a volumetric overlap approach. Our machine-learning based method demonstrates superior performance to two other competitive algorithmic strategies. DeepSite is freely available at www.playmolecule.org. Users can submit either a PDB ID or PDB file for pocket detection to our NVIDIA GPU-equipped servers through a WebGL graphical interface. gianni.defabritiis@upf.edu. Supplementary data are available at Bioinformatics online. © The Author (2017). Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Fast learning method for convolutional neural networks using extreme learning machine and its application to lane detection.

PubMed

Kim, Jihun; Kim, Jonghong; Jang, Gil-Jin; Lee, Minho

2017-03-01

Deep learning has received significant attention recently as a promising solution to many problems in the area of artificial intelligence. Among several deep learning architectures, convolutional neural networks (CNNs) demonstrate superior performance when compared to other machine learning methods in the applications of object detection and recognition. We use a CNN for image enhancement and the detection of driving lanes on motorways. In general, the process of lane detection consists of edge extraction and line detection. A CNN can be used to enhance the input images before lane detection by excluding noise and obstacles that are irrelevant to the edge detection result. However, training conventional CNNs requires considerable computation and a big dataset. Therefore, we suggest a new learning algorithm for CNNs using an extreme learning machine (ELM). The ELM is a fast learning method used to calculate network weights between output and hidden layers in a single iteration and thus, can dramatically reduce learning time while producing accurate results with minimal training data. A conventional ELM can be applied to networks with a single hidden layer; as such, we propose a stacked ELM architecture in the CNN framework. Further, we modify the backpropagation algorithm to find the targets of hidden layers and effectively learn network weights while maintaining performance. Experimental results confirm that the proposed method is effective in reducing learning time and improving performance. Copyright © 2016 Elsevier Ltd. All rights reserved.

A convolutional neural network-based screening tool for X-ray serial crystallography

PubMed Central

Ke, Tsung-Wei; Brewster, Aaron S.; Yu, Stella X.; Ushizima, Daniela; Yang, Chao; Sauter, Nicholas K.

2018-01-01

A new tool is introduced for screening macromolecular X-ray crystallography diffraction images produced at an X-ray free-electron laser light source. Based on a data-driven deep learning approach, the proposed tool executes a convolutional neural network to detect Bragg spots. Automatic image processing algorithms described can enable the classification of large data sets, acquired under realistic conditions consisting of noisy data with experimental artifacts. Outcomes are compared for different data regimes, including samples from multiple instruments and differing amounts of training data for neural network optimization. PMID:29714177

A convolutional neural network-based screening tool for X-ray serial crystallography.

PubMed

Ke, Tsung Wei; Brewster, Aaron S; Yu, Stella X; Ushizima, Daniela; Yang, Chao; Sauter, Nicholas K

2018-05-01

A new tool is introduced for screening macromolecular X-ray crystallography diffraction images produced at an X-ray free-electron laser light source. Based on a data-driven deep learning approach, the proposed tool executes a convolutional neural network to detect Bragg spots. Automatic image processing algorithms described can enable the classification of large data sets, acquired under realistic conditions consisting of noisy data with experimental artifacts. Outcomes are compared for different data regimes, including samples from multiple instruments and differing amounts of training data for neural network optimization. open access.

A convolutional neural network-based screening tool for X-ray serial crystallography

DOE PAGES

Ke, Tsung-Wei; Brewster, Aaron S.; Yu, Stella X.; ...

2018-04-24

A new tool is introduced for screening macromolecular X-ray crystallography diffraction images produced at an X-ray free-electron laser light source. Based on a data-driven deep learning approach, the proposed tool executes a convolutional neural network to detect Bragg spots. Automatic image processing algorithms described can enable the classification of large data sets, acquired under realistic conditions consisting of noisy data with experimental artifacts. Outcomes are compared for different data regimes, including samples from multiple instruments and differing amounts of training data for neural network optimization.

A convolutional neural network-based screening tool for X-ray serial crystallography

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

Ke, Tsung-Wei; Brewster, Aaron S.; Yu, Stella X.

A new tool is introduced for screening macromolecular X-ray crystallography diffraction images produced at an X-ray free-electron laser light source. Based on a data-driven deep learning approach, the proposed tool executes a convolutional neural network to detect Bragg spots. Automatic image processing algorithms described can enable the classification of large data sets, acquired under realistic conditions consisting of noisy data with experimental artifacts. Outcomes are compared for different data regimes, including samples from multiple instruments and differing amounts of training data for neural network optimization.

An Efficient Supervised Training Algorithm for Multilayer Spiking Neural Networks

PubMed Central

Xie, Xiurui; Qu, Hong; Liu, Guisong; Zhang, Malu; Kurths, Jürgen

2016-01-01

The spiking neural networks (SNNs) are the third generation of neural networks and perform remarkably well in cognitive tasks such as pattern recognition. The spike emitting and information processing mechanisms found in biological cognitive systems motivate the application of the hierarchical structure and temporal encoding mechanism in spiking neural networks, which have exhibited strong computational capability. However, the hierarchical structure and temporal encoding approach require neurons to process information serially in space and time respectively, which reduce the training efficiency significantly. For training the hierarchical SNNs, most existing methods are based on the traditional back-propagation algorithm, inheriting its drawbacks of the gradient diffusion and the sensitivity on parameters. To keep the powerful computation capability of the hierarchical structure and temporal encoding mechanism, but to overcome the low efficiency of the existing algorithms, a new training algorithm, the Normalized Spiking Error Back Propagation (NSEBP) is proposed in this paper. In the feedforward calculation, the output spike times are calculated by solving the quadratic function in the spike response model instead of detecting postsynaptic voltage states at all time points in traditional algorithms. Besides, in the feedback weight modification, the computational error is propagated to previous layers by the presynaptic spike jitter instead of the gradient decent rule, which realizes the layer-wised training. Furthermore, our algorithm investigates the mathematical relation between the weight variation and voltage error change, which makes the normalization in the weight modification applicable. Adopting these strategies, our algorithm outperforms the traditional SNN multi-layer algorithms in terms of learning efficiency and parameter sensitivity, that are also demonstrated by the comprehensive experimental results in this paper. PMID:27044001

Constrained Deep Weak Supervision for Histopathology Image Segmentation.

PubMed

Jia, Zhipeng; Huang, Xingyi; Chang, Eric I-Chao; Xu, Yan

2017-11-01

In this paper, we develop a new weakly supervised learning algorithm to learn to segment cancerous regions in histopathology images. This paper is under a multiple instance learning (MIL) framework with a new formulation, deep weak supervision (DWS); we also propose an effective way to introduce constraints to our neural networks to assist the learning process. The contributions of our algorithm are threefold: 1) we build an end-to-end learning system that segments cancerous regions with fully convolutional networks (FCNs) in which image-to-image weakly-supervised learning is performed; 2) we develop a DWS formulation to exploit multi-scale learning under weak supervision within FCNs; and 3) constraints about positive instances are introduced in our approach to effectively explore additional weakly supervised information that is easy to obtain and enjoy a significant boost to the learning process. The proposed algorithm, abbreviated as DWS-MIL, is easy to implement and can be trained efficiently. Our system demonstrates the state-of-the-art results on large-scale histopathology image data sets and can be applied to various applications in medical imaging beyond histopathology images, such as MRI, CT, and ultrasound images.

Applying Gradient Descent in Convolutional Neural Networks

NASA Astrophysics Data System (ADS)

Cui, Nan

2018-04-01

With the development of the integrated circuit and computer science, people become caring more about solving practical issues via information technologies. Along with that, a new subject called Artificial Intelligent (AI) comes up. One popular research interest of AI is about recognition algorithm. In this paper, one of the most common algorithms, Convolutional Neural Networks (CNNs) will be introduced, for image recognition. Understanding its theory and structure is of great significance for every scholar who is interested in this field. Convolution Neural Network is an artificial neural network which combines the mathematical method of convolution and neural network. The hieratical structure of CNN provides it reliable computer speed and reasonable error rate. The most significant characteristics of CNNs are feature extraction, weight sharing and dimension reduction. Meanwhile, combining with the Back Propagation (BP) mechanism and the Gradient Descent (GD) method, CNNs has the ability to self-study and in-depth learning. Basically, BP provides an opportunity for backwardfeedback for enhancing reliability and GD is used for self-training process. This paper mainly discusses the CNN and the related BP and GD algorithms, including the basic structure and function of CNN, details of each layer, the principles and features of BP and GD, and some examples in practice with a summary in the end.

Comparison of four machine learning algorithms for their applicability in satellite-based optical rainfall retrievals

NASA Astrophysics Data System (ADS)

Meyer, Hanna; Kühnlein, Meike; Appelhans, Tim; Nauss, Thomas

2016-03-01

Machine learning (ML) algorithms have successfully been demonstrated to be valuable tools in satellite-based rainfall retrievals which show the practicability of using ML algorithms when faced with high dimensional and complex data. Moreover, recent developments in parallel computing with ML present new possibilities for training and prediction speed and therefore make their usage in real-time systems feasible. This study compares four ML algorithms - random forests (RF), neural networks (NNET), averaged neural networks (AVNNET) and support vector machines (SVM) - for rainfall area detection and rainfall rate assignment using MSG SEVIRI data over Germany. Satellite-based proxies for cloud top height, cloud top temperature, cloud phase and cloud water path serve as predictor variables. The results indicate an overestimation of rainfall area delineation regardless of the ML algorithm (averaged bias = 1.8) but a high probability of detection ranging from 81% (SVM) to 85% (NNET). On a 24-hour basis, the performance of the rainfall rate assignment yielded R2 values between 0.39 (SVM) and 0.44 (AVNNET). Though the differences in the algorithms' performance were rather small, NNET and AVNNET were identified as the most suitable algorithms. On average, they demonstrated the best performance in rainfall area delineation as well as in rainfall rate assignment. NNET's computational speed is an additional advantage in work with large datasets such as in remote sensing based rainfall retrievals. However, since no single algorithm performed considerably better than the others we conclude that further research in providing suitable predictors for rainfall is of greater necessity than an optimization through the choice of the ML algorithm.

Unsupervised Learning of Overlapping Image Components Using Divisive Input Modulation

PubMed Central

Spratling, M. W.; De Meyer, K.; Kompass, R.

2009-01-01

This paper demonstrates that nonnegative matrix factorisation is mathematically related to a class of neural networks that employ negative feedback as a mechanism of competition. This observation inspires a novel learning algorithm which we call Divisive Input Modulation (DIM). The proposed algorithm provides a mathematically simple and computationally efficient method for the unsupervised learning of image components, even in conditions where these elementary features overlap considerably. To test the proposed algorithm, a novel artificial task is introduced which is similar to the frequently-used bars problem but employs squares rather than bars to increase the degree of overlap between components. Using this task, we investigate how the proposed method performs on the parsing of artificial images composed of overlapping features, given the correct representation of the individual components; and secondly, we investigate how well it can learn the elementary components from artificial training images. We compare the performance of the proposed algorithm with its predecessors including variations on these algorithms that have produced state-of-the-art performance on the bars problem. The proposed algorithm is more successful than its predecessors in dealing with overlap and occlusion in the artificial task that has been used to assess performance. PMID:19424442

Diagnosis and prediction of periodontally compromised teeth using a deep learning-based convolutional neural network algorithm.

PubMed

Lee, Jae-Hong; Kim, Do-Hyung; Jeong, Seong-Nyum; Choi, Seong-Ho

2018-04-01

The aim of the current study was to develop a computer-assisted detection system based on a deep convolutional neural network (CNN) algorithm and to evaluate the potential usefulness and accuracy of this system for the diagnosis and prediction of periodontally compromised teeth (PCT). Combining pretrained deep CNN architecture and a self-trained network, periapical radiographic images were used to determine the optimal CNN algorithm and weights. The diagnostic and predictive accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, area under the ROC curve, confusion matrix, and 95% confidence intervals (CIs) were calculated using our deep CNN algorithm, based on a Keras framework in Python. The periapical radiographic dataset was split into training (n=1,044), validation (n=348), and test (n=348) datasets. With the deep learning algorithm, the diagnostic accuracy for PCT was 81.0% for premolars and 76.7% for molars. Using 64 premolars and 64 molars that were clinically diagnosed as severe PCT, the accuracy of predicting extraction was 82.8% (95% CI, 70.1%-91.2%) for premolars and 73.4% (95% CI, 59.9%-84.0%) for molars. We demonstrated that the deep CNN algorithm was useful for assessing the diagnosis and predictability of PCT. Therefore, with further optimization of the PCT dataset and improvements in the algorithm, a computer-aided detection system can be expected to become an effective and efficient method of diagnosing and predicting PCT.

Comment on 'Deep convolutional neural network with transfer learning for rectum toxicity prediction in cervical cancer radiotherapy: a feasibility study'.

PubMed

Valdes, Gilmer; Interian, Yannet

2018-03-15

The application of machine learning (ML) presents tremendous opportunities for the field of oncology, thus we read 'Deep convolutional neural network with transfer learning for rectum toxicity prediction in cervical cancer radiotherapy: a feasibility study' with great interest. In this article, the authors used state of the art techniques: a pre-trained convolutional neural network (VGG-16 CNN), transfer learning, data augmentation, drop out and early stopping, all of which are directly responsible for the success and the excitement that these algorithms have created in other fields. We believe that the use of these techniques can offer tremendous opportunities in the field of Medical Physics and as such we would like to praise the authors for their pioneering application to the field of Radiation Oncology. That being said, given that the field of Medical Physics has unique characteristics that differentiate us from those fields where these techniques have been applied successfully, we would like to raise some points for future discussion and follow up studies that could help the community understand the limitations and nuances of deep learning techniques.

Training radial basis function networks for wind speed prediction using PSO enhanced differential search optimizer

PubMed Central

2018-01-01

This paper presents an integrated hybrid optimization algorithm for training the radial basis function neural network (RBF NN). Training of neural networks is still a challenging exercise in machine learning domain. Traditional training algorithms in general suffer and trap in local optima and lead to premature convergence, which makes them ineffective when applied for datasets with diverse features. Training algorithms based on evolutionary computations are becoming popular due to their robust nature in overcoming the drawbacks of the traditional algorithms. Accordingly, this paper proposes a hybrid training procedure with differential search (DS) algorithm functionally integrated with the particle swarm optimization (PSO). To surmount the local trapping of the search procedure, a new population initialization scheme is proposed using Logistic chaotic sequence, which enhances the population diversity and aid the search capability. To demonstrate the effectiveness of the proposed RBF hybrid training algorithm, experimental analysis on publicly available 7 benchmark datasets are performed. Subsequently, experiments were conducted on a practical application case for wind speed prediction to expound the superiority of the proposed RBF training algorithm in terms of prediction accuracy. PMID:29768463

Training radial basis function networks for wind speed prediction using PSO enhanced differential search optimizer.

PubMed

Rani R, Hannah Jessie; Victoire T, Aruldoss Albert

2018-01-01

This paper presents an integrated hybrid optimization algorithm for training the radial basis function neural network (RBF NN). Training of neural networks is still a challenging exercise in machine learning domain. Traditional training algorithms in general suffer and trap in local optima and lead to premature convergence, which makes them ineffective when applied for datasets with diverse features. Training algorithms based on evolutionary computations are becoming popular due to their robust nature in overcoming the drawbacks of the traditional algorithms. Accordingly, this paper proposes a hybrid training procedure with differential search (DS) algorithm functionally integrated with the particle swarm optimization (PSO). To surmount the local trapping of the search procedure, a new population initialization scheme is proposed using Logistic chaotic sequence, which enhances the population diversity and aid the search capability. To demonstrate the effectiveness of the proposed RBF hybrid training algorithm, experimental analysis on publicly available 7 benchmark datasets are performed. Subsequently, experiments were conducted on a practical application case for wind speed prediction to expound the superiority of the proposed RBF training algorithm in terms of prediction accuracy.

Cell segmentation in histopathological images with deep learning algorithms by utilizing spatial relationships.

PubMed

Hatipoglu, Nuh; Bilgin, Gokhan

2017-10-01

In many computerized methods for cell detection, segmentation, and classification in digital histopathology that have recently emerged, the task of cell segmentation remains a chief problem for image processing in designing computer-aided diagnosis (CAD) systems. In research and diagnostic studies on cancer, pathologists can use CAD systems as second readers to analyze high-resolution histopathological images. Since cell detection and segmentation are critical for cancer grade assessments, cellular and extracellular structures should primarily be extracted from histopathological images. In response, we sought to identify a useful cell segmentation approach with histopathological images that uses not only prominent deep learning algorithms (i.e., convolutional neural networks, stacked autoencoders, and deep belief networks), but also spatial relationships, information of which is critical for achieving better cell segmentation results. To that end, we collected cellular and extracellular samples from histopathological images by windowing in small patches with various sizes. In experiments, the segmentation accuracies of the methods used improved as the window sizes increased due to the addition of local spatial and contextual information. Once we compared the effects of training sample size and influence of window size, results revealed that the deep learning algorithms, especially convolutional neural networks and partly stacked autoencoders, performed better than conventional methods in cell segmentation.

Application of back-propagation artificial neural network (ANN) to predict crystallite size and band gap energy of ZnO quantum dots

NASA Astrophysics Data System (ADS)

Pelicano, Christian Mark; Rapadas, Nick; Cagatan, Gerard; Magdaluyo, Eduardo

2017-12-01

Herein, the crystallite size and band gap energy of zinc oxide (ZnO) quantum dots were predicted using artificial neural network (ANN). Three input factors including reagent ratio, growth time, and growth temperature were examined with respect to crystallite size and band gap energy as response factors. The generated results from neural network model were then compared with the experimental results. Experimental crystallite size and band gap energy of ZnO quantum dots were measured from TEM images and absorbance spectra, respectively. The Levenberg-Marquardt (LM) algorithm was used as the learning algorithm for the ANN model. The performance of the ANN model was then assessed through mean square error (MSE) and regression values. Based on the results, the ANN modelling results are in good agreement with the experimental data.

Vanishing points detection using combination of fast Hough transform and deep learning

NASA Astrophysics Data System (ADS)

Sheshkus, Alexander; Ingacheva, Anastasia; Nikolaev, Dmitry

2018-04-01

In this paper we propose a novel method for vanishing points detection based on convolutional neural network (CNN) approach and fast Hough transform algorithm. We show how to determine fast Hough transform neural network layer and how to use it in order to increase usability of the neural network approach to the vanishing point detection task. Our algorithm includes CNN with consequence of convolutional and fast Hough transform layers. We are building estimator for distribution of possible vanishing points in the image. This distribution can be used to find candidates of vanishing point. We provide experimental results from tests of suggested method using images collected from videos of road trips. Our approach shows stable result on test images with different projective distortions and noise. Described approach can be effectively implemented for mobile GPU and CPU.

Methods for reducing interference in the Complementary Learning Systems model: oscillating inhibition and autonomous memory rehearsal.

PubMed

Norman, Kenneth A; Newman, Ehren L; Perotte, Adler J

2005-11-01

The stability-plasticity problem (i.e. how the brain incorporates new information into its model of the world, while at the same time preserving existing knowledge) has been at the forefront of computational memory research for several decades. In this paper, we critically evaluate how well the Complementary Learning Systems theory of hippocampo-cortical interactions addresses the stability-plasticity problem. We identify two major challenges for the model: Finding a learning algorithm for cortex and hippocampus that enacts selective strengthening of weak memories, and selective punishment of competing memories; and preventing catastrophic forgetting in the case of non-stationary environments (i.e. when items are temporarily removed from the training set). We then discuss potential solutions to these problems: First, we describe a recently developed learning algorithm that leverages neural oscillations to find weak parts of memories (so they can be strengthened) and strong competitors (so they can be punished), and we show how this algorithm outperforms other learning algorithms (CPCA Hebbian learning and Leabra at memorizing overlapping patterns. Second, we describe how autonomous re-activation of memories (separately in cortex and hippocampus) during REM sleep, coupled with the oscillating learning algorithm, can reduce the rate of forgetting of input patterns that are no longer present in the environment. We then present a simple demonstration of how this process can prevent catastrophic interference in an AB-AC learning paradigm.

Analog hardware for delta-backpropagation neural networks

NASA Technical Reports Server (NTRS)

Eberhardt, Silvio P. (Inventor)

1992-01-01

This is a fully parallel analog backpropagation learning processor which comprises a plurality of programmable resistive memory elements serving as synapse connections whose values can be weighted during learning with buffer amplifiers, summing circuits, and sample-and-hold circuits arranged in a plurality of neuron layers in accordance with delta-backpropagation algorithms modified so as to control weight changes due to circuit drift.

Prediction of Air Pollutants Concentration Based on an Extreme Learning Machine: The Case of Hong Kong

PubMed Central

Zhang, Jiangshe; Ding, Weifu

2017-01-01

With the development of the economy and society all over the world, most metropolitan cities are experiencing elevated concentrations of ground-level air pollutants. It is urgent to predict and evaluate the concentration of air pollutants for some local environmental or health agencies. Feed-forward artificial neural networks have been widely used in the prediction of air pollutants concentration. However, there are some drawbacks, such as the low convergence rate and the local minimum. The extreme learning machine for single hidden layer feed-forward neural networks tends to provide good generalization performance at an extremely fast learning speed. The major sources of air pollutants in Hong Kong are mobile, stationary, and from trans-boundary sources. We propose predicting the concentration of air pollutants by the use of trained extreme learning machines based on the data obtained from eight air quality parameters in two monitoring stations, including Sham Shui Po and Tap Mun in Hong Kong for six years. The experimental results show that our proposed algorithm performs better on the Hong Kong data both quantitatively and qualitatively. Particularly, our algorithm shows better predictive ability, with R2 increased and root mean square error values decreased respectively. PMID:28125034

Fully automated, deep learning segmentation of oxygen-induced retinopathy images

PubMed Central

Xiao, Sa; Bucher, Felicitas; Wu, Yue; Rokem, Ariel; Lee, Cecilia S.; Marra, Kyle V.; Fallon, Regis; Diaz-Aguilar, Sophia; Aguilar, Edith; Friedlander, Martin; Lee, Aaron Y.

2017-01-01

Oxygen-induced retinopathy (OIR) is a widely used model to study ischemia-driven neovascularization (NV) in the retina and to serve in proof-of-concept studies in evaluating antiangiogenic drugs for ocular, as well as nonocular, diseases. The primary parameters that are analyzed in this mouse model include the percentage of retina with vaso-obliteration (VO) and NV areas. However, quantification of these two key variables comes with a great challenge due to the requirement of human experts to read the images. Human readers are costly, time-consuming, and subject to bias. Using recent advances in machine learning and computer vision, we trained deep learning neural networks using over a thousand segmentations to fully automate segmentation in OIR images. While determining the percentage area of VO, our algorithm achieved a similar range of correlation coefficients to that of expert inter-human correlation coefficients. In addition, our algorithm achieved a higher range of correlation coefficients compared with inter-expert correlation coefficients for quantification of the percentage area of neovascular tufts. In summary, we have created an open-source, fully automated pipeline for the quantification of key values of OIR images using deep learning neural networks. PMID:29263301

Photoacoustic image reconstruction via deep learning

NASA Astrophysics Data System (ADS)

Antholzer, Stephan; Haltmeier, Markus; Nuster, Robert; Schwab, Johannes

2018-02-01

Applying standard algorithms to sparse data problems in photoacoustic tomography (PAT) yields low-quality images containing severe under-sampling artifacts. To some extent, these artifacts can be reduced by iterative image reconstruction algorithms which allow to include prior knowledge such as smoothness, total variation (TV) or sparsity constraints. These algorithms tend to be time consuming as the forward and adjoint problems have to be solved repeatedly. Further, iterative algorithms have additional drawbacks. For example, the reconstruction quality strongly depends on a-priori model assumptions about the objects to be recovered, which are often not strictly satisfied in practical applications. To overcome these issues, in this paper, we develop direct and efficient reconstruction algorithms based on deep learning. As opposed to iterative algorithms, we apply a convolutional neural network, whose parameters are trained before the reconstruction process based on a set of training data. For actual image reconstruction, a single evaluation of the trained network yields the desired result. Our presented numerical results (using two different network architectures) demonstrate that the proposed deep learning approach reconstructs images with a quality comparable to state of the art iterative reconstruction methods.

Event-Driven Random Back-Propagation: Enabling Neuromorphic Deep Learning Machines

PubMed Central

Neftci, Emre O.; Augustine, Charles; Paul, Somnath; Detorakis, Georgios

2017-01-01

An ongoing challenge in neuromorphic computing is to devise general and computationally efficient models of inference and learning which are compatible with the spatial and temporal constraints of the brain. One increasingly popular and successful approach is to take inspiration from inference and learning algorithms used in deep neural networks. However, the workhorse of deep learning, the gradient descent Gradient Back Propagation (BP) rule, often relies on the immediate availability of network-wide information stored with high-precision memory during learning, and precise operations that are difficult to realize in neuromorphic hardware. Remarkably, recent work showed that exact backpropagated gradients are not essential for learning deep representations. Building on these results, we demonstrate an event-driven random BP (eRBP) rule that uses an error-modulated synaptic plasticity for learning deep representations. Using a two-compartment Leaky Integrate & Fire (I&F) neuron, the rule requires only one addition and two comparisons for each synaptic weight, making it very suitable for implementation in digital or mixed-signal neuromorphic hardware. Our results show that using eRBP, deep representations are rapidly learned, achieving classification accuracies on permutation invariant datasets comparable to those obtained in artificial neural network simulations on GPUs, while being robust to neural and synaptic state quantizations during learning. PMID:28680387

Event-Driven Random Back-Propagation: Enabling Neuromorphic Deep Learning Machines.

PubMed

Neftci, Emre O; Augustine, Charles; Paul, Somnath; Detorakis, Georgios

2017-01-01

An ongoing challenge in neuromorphic computing is to devise general and computationally efficient models of inference and learning which are compatible with the spatial and temporal constraints of the brain. One increasingly popular and successful approach is to take inspiration from inference and learning algorithms used in deep neural networks. However, the workhorse of deep learning, the gradient descent Gradient Back Propagation (BP) rule, often relies on the immediate availability of network-wide information stored with high-precision memory during learning, and precise operations that are difficult to realize in neuromorphic hardware. Remarkably, recent work showed that exact backpropagated gradients are not essential for learning deep representations. Building on these results, we demonstrate an event-driven random BP (eRBP) rule that uses an error-modulated synaptic plasticity for learning deep representations. Using a two-compartment Leaky Integrate & Fire (I&F) neuron, the rule requires only one addition and two comparisons for each synaptic weight, making it very suitable for implementation in digital or mixed-signal neuromorphic hardware. Our results show that using eRBP, deep representations are rapidly learned, achieving classification accuracies on permutation invariant datasets comparable to those obtained in artificial neural network simulations on GPUs, while being robust to neural and synaptic state quantizations during learning.

Feedback error learning control of magnetic satellites using type-2 fuzzy neural networks with elliptic membership functions.

PubMed

Khanesar, Mojtaba Ahmadieh; Kayacan, Erdal; Reyhanoglu, Mahmut; Kaynak, Okyay

2015-04-01

A novel type-2 fuzzy membership function (MF) in the form of an ellipse has recently been proposed in literature, the parameters of which that represent uncertainties are de-coupled from its parameters that determine the center and the support. This property has enabled the proposers to make an analytical comparison of the noise rejection capabilities of type-1 fuzzy logic systems with its type-2 counterparts. In this paper, a sliding mode control theory-based learning algorithm is proposed for an interval type-2 fuzzy logic system which benefits from elliptic type-2 fuzzy MFs. The learning is based on the feedback error learning method and not only the stability of the learning is proved but also the stability of the overall system is shown by adding an additional component to the control scheme to ensure robustness. In order to test the efficiency and efficacy of the proposed learning and the control algorithm, the trajectory tracking problem of a magnetic rigid spacecraft is studied. The simulations results show that the proposed control algorithm gives better performance results in terms of a smaller steady state error and a faster transient response as compared to conventional control algorithms.

Entity recognition in the biomedical domain using a hybrid approach.

PubMed

Basaldella, Marco; Furrer, Lenz; Tasso, Carlo; Rinaldi, Fabio

2017-11-09

This article describes a high-recall, high-precision approach for the extraction of biomedical entities from scientific articles. The approach uses a two-stage pipeline, combining a dictionary-based entity recognizer with a machine-learning classifier. First, the OGER entity recognizer, which has a bias towards high recall, annotates the terms that appear in selected domain ontologies. Subsequently, the Distiller framework uses this information as a feature for a machine learning algorithm to select the relevant entities only. For this step, we compare two different supervised machine-learning algorithms: Conditional Random Fields and Neural Networks. In an in-domain evaluation using the CRAFT corpus, we test the performance of the combined systems when recognizing chemicals, cell types, cellular components, biological processes, molecular functions, organisms, proteins, and biological sequences. Our best system combines dictionary-based candidate generation with Neural-Network-based filtering. It achieves an overall precision of 86% at a recall of 60% on the named entity recognition task, and a precision of 51% at a recall of 49% on the concept recognition task. These results are to our knowledge the best reported so far in this particular task.

Study of the convergence behavior of the complex kernel least mean square algorithm.

PubMed

Paul, Thomas K; Ogunfunmi, Tokunbo

2013-09-01

The complex kernel least mean square (CKLMS) algorithm is recently derived and allows for online kernel adaptive learning for complex data. Kernel adaptive methods can be used in finding solutions for neural network and machine learning applications. The derivation of CKLMS involved the development of a modified Wirtinger calculus for Hilbert spaces to obtain the cost function gradient. We analyze the convergence of the CKLMS with different kernel forms for complex data. The expressions obtained enable us to generate theory-predicted mean-square error curves considering the circularity of the complex input signals and their effect on nonlinear learning. Simulations are used for verifying the analysis results.

A new learning algorithm for a fully connected neuro-fuzzy inference system.

PubMed

Chen, C L Philip; Wang, Jing; Wang, Chi-Hsu; Chen, Long

2014-10-01

A traditional neuro-fuzzy system is transformed into an equivalent fully connected three layer neural network (NN), namely, the fully connected neuro-fuzzy inference systems (F-CONFIS). The F-CONFIS differs from traditional NNs by its dependent and repeated weights between input and hidden layers and can be considered as the variation of a kind of multilayer NN. Therefore, an efficient learning algorithm for the F-CONFIS to cope these repeated weights is derived. Furthermore, a dynamic learning rate is proposed for neuro-fuzzy systems via F-CONFIS where both premise (hidden) and consequent portions are considered. Several simulation results indicate that the proposed approach achieves much better accuracy and fast convergence.

Real-tiem Adaptive Control Scheme for Superior Plasma Confinement

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

Alexander Trunov, Ph.D.

2001-06-01

During this Phase I project, IOS, in collaboration with our subcontractors at General Atomics, Inc., acquired and analyzed measurement data on various plasma equilibrium modes. We developed a Matlab-based toolbox consisting of linear and neural network approximators that are capable of learning and predicting, with accuracy, the behavior of plasma parameters. We also began development of the control algorithm capable of using the model of the plasma obtained by the neural network approximator.

A Pruning Neural Network Model in Credit Classification Analysis

PubMed Central

Tang, Yajiao; Ji, Junkai; Dai, Hongwei; Yu, Yang; Todo, Yuki

2018-01-01

Nowadays, credit classification models are widely applied because they can help financial decision-makers to handle credit classification issues. Among them, artificial neural networks (ANNs) have been widely accepted as the convincing methods in the credit industry. In this paper, we propose a pruning neural network (PNN) and apply it to solve credit classification problem by adopting the well-known Australian and Japanese credit datasets. The model is inspired by synaptic nonlinearity of a dendritic tree in a biological neural model. And it is trained by an error back-propagation algorithm. The model is capable of realizing a neuronal pruning function by removing the superfluous synapses and useless dendrites and forms a tidy dendritic morphology at the end of learning. Furthermore, we utilize logic circuits (LCs) to simulate the dendritic structures successfully which makes PNN be implemented on the hardware effectively. The statistical results of our experiments have verified that PNN obtains superior performance in comparison with other classical algorithms in terms of accuracy and computational efficiency. PMID:29606961

Comparing deep neural network and other machine learning algorithms for stroke prediction in a large-scale population-based electronic medical claims database.

PubMed

Chen-Ying Hung; Wei-Chen Chen; Po-Tsun Lai; Ching-Heng Lin; Chi-Chun Lee

2017-07-01

Electronic medical claims (EMCs) can be used to accurately predict the occurrence of a variety of diseases, which can contribute to precise medical interventions. While there is a growing interest in the application of machine learning (ML) techniques to address clinical problems, the use of deep-learning in healthcare have just gained attention recently. Deep learning, such as deep neural network (DNN), has achieved impressive results in the areas of speech recognition, computer vision, and natural language processing in recent years. However, deep learning is often difficult to comprehend due to the complexities in its framework. Furthermore, this method has not yet been demonstrated to achieve a better performance comparing to other conventional ML algorithms in disease prediction tasks using EMCs. In this study, we utilize a large population-based EMC database of around 800,000 patients to compare DNN with three other ML approaches for predicting 5-year stroke occurrence. The result shows that DNN and gradient boosting decision tree (GBDT) can result in similarly high prediction accuracies that are better compared to logistic regression (LR) and support vector machine (SVM) approaches. Meanwhile, DNN achieves optimal results by using lesser amounts of patient data when comparing to GBDT method.

Machine learning algorithm accurately detects fMRI signature of vulnerability to major depression.

PubMed

Sato, João R; Moll, Jorge; Green, Sophie; Deakin, John F W; Thomaz, Carlos E; Zahn, Roland

2015-08-30

Standard functional magnetic resonance imaging (fMRI) analyses cannot assess the potential of a neuroimaging signature as a biomarker to predict individual vulnerability to major depression (MD). Here, we use machine learning for the first time to address this question. Using a recently identified neural signature of guilt-selective functional disconnection, the classification algorithm was able to distinguish remitted MD from control participants with 78.3% accuracy. This demonstrates the high potential of our fMRI signature as a biomarker of MD vulnerability. Crown Copyright © 2015. Published by Elsevier Ireland Ltd. All rights reserved.

Artificial Neural Network Approach in Laboratory Test Reporting:  Learning Algorithms.

PubMed

Demirci, Ferhat; Akan, Pinar; Kume, Tuncay; Sisman, Ali Riza; Erbayraktar, Zubeyde; Sevinc, Suleyman

2016-08-01

In the field of laboratory medicine, minimizing errors and establishing standardization is only possible by predefined processes. The aim of this study was to build an experimental decision algorithm model open to improvement that would efficiently and rapidly evaluate the results of biochemical tests with critical values by evaluating multiple factors concurrently. The experimental model was built by Weka software (Weka, Waikato, New Zealand) based on the artificial neural network method. Data were received from Dokuz Eylül University Central Laboratory. "Training sets" were developed for our experimental model to teach the evaluation criteria. After training the system, "test sets" developed for different conditions were used to statistically assess the validity of the model. After developing the decision algorithm with three iterations of training, no result was verified that was refused by the laboratory specialist. The sensitivity of the model was 91% and specificity was 100%. The estimated κ score was 0.950. This is the first study based on an artificial neural network to build an experimental assessment and decision algorithm model. By integrating our trained algorithm model into a laboratory information system, it may be possible to reduce employees' workload without compromising patient safety. © American Society for Clinical Pathology, 2016. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.

Towards autonomous neuroprosthetic control using Hebbian reinforcement learning.

PubMed

Mahmoudi, Babak; Pohlmeyer, Eric A; Prins, Noeline W; Geng, Shijia; Sanchez, Justin C

2013-12-01

Our goal was to design an adaptive neuroprosthetic controller that could learn the mapping from neural states to prosthetic actions and automatically adjust adaptation using only a binary evaluative feedback as a measure of desirability/undesirability of performance. Hebbian reinforcement learning (HRL) in a connectionist network was used for the design of the adaptive controller. The method combines the efficiency of supervised learning with the generality of reinforcement learning. The convergence properties of this approach were studied using both closed-loop control simulations and open-loop simulations that used primate neural data from robot-assisted reaching tasks. The HRL controller was able to perform classification and regression tasks using its episodic and sequential learning modes, respectively. In our experiments, the HRL controller quickly achieved convergence to an effective control policy, followed by robust performance. The controller also automatically stopped adapting the parameters after converging to a satisfactory control policy. Additionally, when the input neural vector was reorganized, the controller resumed adaptation to maintain performance. By estimating an evaluative feedback directly from the user, the HRL control algorithm may provide an efficient method for autonomous adaptation of neuroprosthetic systems. This method may enable the user to teach the controller the desired behavior using only a simple feedback signal.

A Novel Extreme Learning Machine Classification Model for e-Nose Application Based on the Multiple Kernel Approach.

PubMed

Jian, Yulin; Huang, Daoyu; Yan, Jia; Lu, Kun; Huang, Ying; Wen, Tailai; Zeng, Tanyue; Zhong, Shijie; Xie, Qilong

2017-06-19

A novel classification model, named the quantum-behaved particle swarm optimization (QPSO)-based weighted multiple kernel extreme learning machine (QWMK-ELM), is proposed in this paper. Experimental validation is carried out with two different electronic nose (e-nose) datasets. Being different from the existing multiple kernel extreme learning machine (MK-ELM) algorithms, the combination coefficients of base kernels are regarded as external parameters of single-hidden layer feedforward neural networks (SLFNs). The combination coefficients of base kernels, the model parameters of each base kernel, and the regularization parameter are optimized by QPSO simultaneously before implementing the kernel extreme learning machine (KELM) with the composite kernel function. Four types of common single kernel functions (Gaussian kernel, polynomial kernel, sigmoid kernel, and wavelet kernel) are utilized to constitute different composite kernel functions. Moreover, the method is also compared with other existing classification methods: extreme learning machine (ELM), kernel extreme learning machine (KELM), k-nearest neighbors (KNN), support vector machine (SVM), multi-layer perceptron (MLP), radical basis function neural network (RBFNN), and probabilistic neural network (PNN). The results have demonstrated that the proposed QWMK-ELM outperforms the aforementioned methods, not only in precision, but also in efficiency for gas classification.

The fatigue life prediction of aluminium alloy using genetic algorithm and neural network

NASA Astrophysics Data System (ADS)

Susmikanti, Mike

2013-09-01

The behavior of the fatigue life of the industrial materials is very important. In many cases, the material with experiencing fatigue life cannot be avoided, however, there are many ways to control their behavior. Many investigations of the fatigue life phenomena of alloys have been done, but it is high cost and times consuming computation. This paper report the modeling and simulation approaches to predict the fatigue life behavior of Aluminum Alloys and resolves some problems of computation. First, the simulation using genetic algorithm was utilized to optimize the load to obtain the stress values. These results can be used to provide N-cycle fatigue life of the material. Furthermore, the experimental data was applied as input data in the neural network learning, while the samples data were applied for testing of the training data. Finally, the multilayer perceptron algorithm is applied to predict whether the given data sets in accordance with the fatigue life of the alloy. To achieve rapid convergence, the Levenberg-Marquardt algorithm was also employed. The simulations results shows that the fatigue behaviors of aluminum under pressure can be predicted. In addition, implementation of neural networks successfully identified a model for material fatigue life.

Spatial and temporal multiyear sea ice distributions in the Arctic: A neural network analysis of SSM/I data, 1988-2001

USGS Publications Warehouse

Belchansky, G.I.; Douglas, David C.; Alpatsky, I.V.; Platonov, Nikita G.

2004-01-01

Arctic multiyear sea ice concentration maps for January 1988-2001 were generated from SSM/I brightness temperatures (19H, 19V, and 37V) using modified multiple layer perceptron neural networks. Learning data for the neural networks were extracted from ice maps derived from Okean and ERS satellite imagery to capitalize on the stability of active radar multiyear ice signatures. Evaluations of three learning algorithms and several topologies indicated that networks constructed with error back propagation learning and 3-20-1 topology produced the most consistent and physically plausible results. Operational neural networks were developed specifically with January learning data, and then used to estimate daily multiyear ice concentrations from daily-averaged SSM/I brightness temperatures during January. Monthly mean maps were produced for analysis by averaging the respective daily estimates. The 14-year series of January multiyear ice distributions revealed dense and persistent cover in the central Arctic surrounded by expansive regions of highly fluctuating interannual cover. Estimates of total multiyear ice area by the neural network were intermediate to those of other passive microwave algorithms, but annual fluctuations and trends were similar among all algorithms. When compared to Radarsat estimates of multiyear ice concentration in the Beaufort and Chukchi Seas (1997-1999), average discrepancies were small (0.9-2.5%) and spatial coherency was reasonable, indicating the neural network's Okean and ERS learning data facilitated passive microwave inversion that emulated backscatter signatures. During 1988-2001, total January multiyear ice area declined at a significant linear rate of -54.3 x 103 km2/yr-1 (-1.4%/yr-1). The most persistent and extensive decline in multiyear ice concentration (-3.3%/yr-1) occurred in the southern Beaufort and Chukchi Seas. In autumn 1996, a large multiyear ice recruitment of over 106 km2 (mostly in the Siberian Arctic) fully replenished the previous 8-year decline in total area, but it was followed by an accelerated and compensatory decline during the subsequent 4 years. Seventy-five percent of the interannual variation in January multiyear sea ice area was explained by linear regression on two atmospheric parameters: the previous inter's (JFM) Arctic Oscillation index as a proxy to melt duration and the previous year's average sea level pressure gradient across the Fram Strait as a proxy to annual ice export. Consecutive year changes (1994-2001) in January multiyear ice volume were significantly correlated with duration of the intervening melt season (R2 = 0.73, -80.0 km3/d-1), emphasizing a large thermodynamic influence on the Arctic's mass sea ice balance during summers with anomalous melt durations.

Neural networks and applications tutorial

NASA Astrophysics Data System (ADS)

Guyon, I.

1991-09-01

The importance of neural networks has grown dramatically during this decade. While only a few years ago they were primarily of academic interest, now dozens of companies and many universities are investigating the potential use of these systems and products are beginning to appear. The idea of building a machine whose architecture is inspired by that of the brain has roots which go far back in history. Nowadays, technological advances of computers and the availability of custom integrated circuits, permit simulations of hundreds or even thousands of neurons. In conjunction, the growing interest in learning machines, non-linear dynamics and parallel computation spurred renewed attention in artificial neural networks. Many tentative applications have been proposed, including decision systems (associative memories, classifiers, data compressors and optimizers), or parametric models for signal processing purposes (system identification, automatic control, noise canceling, etc.). While they do not always outperform standard methods, neural network approaches are already used in some real world applications for pattern recognition and signal processing tasks. The tutorial is divided into six lectures, that where presented at the Third Graduate Summer Course on Computational Physics (September 3-7, 1990) on Parallel Architectures and Applications, organized by the European Physical Society: (1) Introduction: machine learning and biological computation. (2) Adaptive artificial neurons (perceptron, ADALINE, sigmoid units, etc.): learning rules and implementations. (3) Neural network systems: architectures, learning algorithms. (4) Applications: pattern recognition, signal processing, etc. (5) Elements of learning theory: how to build networks which generalize. (6) A case study: a neural network for on-line recognition of handwritten alphanumeric characters.

Adaptive fuzzy leader clustering of complex data sets in pattern recognition

NASA Technical Reports Server (NTRS)

Newton, Scott C.; Pemmaraju, Surya; Mitra, Sunanda

1992-01-01

A modular, unsupervised neural network architecture for clustering and classification of complex data sets is presented. The adaptive fuzzy leader clustering (AFLC) architecture is a hybrid neural-fuzzy system that learns on-line in a stable and efficient manner. The initial classification is performed in two stages: a simple competitive stage and a distance metric comparison stage. The cluster prototypes are then incrementally updated by relocating the centroid positions from fuzzy C-means system equations for the centroids and the membership values. The AFLC algorithm is applied to the Anderson Iris data and laser-luminescent fingerprint image data. It is concluded that the AFLC algorithm successfully classifies features extracted from real data, discrete or continuous.

APPLICATION OF NEURAL NETWORK ALGORITHMS FOR BPM LINEARIZATION

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

Musson, John C.; Seaton, Chad; Spata, Mike F.

2012-11-01

Stripline BPM sensors contain inherent non-linearities, as a result of field distortions from the pickup elements. Many methods have been devised to facilitate corrections, often employing polynomial fitting. The cost of computation makes real-time correction difficult, particulalry when integer math is utilized. The application of neural-network technology, particularly the multi-layer perceptron algorithm, is proposed as an efficient alternative for electrode linearization. A process of supervised learning is initially used to determine the weighting coefficients, which are subsequently applied to the incoming electrode data. A non-linear layer, known as an activation layer, is responsible for the removal of saturation effects. Implementationmore » of a perceptron in an FPGA-based software-defined radio (SDR) is presented, along with performance comparisons. In addition, efficient calculation of the sigmoidal activation function via the CORDIC algorithm is presented.« less

A recurrent self-organizing neural fuzzy inference network.

PubMed

Juang, C F; Lin, C T

1999-01-01

A recurrent self-organizing neural fuzzy inference network (RSONFIN) is proposed in this paper. The RSONFIN is inherently a recurrent multilayered connectionist network for realizing the basic elements and functions of dynamic fuzzy inference, and may be considered to be constructed from a series of dynamic fuzzy rules. The temporal relations embedded in the network are built by adding some feedback connections representing the memory elements to a feedforward neural fuzzy network. Each weight as well as node in the RSONFIN has its own meaning and represents a special element in a fuzzy rule. There are no hidden nodes (i.e., no membership functions and fuzzy rules) initially in the RSONFIN. They are created on-line via concurrent structure identification (the construction of dynamic fuzzy if-then rules) and parameter identification (the tuning of the free parameters of membership functions). The structure learning together with the parameter learning forms a fast learning algorithm for building a small, yet powerful, dynamic neural fuzzy network. Two major characteristics of the RSONFIN can thus be seen: 1) the recurrent property of the RSONFIN makes it suitable for dealing with temporal problems and 2) no predetermination, like the number of hidden nodes, must be given, since the RSONFIN can find its optimal structure and parameters automatically and quickly. Moreover, to reduce the number of fuzzy rules generated, a flexible input partition method, the aligned clustering-based algorithm, is proposed. Various simulations on temporal problems are done and performance comparisons with some existing recurrent networks are also made. Efficiency of the RSONFIN is verified from these results.

Mastering the game of Go with deep neural networks and tree search

NASA Astrophysics Data System (ADS)

Silver, David; Huang, Aja; Maddison, Chris J.; Guez, Arthur; Sifre, Laurent; van den Driessche, George; Schrittwieser, Julian; Antonoglou, Ioannis; Panneershelvam, Veda; Lanctot, Marc; Dieleman, Sander; Grewe, Dominik; Nham, John; Kalchbrenner, Nal; Sutskever, Ilya; Lillicrap, Timothy; Leach, Madeleine; Kavukcuoglu, Koray; Graepel, Thore; Hassabis, Demis

2016-01-01

The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves. Here we introduce a new approach to computer Go that uses ‘value networks’ to evaluate board positions and ‘policy networks’ to select moves. These deep neural networks are trained by a novel combination of supervised learning from human expert games, and reinforcement learning from games of self-play. Without any lookahead search, the neural networks play Go at the level of state-of-the-art Monte Carlo tree search programs that simulate thousands of random games of self-play. We also introduce a new search algorithm that combines Monte Carlo simulation with value and policy networks. Using this search algorithm, our program AlphaGo achieved a 99.8% winning rate against other Go programs, and defeated the human European Go champion by 5 games to 0. This is the first time that a computer program has defeated a human professional player in the full-sized game of Go, a feat previously thought to be at least a decade away.

Mastering the game of Go with deep neural networks and tree search.

PubMed

Silver, David; Huang, Aja; Maddison, Chris J; Guez, Arthur; Sifre, Laurent; van den Driessche, George; Schrittwieser, Julian; Antonoglou, Ioannis; Panneershelvam, Veda; Lanctot, Marc; Dieleman, Sander; Grewe, Dominik; Nham, John; Kalchbrenner, Nal; Sutskever, Ilya; Lillicrap, Timothy; Leach, Madeleine; Kavukcuoglu, Koray; Graepel, Thore; Hassabis, Demis

2016-01-28

The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves. Here we introduce a new approach to computer Go that uses 'value networks' to evaluate board positions and 'policy networks' to select moves. These deep neural networks are trained by a novel combination of supervised learning from human expert games, and reinforcement learning from games of self-play. Without any lookahead search, the neural networks play Go at the level of state-of-the-art Monte Carlo tree search programs that simulate thousands of random games of self-play. We also introduce a new search algorithm that combines Monte Carlo simulation with value and policy networks. Using this search algorithm, our program AlphaGo achieved a 99.8% winning rate against other Go programs, and defeated the human European Go champion by 5 games to 0. This is the first time that a computer program has defeated a human professional player in the full-sized game of Go, a feat previously thought to be at least a decade away.

Medical subdomain classification of clinical notes using a machine learning-based natural language processing approach.

PubMed

Weng, Wei-Hung; Wagholikar, Kavishwar B; McCray, Alexa T; Szolovits, Peter; Chueh, Henry C

2017-12-01

The medical subdomain of a clinical note, such as cardiology or neurology, is useful content-derived metadata for developing machine learning downstream applications. To classify the medical subdomain of a note accurately, we have constructed a machine learning-based natural language processing (NLP) pipeline and developed medical subdomain classifiers based on the content of the note. We constructed the pipeline using the clinical NLP system, clinical Text Analysis and Knowledge Extraction System (cTAKES), the Unified Medical Language System (UMLS) Metathesaurus, Semantic Network, and learning algorithms to extract features from two datasets - clinical notes from Integrating Data for Analysis, Anonymization, and Sharing (iDASH) data repository (n = 431) and Massachusetts General Hospital (MGH) (n = 91,237), and built medical subdomain classifiers with different combinations of data representation methods and supervised learning algorithms. We evaluated the performance of classifiers and their portability across the two datasets. The convolutional recurrent neural network with neural word embeddings trained-medical subdomain classifier yielded the best performance measurement on iDASH and MGH datasets with area under receiver operating characteristic curve (AUC) of 0.975 and 0.991, and F1 scores of 0.845 and 0.870, respectively. Considering better clinical interpretability, linear support vector machine-trained medical subdomain classifier using hybrid bag-of-words and clinically relevant UMLS concepts as the feature representation, with term frequency-inverse document frequency (tf-idf)-weighting, outperformed other shallow learning classifiers on iDASH and MGH datasets with AUC of 0.957 and 0.964, and F1 scores of 0.932 and 0.934 respectively. We trained classifiers on one dataset, applied to the other dataset and yielded the threshold of F1 score of 0.7 in classifiers for half of the medical subdomains we studied. Our study shows that a supervised learning-based NLP approach is useful to develop medical subdomain classifiers. The deep learning algorithm with distributed word representation yields better performance yet shallow learning algorithms with the word and concept representation achieves comparable performance with better clinical interpretability. Portable classifiers may also be used across datasets from different institutions.

A smart-pixel holographic competitive learning network

NASA Astrophysics Data System (ADS)

Slagle, Timothy Michael

Neural networks are adaptive classifiers which modify their decision boundaries based on feedback from externally- or internally-generated error signals. Optics is an attractive technology for neural network implementation because it offers the possibility of parallel, nearly instantaneous computation of the weighted neuron inputs by the propagation of light through the optical system. Using current optical device technology, system performance levels of 3 × 1011 connection updates per second can be achieved. This thesis presents an architecture for an optical competitive learning network which offers advantages over previous optical implementations, including smart-pixel-based optical neurons, phase- conjugate self-alignment of a single neuron plane, and high-density, parallel-access weight storage, interconnection, and learning in a volume hologram. The competitive learning algorithm with modifications for optical implementation is described, and algorithm simulations are performed for an example problem. The optical competitive learning architecture is then introduced. The optical system is simulated using the ``beamprop'' algorithm at the level of light propagating through the system components, and results showing competitive learning operation in agreement with the algorithm simulations are presented. The optical competitive learning requires a non-linear, non-local ``winner-take-all'' (WTA) neuron function. Custom-designed smart-pixel WTA neuron arrays were fabricated using CMOS VLSI/liquid crystal technology. Results of laboratory tests of the WTA arrays' switching characteristics, time response, and uniformity are then presented. The system uses a phase-conjugate mirror to write the self-aligning interconnection weight holograms, and energy gain is required from the reflection to minimize erasure of the existing weights. An experimental system for characterizing the PCM response is described. Useful gains of 20 were obtained with a polarization-multiplexed PCM readout, and gains of up to 60 were observed when a time-sequential read-out technique was used. Finally, the optical competitive learning laboratory system is described, including some necessary modifications to the previous architectures, and the data acquisition and control system developed for the system. Experimental results showing phase conjugation of the WTA outputs, holographic interconnect storage, associative storage between input images and WTA neuron outputs, and WTA array switching are presented, demonstrating the functions necessary for the operation of the optical learning system.

Multi-Objective Reinforcement Learning-based Deep Neural Networks for Cognitive Space Communications

NASA Technical Reports Server (NTRS)

Ferreria, Paulo; Paffenroth, Randy; Wyglinski, Alexander M.; Hackett, Timothy; Bilen, Sven; Reinhart, Richard; Mortensen, Dale

2017-01-01

Future communication subsystems of space exploration missions can potentially benefit from software-defined radios (SDRs) controlled by machine learning algorithms. In this paper, we propose a novel hybrid radio resource allocation management control algorithm that integrates multi-objective reinforcement learning and deep artificial neural networks. The objective is to efficiently manage communications system resources by monitoring performance functions with common dependent variables that result in conflicting goals. The uncertainty in the performance of thousands of different possible combinations of radio parameters makes the trade-off between exploration and exploitation in reinforcement learning (RL) much more challenging for future critical space-based missions. Thus, the system should spend as little time as possible on exploring actions, and whenever it explores an action, it should perform at acceptable levels most of the time. The proposed approach enables on-line learning by interactions with the environment and restricts poor resource allocation performance through virtual environment exploration. Improvements in the multiobjective performance can be achieved via transmitter parameter adaptation on a packet-basis, with poorly predicted performance promptly resulting in rejected decisions. Simulations presented in this work considered the DVB-S2 standard adaptive transmitter parameters and additional ones expected to be present in future adaptive radio systems. Performance results are provided by analysis of the proposed hybrid algorithm when operating across a satellite communication channel from Earth to GEO orbit during clear sky conditions. The proposed approach constitutes part of the core cognitive engine proof-of-concept to be delivered to the NASA Glenn Research Center SCaN Testbed located onboard the International Space Station.

Multi-Objective Reinforcement Learning-Based Deep Neural Networks for Cognitive Space Communications

NASA Technical Reports Server (NTRS)

Ferreria, Paulo Victor R.; Paffenroth, Randy; Wyglinski, Alexander M.; Hackett, Timothy M.; Bilen, Sven G.; Reinhart, Richard C.; Mortensen, Dale J.

2017-01-01

Future communication subsystems of space exploration missions can potentially benefit from software-defined radios (SDRs) controlled by machine learning algorithms. In this paper, we propose a novel hybrid radio resource allocation management control algorithm that integrates multi-objective reinforcement learning and deep artificial neural networks. The objective is to efficiently manage communications system resources by monitoring performance functions with common dependent variables that result in conflicting goals. The uncertainty in the performance of thousands of different possible combinations of radio parameters makes the trade-off between exploration and exploitation in reinforcement learning (RL) much more challenging for future critical space-based missions. Thus, the system should spend as little time as possible on exploring actions, and whenever it explores an action, it should perform at acceptable levels most of the time. The proposed approach enables on-line learning by interactions with the environment and restricts poor resource allocation performance through virtual environment exploration. Improvements in the multiobjective performance can be achieved via transmitter parameter adaptation on a packet-basis, with poorly predicted performance promptly resulting in rejected decisions. Simulations presented in this work considered the DVB-S2 standard adaptive transmitter parameters and additional ones expected to be present in future adaptive radio systems. Performance results are provided by analysis of the proposed hybrid algorithm when operating across a satellite communication channel from Earth to GEO orbit during clear sky conditions. The proposed approach constitutes part of the core cognitive engine proof-of-concept to be delivered to the NASA Glenn Research Center SCaN Testbed located onboard the International Space Station.

A reductionist approach to the analysis of learning in brain-computer interfaces.

PubMed

Danziger, Zachary

2014-04-01

The complexity and scale of brain-computer interface (BCI) studies limit our ability to investigate how humans learn to use BCI systems. It also limits our capacity to develop adaptive algorithms needed to assist users with their control. Adaptive algorithm development is forced offline and typically uses static data sets. But this is a poor substitute for the online, dynamic environment where algorithms are ultimately deployed and interact with an adapting user. This work evaluates a paradigm that simulates the control problem faced by human subjects when controlling a BCI, but which avoids the many complications associated with full-scale BCI studies. Biological learners can be studied in a reductionist way as they solve BCI-like control problems, and machine learning algorithms can be developed and tested in closed loop with the subjects before being translated to full BCIs. The method is to map 19 joint angles of the hand (representing neural signals) to the position of a 2D cursor which must be piloted to displayed targets (a typical BCI task). An investigation is presented on how closely the joint angle method emulates BCI systems; a novel learning algorithm is evaluated, and a performance difference between genders is discussed.

Toward an Improvement of the Analysis of Neural Coding.

PubMed

Alegre-Cortés, Javier; Soto-Sánchez, Cristina; Albarracín, Ana L; Farfán, Fernando D; Val-Calvo, Mikel; Ferrandez, José M; Fernandez, Eduardo

2017-01-01

Machine learning and artificial intelligence have strong roots on principles of neural computation. Some examples are the structure of the first perceptron, inspired in the retina, neuroprosthetics based on ganglion cell recordings or Hopfield networks. In addition, machine learning provides a powerful set of tools to analyze neural data, which has already proved its efficacy in so distant fields of research as speech recognition, behavioral states classification, or LFP recordings. However, despite the huge technological advances in neural data reduction of dimensionality, pattern selection, and clustering during the last years, there has not been a proportional development of the analytical tools used for Time-Frequency (T-F) analysis in neuroscience. Bearing this in mind, we introduce the convenience of using non-linear, non-stationary tools, EMD algorithms in particular, for the transformation of the oscillatory neural data (EEG, EMG, spike oscillations…) into the T-F domain prior to its analysis with machine learning tools. We support that to achieve meaningful conclusions, the transformed data we analyze has to be as faithful as possible to the original recording, so that the transformations forced into the data due to restrictions in the T-F computation are not extended to the results of the machine learning analysis. Moreover, bioinspired computation such as brain-machine interface may be enriched from a more precise definition of neuronal coding where non-linearities of the neuronal dynamics are considered.

A new neural observer for an anaerobic bioreactor.

PubMed

Belmonte-Izquierdo, R; Carlos-Hernandez, S; Sanchez, E N

2010-02-01

In this paper, a recurrent high order neural observer (RHONO) for anaerobic processes is proposed. The main objective is to estimate variables of methanogenesis: biomass, substrate and inorganic carbon in a completely stirred tank reactor (CSTR). The recurrent high order neural network (RHONN) structure is based on the hyperbolic tangent as activation function. The learning algorithm is based on an extended Kalman filter (EKF). The applicability of the proposed scheme is illustrated via simulation. A validation using real data from a lab scale process is included. Thus, this observer can be successfully implemented for control purposes.

Proceedings of the Third International Workshop on Neural Networks and Fuzzy Logic, volume 2

NASA Technical Reports Server (NTRS)

Culbert, Christopher J. (Editor)

1993-01-01

Papers presented at the Neural Networks and Fuzzy Logic Workshop sponsored by the National Aeronautics and Space Administration and cosponsored by the University of Houston, Clear Lake, held 1-3 Jun. 1992 at the Lyndon B. Johnson Space Center in Houston, Texas are included. During the three days approximately 50 papers were presented. Technical topics addressed included adaptive systems; learning algorithms; network architectures; vision; robotics; neurobiological connections; speech recognition and synthesis; fuzzy set theory and application, control and dynamics processing; space applications; fuzzy logic and neural network computers; approximate reasoning; and multiobject decision making.

Target recognition based on convolutional neural network

NASA Astrophysics Data System (ADS)

Wang, Liqiang; Wang, Xin; Xi, Fubiao; Dong, Jian

2017-11-01

One of the important part of object target recognition is the feature extraction, which can be classified into feature extraction and automatic feature extraction. The traditional neural network is one of the automatic feature extraction methods, while it causes high possibility of over-fitting due to the global connection. The deep learning algorithm used in this paper is a hierarchical automatic feature extraction method, trained with the layer-by-layer convolutional neural network (CNN), which can extract the features from lower layers to higher layers. The features are more discriminative and it is beneficial to the object target recognition.

Analysis Resilient Algorithm on Artificial Neural Network Backpropagation

NASA Astrophysics Data System (ADS)

Saputra, Widodo; Tulus; Zarlis, Muhammad; Widia Sembiring, Rahmat; Hartama, Dedy

2017-12-01

Prediction required by decision makers to anticipate future planning. Artificial Neural Network (ANN) Backpropagation is one of method. This method however still has weakness, for long training time. This is a reason to improve a method to accelerate the training. One of Artificial Neural Network (ANN) Backpropagation method is a resilient method. Resilient method of changing weights and bias network with direct adaptation process of weighting based on local gradient information from every learning iteration. Predicting data result of Istanbul Stock Exchange training getting better. Mean Square Error (MSE) value is getting smaller and increasing accuracy.

GA-based fuzzy reinforcement learning for control of a magnetic bearing system.

PubMed

Lin, C T; Jou, C P

2000-01-01

This paper proposes a TD (temporal difference) and GA (genetic algorithm)-based reinforcement (TDGAR) learning method and applies it to the control of a real magnetic bearing system. The TDGAR learning scheme is a new hybrid GA, which integrates the TD prediction method and the GA to perform the reinforcement learning task. The TDGAR learning system is composed of two integrated feedforward networks. One neural network acts as a critic network to guide the learning of the other network (the action network) which determines the outputs (actions) of the TDGAR learning system. The action network can be a normal neural network or a neural fuzzy network. Using the TD prediction method, the critic network can predict the external reinforcement signal and provide a more informative internal reinforcement signal to the action network. The action network uses the GA to adapt itself according to the internal reinforcement signal. The key concept of the TDGAR learning scheme is to formulate the internal reinforcement signal as the fitness function for the GA such that the GA can evaluate the candidate solutions (chromosomes) regularly, even during periods without external feedback from the environment. This enables the GA to proceed to new generations regularly without waiting for the arrival of the external reinforcement signal. This can usually accelerate the GA learning since a reinforcement signal may only be available at a time long after a sequence of actions has occurred in the reinforcement learning problem. The proposed TDGAR learning system has been used to control an active magnetic bearing (AMB) system in practice. A systematic design procedure is developed to achieve successful integration of all the subsystems including magnetic suspension, mechanical structure, and controller training. The results show that the TDGAR learning scheme can successfully find a neural controller or a neural fuzzy controller for a self-designed magnetic bearing system.

An Ensemble Deep Convolutional Neural Network Model with Improved D-S Evidence Fusion for Bearing Fault Diagnosis.

PubMed

Li, Shaobo; Liu, Guokai; Tang, Xianghong; Lu, Jianguang; Hu, Jianjun

2017-07-28

Intelligent machine health monitoring and fault diagnosis are becoming increasingly important for modern manufacturing industries. Current fault diagnosis approaches mostly depend on expert-designed features for building prediction models. In this paper, we proposed IDSCNN, a novel bearing fault diagnosis algorithm based on ensemble deep convolutional neural networks and an improved Dempster-Shafer theory based evidence fusion. The convolutional neural networks take the root mean square (RMS) maps from the FFT (Fast Fourier Transformation) features of the vibration signals from two sensors as inputs. The improved D-S evidence theory is implemented via distance matrix from evidences and modified Gini Index. Extensive evaluations of the IDSCNN on the Case Western Reserve Dataset showed that our IDSCNN algorithm can achieve better fault diagnosis performance than existing machine learning methods by fusing complementary or conflicting evidences from different models and sensors and adapting to different load conditions.

An Ensemble Deep Convolutional Neural Network Model with Improved D-S Evidence Fusion for Bearing Fault Diagnosis

PubMed Central

Li, Shaobo; Liu, Guokai; Tang, Xianghong; Lu, Jianguang

2017-01-01

Intelligent machine health monitoring and fault diagnosis are becoming increasingly important for modern manufacturing industries. Current fault diagnosis approaches mostly depend on expert-designed features for building prediction models. In this paper, we proposed IDSCNN, a novel bearing fault diagnosis algorithm based on ensemble deep convolutional neural networks and an improved Dempster–Shafer theory based evidence fusion. The convolutional neural networks take the root mean square (RMS) maps from the FFT (Fast Fourier Transformation) features of the vibration signals from two sensors as inputs. The improved D-S evidence theory is implemented via distance matrix from evidences and modified Gini Index. Extensive evaluations of the IDSCNN on the Case Western Reserve Dataset showed that our IDSCNN algorithm can achieve better fault diagnosis performance than existing machine learning methods by fusing complementary or conflicting evidences from different models and sensors and adapting to different load conditions. PMID:28788099

Computational model for perception of objects and motions.

PubMed

Yang, WenLu; Zhang, LiQing; Ma, LiBo

2008-06-01

Perception of objects and motions in the visual scene is one of the basic problems in the visual system. There exist 'What' and 'Where' pathways in the superior visual cortex, starting from the simple cells in the primary visual cortex. The former is able to perceive objects such as forms, color, and texture, and the latter perceives 'where', for example, velocity and direction of spatial movement of objects. This paper explores brain-like computational architectures of visual information processing. We propose a visual perceptual model and computational mechanism for training the perceptual model. The computational model is a three-layer network. The first layer is the input layer which is used to receive the stimuli from natural environments. The second layer is designed for representing the internal neural information. The connections between the first layer and the second layer, called the receptive fields of neurons, are self-adaptively learned based on principle of sparse neural representation. To this end, we introduce Kullback-Leibler divergence as the measure of independence between neural responses and derive the learning algorithm based on minimizing the cost function. The proposed algorithm is applied to train the basis functions, namely receptive fields, which are localized, oriented, and bandpassed. The resultant receptive fields of neurons in the second layer have the characteristics resembling that of simple cells in the primary visual cortex. Based on these basis functions, we further construct the third layer for perception of what and where in the superior visual cortex. The proposed model is able to perceive objects and their motions with a high accuracy and strong robustness against additive noise. Computer simulation results in the final section show the feasibility of the proposed perceptual model and high efficiency of the learning algorithm.

Associative memory in an analog iterated-map neural network

NASA Astrophysics Data System (ADS)

Marcus, C. M.; Waugh, F. R.; Westervelt, R. M.

1990-03-01

The behavior of an analog neural network with parallel dynamics is studied analytically and numerically for two associative-memory learning algorithms, the Hebb rule and the pseudoinverse rule. Phase diagrams in the parameter space of analog gain β and storage ratio α are presented. For both learning rules, the networks have large ``recall'' phases in which retrieval states exist and convergence to a fixed point is guaranteed by a global stability criterion. We also demonstrate numerically that using a reduced analog gain increases the probability of recall starting from a random initial state. This phenomenon is comparable to thermal annealing used to escape local minima but has the advantage of being deterministic, and therefore easily implemented in electronic hardware. Similarities and differences between analog neural networks and networks with two-state neurons at finite temperature are also discussed.

Deep Constrained Siamese Hash Coding Network and Load-Balanced Locality-Sensitive Hashing for Near Duplicate Image Detection.

PubMed

Hu, Weiming; Fan, Yabo; Xing, Junliang; Sun, Liang; Cai, Zhaoquan; Maybank, Stephen

2018-09-01

We construct a new efficient near duplicate image detection method using a hierarchical hash code learning neural network and load-balanced locality-sensitive hashing (LSH) indexing. We propose a deep constrained siamese hash coding neural network combined with deep feature learning. Our neural network is able to extract effective features for near duplicate image detection. The extracted features are used to construct a LSH-based index. We propose a load-balanced LSH method to produce load-balanced buckets in the hashing process. The load-balanced LSH significantly reduces the query time. Based on the proposed load-balanced LSH, we design an effective and feasible algorithm for near duplicate image detection. Extensive experiments on three benchmark data sets demonstrate the effectiveness of our deep siamese hash encoding network and load-balanced LSH.

Neuroprosthetic Decoder Training as Imitation Learning.

PubMed

Merel, Josh; Carlson, David; Paninski, Liam; Cunningham, John P

2016-05-01

Neuroprosthetic brain-computer interfaces function via an algorithm which decodes neural activity of the user into movements of an end effector, such as a cursor or robotic arm. In practice, the decoder is often learned by updating its parameters while the user performs a task. When the user's intention is not directly observable, recent methods have demonstrated value in training the decoder against a surrogate for the user's intended movement. Here we show that training a decoder in this way is a novel variant of an imitation learning problem, where an oracle or expert is employed for supervised training in lieu of direct observations, which are not available. Specifically, we describe how a generic imitation learning meta-algorithm, dataset aggregation (DAgger), can be adapted to train a generic brain-computer interface. By deriving existing learning algorithms for brain-computer interfaces in this framework, we provide a novel analysis of regret (an important metric of learning efficacy) for brain-computer interfaces. This analysis allows us to characterize the space of algorithmic variants and bounds on their regret rates. Existing approaches for decoder learning have been performed in the cursor control setting, but the available design principles for these decoders are such that it has been impossible to scale them to naturalistic settings. Leveraging our findings, we then offer an algorithm that combines imitation learning with optimal control, which should allow for training of arbitrary effectors for which optimal control can generate goal-oriented control. We demonstrate this novel and general BCI algorithm with simulated neuroprosthetic control of a 26 degree-of-freedom model of an arm, a sophisticated and realistic end effector.

Sequential Nonlinear Learning for Distributed Multiagent Systems via Extreme Learning Machines.

PubMed

Vanli, Nuri Denizcan; Sayin, Muhammed O; Delibalta, Ibrahim; Kozat, Suleyman Serdar

2017-03-01

We study online nonlinear learning over distributed multiagent systems, where each agent employs a single hidden layer feedforward neural network (SLFN) structure to sequentially minimize arbitrary loss functions. In particular, each agent trains its own SLFN using only the data that is revealed to itself. On the other hand, the aim of the multiagent system is to train the SLFN at each agent as well as the optimal centralized batch SLFN that has access to all the data, by exchanging information between neighboring agents. We address this problem by introducing a distributed subgradient-based extreme learning machine algorithm. The proposed algorithm provides guaranteed upper bounds on the performance of the SLFN at each agent and shows that each of these individual SLFNs asymptotically achieves the performance of the optimal centralized batch SLFN. Our performance guarantees explicitly distinguish the effects of data- and network-dependent parameters on the convergence rate of the proposed algorithm. The experimental results illustrate that the proposed algorithm achieves the oracle performance significantly faster than the state-of-the-art methods in the machine learning and signal processing literature. Hence, the proposed method is highly appealing for the applications involving big data.

Cascade process modeling with mechanism-based hierarchical neural networks.

PubMed

Cong, Qiumei; Yu, Wen; Chai, Tianyou

2010-02-01

Cascade process, such as wastewater treatment plant, includes many nonlinear sub-systems and many variables. When the number of sub-systems is big, the input-output relation in the first block and the last block cannot represent the whole process. In this paper we use two techniques to overcome the above problem. Firstly we propose a new neural model: hierarchical neural networks to identify the cascade process; then we use serial structural mechanism model based on the physical equations to connect with neural model. A stable learning algorithm and theoretical analysis are given. Finally, this method is used to model a wastewater treatment plant. Real operational data of wastewater treatment plant is applied to illustrate the modeling approach.

Efficient Prediction of Low-Visibility Events at Airports Using Machine-Learning Regression

NASA Astrophysics Data System (ADS)

Cornejo-Bueno, L.; Casanova-Mateo, C.; Sanz-Justo, J.; Cerro-Prada, E.; Salcedo-Sanz, S.

2017-11-01

We address the prediction of low-visibility events at airports using machine-learning regression. The proposed model successfully forecasts low-visibility events in terms of the runway visual range at the airport, with the use of support-vector regression, neural networks (multi-layer perceptrons and extreme-learning machines) and Gaussian-process algorithms. We assess the performance of these algorithms based on real data collected at the Valladolid airport, Spain. We also propose a study of the atmospheric variables measured at a nearby tower related to low-visibility atmospheric conditions, since they are considered as the inputs of the different regressors. A pre-processing procedure of these input variables with wavelet transforms is also described. The results show that the proposed machine-learning algorithms are able to predict low-visibility events well. The Gaussian process is the best algorithm among those analyzed, obtaining over 98% of the correct classification rate in low-visibility events when the runway visual range is {>}1000 m, and about 80% under this threshold. The performance of all the machine-learning algorithms tested is clearly affected in extreme low-visibility conditions ({<}500 m). However, we show improved results of all the methods when data from a neighbouring meteorological tower are included, and also with a pre-processing scheme using a wavelet transform. Also presented are results of the algorithm performance in daytime and nighttime conditions, and for different prediction time horizons.

Reinforcement learning in multidimensional environments relies on attention mechanisms.

PubMed

Niv, Yael; Daniel, Reka; Geana, Andra; Gershman, Samuel J; Leong, Yuan Chang; Radulescu, Angela; Wilson, Robert C

2015-05-27

In recent years, ideas from the computational field of reinforcement learning have revolutionized the study of learning in the brain, famously providing new, precise theories of how dopamine affects learning in the basal ganglia. However, reinforcement learning algorithms are notorious for not scaling well to multidimensional environments, as is required for real-world learning. We hypothesized that the brain naturally reduces the dimensionality of real-world problems to only those dimensions that are relevant to predicting reward, and conducted an experiment to assess by what algorithms and with what neural mechanisms this "representation learning" process is realized in humans. Our results suggest that a bilateral attentional control network comprising the intraparietal sulcus, precuneus, and dorsolateral prefrontal cortex is involved in selecting what dimensions are relevant to the task at hand, effectively updating the task representation through trial and error. In this way, cortical attention mechanisms interact with learning in the basal ganglia to solve the "curse of dimensionality" in reinforcement learning. Copyright © 2015 the authors 0270-6474/15/358145-13$15.00/0.

Using Deep Learning to Analyze the Voices of Stars.

NASA Astrophysics Data System (ADS)

Boudreaux, Thomas Macaulay

2018-01-01

With several new large-scale surveys on the horizon, including LSST, TESS, ZTF, and Evryscope, faster and more accurate analysis methods will be required to adequately process the enormous amount of data produced. Deep learning, used in industry for years now, allows for advanced feature detection in minimally prepared datasets at very high speeds; however, despite the advantages of this method, its application to astrophysics has not yet been extensively explored. This dearth may be due to a lack of training data available to researchers. Here we generate synthetic data loosely mimicking the properties of acoustic mode pulsating stars and compare the performance of different deep learning algorithms, including Artifical Neural Netoworks, and Convolutional Neural Networks, in classifing these synthetic data sets as either pulsators, or not observed to vary stars.

Genetic learning in rule-based and neural systems

NASA Technical Reports Server (NTRS)

Smith, Robert E.

1993-01-01

The design of neural networks and fuzzy systems can involve complex, nonlinear, and ill-conditioned optimization problems. Often, traditional optimization schemes are inadequate or inapplicable for such tasks. Genetic Algorithms (GA's) are a class of optimization procedures whose mechanics are based on those of natural genetics. Mathematical arguments show how GAs bring substantial computational leverage to search problems, without requiring the mathematical characteristics often necessary for traditional optimization schemes (e.g., modality, continuity, availability of derivative information, etc.). GA's have proven effective in a variety of search tasks that arise in neural networks and fuzzy systems. This presentation begins by introducing the mechanism and theoretical underpinnings of GA's. GA's are then related to a class of rule-based machine learning systems called learning classifier systems (LCS's). An LCS implements a low-level production-system that uses a GA as its primary rule discovery mechanism. This presentation illustrates how, despite its rule-based framework, an LCS can be thought of as a competitive neural network. Neural network simulator code for an LCS is presented. In this context, the GA is doing more than optimizing and objective function. It is searching for an ecology of hidden nodes with limited connectivity. The GA attempts to evolve this ecology such that effective neural network performance results. The GA is particularly well adapted to this task, given its naturally-inspired basis. The LCS/neural network analogy extends itself to other, more traditional neural networks. Conclusions to the presentation discuss the implications of using GA's in ecological search problems that arise in neural and fuzzy systems.

A novel stock forecasting model based on High-order-fuzzy-fluctuation Trends and Back Propagation Neural Network

PubMed Central

Dai, Zongli; Zhao, Aiwu; He, Jie

2018-01-01

In this paper, we propose a hybrid method to forecast the stock prices called High-order-fuzzy-fluctuation-Trends-based Back Propagation(HTBP)Neural Network model. First, we compare each value of the historical training data with the previous day's value to obtain a fluctuation trend time series (FTTS). On this basis, the FTTS blur into fuzzy time series (FFTS) based on the fluctuation of the increasing, equality, decreasing amplitude and direction. Since the relationship between FFTS and future wave trends is nonlinear, the HTBP neural network algorithm is used to find the mapping rules in the form of self-learning. Finally, the results of the algorithm output are used to predict future fluctuations. The proposed model provides some innovative features:(1)It combines fuzzy set theory and neural network algorithm to avoid overfitting problems existed in traditional models. (2)BP neural network algorithm can intelligently explore the internal rules of the actual existence of sequential data, without the need to analyze the influence factors of specific rules and the path of action. (3)The hybrid modal can reasonably remove noises from the internal rules by proper fuzzy treatment. This paper takes the TAIEX data set of Taiwan stock exchange as an example, and compares and analyzes the prediction performance of the model. The experimental results show that this method can predict the stock market in a very simple way. At the same time, we use this method to predict the Shanghai stock exchange composite index, and further verify the effectiveness and universality of the method. PMID:29420584

A novel stock forecasting model based on High-order-fuzzy-fluctuation Trends and Back Propagation Neural Network.

PubMed

Guan, Hongjun; Dai, Zongli; Zhao, Aiwu; He, Jie

2018-01-01

In this paper, we propose a hybrid method to forecast the stock prices called High-order-fuzzy-fluctuation-Trends-based Back Propagation(HTBP)Neural Network model. First, we compare each value of the historical training data with the previous day's value to obtain a fluctuation trend time series (FTTS). On this basis, the FTTS blur into fuzzy time series (FFTS) based on the fluctuation of the increasing, equality, decreasing amplitude and direction. Since the relationship between FFTS and future wave trends is nonlinear, the HTBP neural network algorithm is used to find the mapping rules in the form of self-learning. Finally, the results of the algorithm output are used to predict future fluctuations. The proposed model provides some innovative features:(1)It combines fuzzy set theory and neural network algorithm to avoid overfitting problems existed in traditional models. (2)BP neural network algorithm can intelligently explore the internal rules of the actual existence of sequential data, without the need to analyze the influence factors of specific rules and the path of action. (3)The hybrid modal can reasonably remove noises from the internal rules by proper fuzzy treatment. This paper takes the TAIEX data set of Taiwan stock exchange as an example, and compares and analyzes the prediction performance of the model. The experimental results show that this method can predict the stock market in a very simple way. At the same time, we use this method to predict the Shanghai stock exchange composite index, and further verify the effectiveness and universality of the method.

Precipitation Nowcast using Deep Recurrent Neural Network

NASA Astrophysics Data System (ADS)

Akbari Asanjan, A.; Yang, T.; Gao, X.; Hsu, K. L.; Sorooshian, S.

2016-12-01

An accurate precipitation nowcast (0-6 hours) with a fine temporal and spatial resolution has always been an important prerequisite for flood warning, streamflow prediction and risk management. Most of the popular approaches used for forecasting precipitation can be categorized into two groups. One type of precipitation forecast relies on numerical modeling of the physical dynamics of atmosphere and another is based on empirical and statistical regression models derived by local hydrologists or meteorologists. Given the recent advances in artificial intelligence, in this study a powerful Deep Recurrent Neural Network, termed as Long Short-Term Memory (LSTM) model, is creatively used to extract the patterns and forecast the spatial and temporal variability of Cloud Top Brightness Temperature (CTBT) observed from GOES satellite. Then, a 0-6 hours precipitation nowcast is produced using a Precipitation Estimation from Remote Sensing Information using Artificial Neural Network (PERSIANN) algorithm, in which the CTBT nowcast is used as the PERSIANN algorithm's raw inputs. Two case studies over the continental U.S. have been conducted that demonstrate the improvement of proposed approach as compared to a classical Feed Forward Neural Network and a couple simple regression models. The advantages and disadvantages of the proposed method are summarized with regard to its capability of pattern recognition through time, handling of vanishing gradient during model learning, and working with sparse data. The studies show that the LSTM model performs better than other methods, and it is able to learn the temporal evolution of the precipitation events through over 1000 time lags. The uniqueness of PERSIANN's algorithm enables an alternative precipitation nowcast approach as demonstrated in this study, in which the CTBT prediction is produced and used as the inputs for generating precipitation nowcast.

Multi-mode energy management strategy for fuel cell electric vehicles based on driving pattern identification using learning vector quantization neural network algorithm

NASA Astrophysics Data System (ADS)

Song, Ke; Li, Feiqiang; Hu, Xiao; He, Lin; Niu, Wenxu; Lu, Sihao; Zhang, Tong

2018-06-01

The development of fuel cell electric vehicles can to a certain extent alleviate worldwide energy and environmental issues. While a single energy management strategy cannot meet the complex road conditions of an actual vehicle, this article proposes a multi-mode energy management strategy for electric vehicles with a fuel cell range extender based on driving condition recognition technology, which contains a patterns recognizer and a multi-mode energy management controller. This paper introduces a learning vector quantization (LVQ) neural network to design the driving patterns recognizer according to a vehicle's driving information. This multi-mode strategy can automatically switch to the genetic algorithm optimized thermostat strategy under specific driving conditions in the light of the differences in condition recognition results. Simulation experiments were carried out based on the model's validity verification using a dynamometer test bench. Simulation results show that the proposed strategy can obtain better economic performance than the single-mode thermostat strategy under dynamic driving conditions.

Seismic waveform classification using deep learning

NASA Astrophysics Data System (ADS)

Kong, Q.; Allen, R. M.

2017-12-01

MyShake is a global smartphone seismic network that harnesses the power of crowdsourcing. It has an Artificial Neural Network (ANN) algorithm running on the phone to distinguish earthquake motion from human activities recorded by the accelerometer on board. Once the ANN detects earthquake-like motion, it sends a 5-min chunk of acceleration data back to the server for further analysis. The time-series data collected contains both earthquake data and human activity data that the ANN confused. In this presentation, we will show the Convolutional Neural Network (CNN) we built under the umbrella of supervised learning to find out the earthquake waveform. The waveforms of the recorded motion could treat easily as images, and by taking the advantage of the power of CNN processing the images, we achieved very high successful rate to select the earthquake waveforms out. Since there are many non-earthquake waveforms than the earthquake waveforms, we also built an anomaly detection algorithm using the CNN. Both these two methods can be easily extended to other waveform classification problems.

Learning-based computing techniques in geoid modeling for precise height transformation

NASA Astrophysics Data System (ADS)

Erol, B.; Erol, S.

2013-03-01

Precise determination of local geoid is of particular importance for establishing height control in geodetic GNSS applications, since the classical leveling technique is too laborious. A geoid model can be accurately obtained employing properly distributed benchmarks having GNSS and leveling observations using an appropriate computing algorithm. Besides the classical multivariable polynomial regression equations (MPRE), this study attempts an evaluation of learning based computing algorithms: artificial neural networks (ANNs), adaptive network-based fuzzy inference system (ANFIS) and especially the wavelet neural networks (WNNs) approach in geoid surface approximation. These algorithms were developed parallel to advances in computer technologies and recently have been used for solving complex nonlinear problems of many applications. However, they are rather new in dealing with precise modeling problem of the Earth gravity field. In the scope of the study, these methods were applied to Istanbul GPS Triangulation Network data. The performances of the methods were assessed considering the validation results of the geoid models at the observation points. In conclusion the ANFIS and WNN revealed higher prediction accuracies compared to ANN and MPRE methods. Beside the prediction capabilities, these methods were also compared and discussed from the practical point of view in conclusions.

Neural network based online simultaneous policy update algorithm for solving the HJI equation in nonlinear H∞ control.

PubMed

Wu, Huai-Ning; Luo, Biao

2012-12-01

It is well known that the nonlinear H∞ state feedback control problem relies on the solution of the Hamilton-Jacobi-Isaacs (HJI) equation, which is a nonlinear partial differential equation that has proven to be impossible to solve analytically. In this paper, a neural network (NN)-based online simultaneous policy update algorithm (SPUA) is developed to solve the HJI equation, in which knowledge of internal system dynamics is not required. First, we propose an online SPUA which can be viewed as a reinforcement learning technique for two players to learn their optimal actions in an unknown environment. The proposed online SPUA updates control and disturbance policies simultaneously; thus, only one iterative loop is needed. Second, the convergence of the online SPUA is established by proving that it is mathematically equivalent to Newton's method for finding a fixed point in a Banach space. Third, we develop an actor-critic structure for the implementation of the online SPUA, in which only one critic NN is needed for approximating the cost function, and a least-square method is given for estimating the NN weight parameters. Finally, simulation studies are provided to demonstrate the effectiveness of the proposed algorithm.

Comment on ‘Deep convolutional neural network with transfer learning for rectum toxicity prediction in cervical cancer radiotherapy: a feasibility study’

NASA Astrophysics Data System (ADS)

Valdes, Gilmer; Interian, Yannet

2018-03-01

The application of machine learning (ML) presents tremendous opportunities for the field of oncology, thus we read ‘Deep convolutional neural network with transfer learning for rectum toxicity prediction in cervical cancer radiotherapy: a feasibility study’ with great interest. In this article, the authors used state of the art techniques: a pre-trained convolutional neural network (VGG-16 CNN), transfer learning, data augmentation, drop out and early stopping, all of which are directly responsible for the success and the excitement that these algorithms have created in other fields. We believe that the use of these techniques can offer tremendous opportunities in the field of Medical Physics and as such we would like to praise the authors for their pioneering application to the field of Radiation Oncology. That being said, given that the field of Medical Physics has unique characteristics that differentiate us from those fields where these techniques have been applied successfully, we would like to raise some points for future discussion and follow up studies that could help the community understand the limitations and nuances of deep learning techniques.

Hybrid Model Based on Genetic Algorithms and SVM Applied to Variable Selection within Fruit Juice Classification

PubMed Central

Fernandez-Lozano, C.; Canto, C.; Gestal, M.; Andrade-Garda, J. M.; Rabuñal, J. R.; Dorado, J.; Pazos, A.

2013-01-01

Given the background of the use of Neural Networks in problems of apple juice classification, this paper aim at implementing a newly developed method in the field of machine learning: the Support Vector Machines (SVM). Therefore, a hybrid model that combines genetic algorithms and support vector machines is suggested in such a way that, when using SVM as a fitness function of the Genetic Algorithm (GA), the most representative variables for a specific classification problem can be selected. PMID:24453933

Application of artificial intelligence to search ground-state geometry of clusters

NASA Astrophysics Data System (ADS)

Lemes, Maurício Ruv; Marim, L. R.; dal Pino, A.

2002-08-01

We introduce a global optimization procedure, the neural-assisted genetic algorithm (NAGA). It combines the power of an artificial neural network (ANN) with the versatility of the genetic algorithm. This method is suitable to solve optimization problems that depend on some kind of heuristics to limit the search space. If a reasonable amount of data is available, the ANN can ``understand'' the problem and provide the genetic algorithm with a selected population of elements that will speed up the search for the optimum solution. We tested the method in a search for the ground-state geometry of silicon clusters. We trained the ANN with information about the geometry and energetics of small silicon clusters. Next, the ANN learned how to restrict the configurational space for larger silicon clusters. For Si10 and Si20, we noticed that the NAGA is at least three times faster than the ``pure'' genetic algorithm. As the size of the cluster increases, it is expected that the gain in terms of time will increase as well.

HD-MTL: Hierarchical Deep Multi-Task Learning for Large-Scale Visual Recognition.

PubMed

Fan, Jianping; Zhao, Tianyi; Kuang, Zhenzhong; Zheng, Yu; Zhang, Ji; Yu, Jun; Peng, Jinye

2017-02-09

In this paper, a hierarchical deep multi-task learning (HD-MTL) algorithm is developed to support large-scale visual recognition (e.g., recognizing thousands or even tens of thousands of atomic object classes automatically). First, multiple sets of multi-level deep features are extracted from different layers of deep convolutional neural networks (deep CNNs), and they are used to achieve more effective accomplishment of the coarseto- fine tasks for hierarchical visual recognition. A visual tree is then learned by assigning the visually-similar atomic object classes with similar learning complexities into the same group, which can provide a good environment for determining the interrelated learning tasks automatically. By leveraging the inter-task relatedness (inter-class similarities) to learn more discriminative group-specific deep representations, our deep multi-task learning algorithm can train more discriminative node classifiers for distinguishing the visually-similar atomic object classes effectively. Our hierarchical deep multi-task learning (HD-MTL) algorithm can integrate two discriminative regularization terms to control the inter-level error propagation effectively, and it can provide an end-to-end approach for jointly learning more representative deep CNNs (for image representation) and more discriminative tree classifier (for large-scale visual recognition) and updating them simultaneously. Our incremental deep learning algorithms can effectively adapt both the deep CNNs and the tree classifier to the new training images and the new object classes. Our experimental results have demonstrated that our HD-MTL algorithm can achieve very competitive results on improving the accuracy rates for large-scale visual recognition.

Study on Temperature and Synthetic Compensation of Piezo-Resistive Differential Pressure Sensors by Coupled Simulated Annealing and Simplex Optimized Kernel Extreme Learning Machine

PubMed Central

Li, Ji; Hu, Guoqing; Zhou, Yonghong; Zou, Chong; Peng, Wei; Alam SM, Jahangir

2017-01-01

As a high performance-cost ratio solution for differential pressure measurement, piezo-resistive differential pressure sensors are widely used in engineering processes. However, their performance is severely affected by the environmental temperature and the static pressure applied to them. In order to modify the non-linear measuring characteristics of the piezo-resistive differential pressure sensor, compensation actions should synthetically consider these two aspects. Advantages such as nonlinear approximation capability, highly desirable generalization ability and computational efficiency make the kernel extreme learning machine (KELM) a practical approach for this critical task. Since the KELM model is intrinsically sensitive to the regularization parameter and the kernel parameter, a searching scheme combining the coupled simulated annealing (CSA) algorithm and the Nelder-Mead simplex algorithm is adopted to find an optimal KLEM parameter set. A calibration experiment at different working pressure levels was conducted within the temperature range to assess the proposed method. In comparison with other compensation models such as the back-propagation neural network (BP), radius basis neural network (RBF), particle swarm optimization optimized support vector machine (PSO-SVM), particle swarm optimization optimized least squares support vector machine (PSO-LSSVM) and extreme learning machine (ELM), the compensation results show that the presented compensation algorithm exhibits a more satisfactory performance with respect to temperature compensation and synthetic compensation problems. PMID:28422080

Study on Temperature and Synthetic Compensation of Piezo-Resistive Differential Pressure Sensors by Coupled Simulated Annealing and Simplex Optimized Kernel Extreme Learning Machine.

PubMed

Li, Ji; Hu, Guoqing; Zhou, Yonghong; Zou, Chong; Peng, Wei; Alam Sm, Jahangir

2017-04-19

As a high performance-cost ratio solution for differential pressure measurement, piezo-resistive differential pressure sensors are widely used in engineering processes. However, their performance is severely affected by the environmental temperature and the static pressure applied to them. In order to modify the non-linear measuring characteristics of the piezo-resistive differential pressure sensor, compensation actions should synthetically consider these two aspects. Advantages such as nonlinear approximation capability, highly desirable generalization ability and computational efficiency make the kernel extreme learning machine (KELM) a practical approach for this critical task. Since the KELM model is intrinsically sensitive to the regularization parameter and the kernel parameter, a searching scheme combining the coupled simulated annealing (CSA) algorithm and the Nelder-Mead simplex algorithm is adopted to find an optimal KLEM parameter set. A calibration experiment at different working pressure levels was conducted within the temperature range to assess the proposed method. In comparison with other compensation models such as the back-propagation neural network (BP), radius basis neural network (RBF), particle swarm optimization optimized support vector machine (PSO-SVM), particle swarm optimization optimized least squares support vector machine (PSO-LSSVM) and extreme learning machine (ELM), the compensation results show that the presented compensation algorithm exhibits a more satisfactory performance with respect to temperature compensation and synthetic compensation problems.

Multivariate cross-classification: applying machine learning techniques to characterize abstraction in neural representations

PubMed Central

Kaplan, Jonas T.; Man, Kingson; Greening, Steven G.

2015-01-01

Here we highlight an emerging trend in the use of machine learning classifiers to test for abstraction across patterns of neural activity. When a classifier algorithm is trained on data from one cognitive context, and tested on data from another, conclusions can be drawn about the role of a given brain region in representing information that abstracts across those cognitive contexts. We call this kind of analysis Multivariate Cross-Classification (MVCC), and review several domains where it has recently made an impact. MVCC has been important in establishing correspondences among neural patterns across cognitive domains, including motor-perception matching and cross-sensory matching. It has been used to test for similarity between neural patterns evoked by perception and those generated from memory. Other work has used MVCC to investigate the similarity of representations for semantic categories across different kinds of stimulus presentation, and in the presence of different cognitive demands. We use these examples to demonstrate the power of MVCC as a tool for investigating neural abstraction and discuss some important methodological issues related to its application. PMID:25859202

Machine learning of neural representations of suicide and emotion concepts identifies suicidal youth

PubMed Central

Just, Marcel Adam; Pan, Lisa; Cherkassky, Vladimir L.; McMakin, Dana; Cha, Christine; Nock, Matthew K.; Brent, David

2017-01-01

The clinical assessment of suicidal risk would be significantly complemented by a biologically-based measure that assesses alterations in the neural representations of concepts related to death and life in people who engage in suicidal ideation. This study used machine-learning algorithms (Gaussian Naïve Bayes) to identify such individuals (17 suicidal ideators vs 17 controls) with high (91%) accuracy, based on their altered fMRI neural signatures of death and life-related concepts. The most discriminating concepts were death, cruelty, trouble, carefree, good, and praise. A similar classification accurately (94%) discriminated 9 suicidal ideators who had made a suicide attempt from 8 who had not. Moreover, a major facet of the concept alterations was the evoked emotion, whose neural signature served as an alternative basis for accurate (85%) group classification. The study establishes a biological, neurocognitive basis for altered concept representations in participants with suicidal ideation, which enables highly accurate group membership classification. PMID:29367952

Vision-based mobile robot navigation through deep convolutional neural networks and end-to-end learning

NASA Astrophysics Data System (ADS)

Zhang, Yachu; Zhao, Yuejin; Liu, Ming; Dong, Liquan; Kong, Lingqin; Liu, Lingling

2017-09-01

In contrast to humans, who use only visual information for navigation, many mobile robots use laser scanners and ultrasonic sensors along with vision cameras to navigate. This work proposes a vision-based robot control algorithm based on deep convolutional neural networks. We create a large 15-layer convolutional neural network learning system and achieve the advanced recognition performance. Our system is trained from end to end to map raw input images to direction in supervised mode. The images of data sets are collected in a wide variety of weather conditions and lighting conditions. Besides, the data sets are augmented by adding Gaussian noise and Salt-and-pepper noise to avoid overfitting. The algorithm is verified by two experiments, which are line tracking and obstacle avoidance. The line tracking experiment is proceeded in order to track the desired path which is composed of straight and curved lines. The goal of obstacle avoidance experiment is to avoid the obstacles indoor. Finally, we get 3.29% error rate on the training set and 5.1% error rate on the test set in the line tracking experiment, 1.8% error rate on the training set and less than 5% error rate on the test set in the obstacle avoidance experiment. During the actual test, the robot can follow the runway centerline outdoor and avoid the obstacle in the room accurately. The result confirms the effectiveness of the algorithm and our improvement in the network structure and train parameters

Neural Mechanism for Stochastic Behavior During a Competitive Game

PubMed Central

Soltani, Alireza; Lee, Daeyeol; Wang, Xiao-Jing

2006-01-01

Previous studies have shown that non-human primates can generate highly stochastic choice behavior, especially when this is required during a competitive interaction with another agent. To understand the neural mechanism of such dynamic choice behavior, we propose a biologically plausible model of decision making endowed with synaptic plasticity that follows a reward-dependent stochastic Hebbian learning rule. This model constitutes a biophysical implementation of reinforcement learning, and it reproduces salient features of behavioral data from an experiment with monkeys playing a matching pennies game. Due to interaction with an opponent and learning dynamics, the model generates quasi-random behavior robustly in spite of intrinsic biases. Furthermore, non-random choice behavior can also emerge when the model plays against a non-interactive opponent, as observed in the monkey experiment. Finally, when combined with a meta-learning algorithm, our model accounts for the slow drift in the animal’s strategy based on a process of reward maximization. PMID:17015181

Reinforcement Learning in Multidimensional Environments Relies on Attention Mechanisms

PubMed Central

Daniel, Reka; Geana, Andra; Gershman, Samuel J.; Leong, Yuan Chang; Radulescu, Angela; Wilson, Robert C.

2015-01-01

In recent years, ideas from the computational field of reinforcement learning have revolutionized the study of learning in the brain, famously providing new, precise theories of how dopamine affects learning in the basal ganglia. However, reinforcement learning algorithms are notorious for not scaling well to multidimensional environments, as is required for real-world learning. We hypothesized that the brain naturally reduces the dimensionality of real-world problems to only those dimensions that are relevant to predicting reward, and conducted an experiment to assess by what algorithms and with what neural mechanisms this “representation learning” process is realized in humans. Our results suggest that a bilateral attentional control network comprising the intraparietal sulcus, precuneus, and dorsolateral prefrontal cortex is involved in selecting what dimensions are relevant to the task at hand, effectively updating the task representation through trial and error. In this way, cortical attention mechanisms interact with learning in the basal ganglia to solve the “curse of dimensionality” in reinforcement learning. PMID:26019331

Evolving autonomous learning in cognitive networks.

PubMed

Sheneman, Leigh; Hintze, Arend

2017-12-01

There are two common approaches for optimizing the performance of a machine: genetic algorithms and machine learning. A genetic algorithm is applied over many generations whereas machine learning works by applying feedback until the system meets a performance threshold. These methods have been previously combined, particularly in artificial neural networks using an external objective feedback mechanism. We adapt this approach to Markov Brains, which are evolvable networks of probabilistic and deterministic logic gates. Prior to this work MB could only adapt from one generation to the other, so we introduce feedback gates which augment their ability to learn during their lifetime. We show that Markov Brains can incorporate these feedback gates in such a way that they do not rely on an external objective feedback signal, but instead can generate internal feedback that is then used to learn. This results in a more biologically accurate model of the evolution of learning, which will enable us to study the interplay between evolution and learning and could be another step towards autonomously learning machines.

Reinforcement learning with Marr.

PubMed

Niv, Yael; Langdon, Angela

2016-10-01

To many, the poster child for David Marr's famous three levels of scientific inquiry is reinforcement learning-a computational theory of reward optimization, which readily prescribes algorithmic solutions that evidence striking resemblance to signals found in the brain, suggesting a straightforward neural implementation. Here we review questions that remain open at each level of analysis, concluding that the path forward to their resolution calls for inspiration across levels, rather than a focus on mutual constraints.

Learning to read aloud: A neural network approach using sparse distributed memory

NASA Technical Reports Server (NTRS)

Joglekar, Umesh Dwarkanath

1989-01-01

An attempt to solve a problem of text-to-phoneme mapping is described which does not appear amenable to solution by use of standard algorithmic procedures. Experiments based on a model of distributed processing are also described. This model (sparse distributed memory (SDM)) can be used in an iterative supervised learning mode to solve the problem. Additional improvements aimed at obtaining better performance are suggested.

A convolutional neural network neutrino event classifier

DOE PAGES

Aurisano, A.; Radovic, A.; Rocco, D.; ...

2016-09-01

Here, convolutional neural networks (CNNs) have been widely applied in the computer vision community to solve complex problems in image recognition and analysis. We describe an application of the CNN technology to the problem of identifying particle interactions in sampling calorimeters used commonly in high energy physics and high energy neutrino physics in particular. Following a discussion of the core concepts of CNNs and recent innovations in CNN architectures related to the field of deep learning, we outline a specific application to the NOvA neutrino detector. This algorithm, CVN (Convolutional Visual Network) identifies neutrino interactions based on their topology withoutmore » the need for detailed reconstruction and outperforms algorithms currently in use by the NOvA collaboration.« less

A convolutional neural network neutrino event classifier

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

Aurisano, A.; Radovic, A.; Rocco, D.

Here, convolutional neural networks (CNNs) have been widely applied in the computer vision community to solve complex problems in image recognition and analysis. We describe an application of the CNN technology to the problem of identifying particle interactions in sampling calorimeters used commonly in high energy physics and high energy neutrino physics in particular. Following a discussion of the core concepts of CNNs and recent innovations in CNN architectures related to the field of deep learning, we outline a specific application to the NOvA neutrino detector. This algorithm, CVN (Convolutional Visual Network) identifies neutrino interactions based on their topology withoutmore » the need for detailed reconstruction and outperforms algorithms currently in use by the NOvA collaboration.« less

Neural CMOS-integrated circuit and its application to data classification.

PubMed

Göknar, Izzet Cem; Yildiz, Merih; Minaei, Shahram; Deniz, Engin

2012-05-01

Implementation and new applications of a tunable complementary metal-oxide-semiconductor-integrated circuit (CMOS-IC) of a recently proposed classifier core-cell (CC) are presented and tested with two different datasets. With two algorithms-one based on Fisher's linear discriminant analysis and the other based on perceptron learning, used to obtain CCs' tunable parameters-the Haberman and Iris datasets are classified. The parameters so obtained are used for hard-classification of datasets with a neural network structured circuit. Classification performance and coefficient calculation times for both algorithms are given. The CC has 6-ns response time and 1.8-mW power consumption. The fabrication parameters used for the IC are taken from CMOS AMS 0.35-μm technology.

Neural Network and Letter Recognition.

NASA Astrophysics Data System (ADS)

Lee, Hue Yeon

Neural net architectures and learning algorithms that recognize hand written 36 alphanumeric characters are studied. The thin line input patterns written in 32 x 32 binary array are used. The system is comprised of two major components, viz. a preprocessing unit and a Recognition unit. The preprocessing unit in turn consists of three layers of neurons; the U-layer, the V-layer, and the C -layer. The functions of the U-layer is to extract local features by template matching. The correlation between the detected local features are considered. Through correlating neurons in a plane with their neighboring neurons, the V-layer would thicken the on-cells or lines that are groups of on-cells of the previous layer. These two correlations would yield some deformation tolerance and some of the rotational tolerance of the system. The C-layer then compresses data through the 'Gabor' transform. Pattern dependent choice of center and wavelengths of 'Gabor' filters is the cause of shift and scale tolerance of the system. Three different learning schemes had been investigated in the recognition unit, namely; the error back propagation learning with hidden units, a simple perceptron learning, and a competitive learning. Their performances were analyzed and compared. Since sometimes the network fails to distinguish between two letters that are inherently similar, additional ambiguity resolving neural nets are introduced on top of the above main neural net. The two dimensional Fourier transform is used as the preprocessing and the perceptron is used as the recognition unit of the ambiguity resolver. One hundred different person's handwriting sets are collected. Some of these are used as the training sets and the remainders are used as the test sets. The correct recognition rate of the system increases with the number of training sets and eventually saturates at a certain value. Similar recognition rates are obtained for the above three different learning algorithms. The minimum error rate, 4.9% is achieved for alphanumeric sets when 50 sets are trained. With the ambiguity resolver, it is reduced to 2.5%. In case that only numeral sets are trained and tested, 2.0% error rate is achieved. When only alphabet sets are considered, the error rate is reduced to 1.1%.

Joint multiple fully connected convolutional neural network with extreme learning machine for hepatocellular carcinoma nuclei grading.

PubMed

Li, Siqi; Jiang, Huiyan; Pang, Wenbo

2017-05-01

Accurate cell grading of cancerous tissue pathological image is of great importance in medical diagnosis and treatment. This paper proposes a joint multiple fully connected convolutional neural network with extreme learning machine (MFC-CNN-ELM) architecture for hepatocellular carcinoma (HCC) nuclei grading. First, in preprocessing stage, each grayscale image patch with the fixed size is obtained using center-proliferation segmentation (CPS) method and the corresponding labels are marked under the guidance of three pathologists. Next, a multiple fully connected convolutional neural network (MFC-CNN) is designed to extract the multi-form feature vectors of each input image automatically, which considers multi-scale contextual information of deep layer maps sufficiently. After that, a convolutional neural network extreme learning machine (CNN-ELM) model is proposed to grade HCC nuclei. Finally, a back propagation (BP) algorithm, which contains a new up-sample method, is utilized to train MFC-CNN-ELM architecture. The experiment comparison results demonstrate that our proposed MFC-CNN-ELM has superior performance compared with related works for HCC nuclei grading. Meanwhile, external validation using ICPR 2014 HEp-2 cell dataset shows the good generalization of our MFC-CNN-ELM architecture. Copyright © 2017 Elsevier Ltd. All rights reserved.

Processing of chromatic information in a deep convolutional neural network.

PubMed

Flachot, Alban; Gegenfurtner, Karl R

2018-04-01

Deep convolutional neural networks are a class of machine-learning algorithms capable of solving non-trivial tasks, such as object recognition, with human-like performance. Little is known about the exact computations that deep neural networks learn, and to what extent these computations are similar to the ones performed by the primate brain. Here, we investigate how color information is processed in the different layers of the AlexNet deep neural network, originally trained on object classification of over 1.2M images of objects in their natural contexts. We found that the color-responsive units in the first layer of AlexNet learned linear features and were broadly tuned to two directions in color space, analogously to what is known of color responsive cells in the primate thalamus. Moreover, these directions are decorrelated and lead to statistically efficient representations, similar to the cardinal directions of the second-stage color mechanisms in primates. We also found, in analogy to the early stages of the primate visual system, that chromatic and achromatic information were segregated in the early layers of the network. Units in the higher layers of AlexNet exhibit on average a lower responsivity for color than units at earlier stages.

Biologically inspired computation and learning in Sensorimotor Systems

NASA Astrophysics Data System (ADS)

Lee, Daniel D.; Seung, H. S.

2001-11-01

Networking systems presently lack the ability to intelligently process the rich multimedia content of the data traffic they carry. Endowing artificial systems with the ability to adapt to changing conditions requires algorithms that can rapidly learn from examples. We demonstrate the application of such learning algorithms on an inexpensive quadruped robot constructed to perform simple sensorimotor tasks. The robot learns to track a particular object by discovering the salient visual and auditory cues unique to that object. The system uses a convolutional neural network that automatically combines color, luminance, motion, and auditory information. The weights of the networks are adjusted using feedback from a teacher to reflect the reliability of the various input channels in the surrounding environment. Additionally, the robot is able to compensate for its own motion by adapting the parameters of a vestibular ocular reflex system.

Forecasting the portuguese stock market time series by using artificial neural networks

NASA Astrophysics Data System (ADS)

Isfan, Monica; Menezes, Rui; Mendes, Diana A.

2010-04-01

In this paper, we show that neural networks can be used to uncover the non-linearity that exists in the financial data. First, we follow a traditional approach by analysing the deterministic/stochastic characteristics of the Portuguese stock market data and some typical features are studied, like the Hurst exponents, among others. We also simulate a BDS test to investigate nonlinearities and the results are as expected: the financial time series do not exhibit linear dependence. Secondly, we trained four types of neural networks for the stock markets and used the models to make forecasts. The artificial neural networks were obtained using a three-layer feed-forward topology and the back-propagation learning algorithm. The quite large number of parameters that must be selected to develop a neural network forecasting model involves some trial and as a consequence the error is not small enough. In order to improve this we use a nonlinear optimization algorithm to minimize the error. Finally, the output of the 4 models is quite similar, leading to a qualitative forecast that we compare with the results of the application of k-nearest-neighbor for the same time series.

Statistical analysis and machine learning algorithms for optical biopsy

NASA Astrophysics Data System (ADS)

Wu, Binlin; Liu, Cheng-hui; Boydston-White, Susie; Beckman, Hugh; Sriramoju, Vidyasagar; Sordillo, Laura; Zhang, Chunyuan; Zhang, Lin; Shi, Lingyan; Smith, Jason; Bailin, Jacob; Alfano, Robert R.

2018-02-01

Analyzing spectral or imaging data collected with various optical biopsy methods is often times difficult due to the complexity of the biological basis. Robust methods that can utilize the spectral or imaging data and detect the characteristic spectral or spatial signatures for different types of tissue is challenging but highly desired. In this study, we used various machine learning algorithms to analyze a spectral dataset acquired from human skin normal and cancerous tissue samples using resonance Raman spectroscopy with 532nm excitation. The algorithms including principal component analysis, nonnegative matrix factorization, and autoencoder artificial neural network are used to reduce dimension of the dataset and detect features. A support vector machine with a linear kernel is used to classify the normal tissue and cancerous tissue samples. The efficacies of the methods are compared.

Machine learning based cloud mask algorithm driven by radiative transfer modeling

NASA Astrophysics Data System (ADS)

Chen, N.; Li, W.; Tanikawa, T.; Hori, M.; Shimada, R.; Stamnes, K. H.

2017-12-01

Cloud detection is a critically important first step required to derive many satellite data products. Traditional threshold based cloud mask algorithms require a complicated design process and fine tuning for each sensor, and have difficulty over snow/ice covered areas. With the advance of computational power and machine learning techniques, we have developed a new algorithm based on a neural network classifier driven by extensive radiative transfer modeling. Statistical validation results obtained by using collocated CALIOP and MODIS data show that its performance is consistent over different ecosystems and significantly better than the MODIS Cloud Mask (MOD35 C6) during the winter seasons over mid-latitude snow covered areas. Simulations using a reduced number of satellite channels also show satisfactory results, indicating its flexibility to be configured for different sensors.

A Novel Extreme Learning Machine Classification Model for e-Nose Application Based on the Multiple Kernel Approach

PubMed Central

Jian, Yulin; Huang, Daoyu; Yan, Jia; Lu, Kun; Huang, Ying; Wen, Tailai; Zeng, Tanyue; Zhong, Shijie; Xie, Qilong

2017-01-01

A novel classification model, named the quantum-behaved particle swarm optimization (QPSO)-based weighted multiple kernel extreme learning machine (QWMK-ELM), is proposed in this paper. Experimental validation is carried out with two different electronic nose (e-nose) datasets. Being different from the existing multiple kernel extreme learning machine (MK-ELM) algorithms, the combination coefficients of base kernels are regarded as external parameters of single-hidden layer feedforward neural networks (SLFNs). The combination coefficients of base kernels, the model parameters of each base kernel, and the regularization parameter are optimized by QPSO simultaneously before implementing the kernel extreme learning machine (KELM) with the composite kernel function. Four types of common single kernel functions (Gaussian kernel, polynomial kernel, sigmoid kernel, and wavelet kernel) are utilized to constitute different composite kernel functions. Moreover, the method is also compared with other existing classification methods: extreme learning machine (ELM), kernel extreme learning machine (KELM), k-nearest neighbors (KNN), support vector machine (SVM), multi-layer perceptron (MLP), radical basis function neural network (RBFNN), and probabilistic neural network (PNN). The results have demonstrated that the proposed QWMK-ELM outperforms the aforementioned methods, not only in precision, but also in efficiency for gas classification. PMID:28629202

Machine learning for outcome prediction of acute ischemic stroke post intra-arterial therapy.

PubMed

Asadi, Hamed; Dowling, Richard; Yan, Bernard; Mitchell, Peter

2014-01-01

Stroke is a major cause of death and disability. Accurately predicting stroke outcome from a set of predictive variables may identify high-risk patients and guide treatment approaches, leading to decreased morbidity. Logistic regression models allow for the identification and validation of predictive variables. However, advanced machine learning algorithms offer an alternative, in particular, for large-scale multi-institutional data, with the advantage of easily incorporating newly available data to improve prediction performance. Our aim was to design and compare different machine learning methods, capable of predicting the outcome of endovascular intervention in acute anterior circulation ischaemic stroke. We conducted a retrospective study of a prospectively collected database of acute ischaemic stroke treated by endovascular intervention. Using SPSS®, MATLAB®, and Rapidminer®, classical statistics as well as artificial neural network and support vector algorithms were applied to design a supervised machine capable of classifying these predictors into potential good and poor outcomes. These algorithms were trained, validated and tested using randomly divided data. We included 107 consecutive acute anterior circulation ischaemic stroke patients treated by endovascular technique. Sixty-six were male and the mean age of 65.3. All the available demographic, procedural and clinical factors were included into the models. The final confusion matrix of the neural network, demonstrated an overall congruency of ∼ 80% between the target and output classes, with favourable receiving operative characteristics. However, after optimisation, the support vector machine had a relatively better performance, with a root mean squared error of 2.064 (SD: ± 0.408). We showed promising accuracy of outcome prediction, using supervised machine learning algorithms, with potential for incorporation of larger multicenter datasets, likely further improving prediction. Finally, we propose that a robust machine learning system can potentially optimise the selection process for endovascular versus medical treatment in the management of acute stroke.

Can surgical simulation be used to train detection and classification of neural networks?

PubMed

Zisimopoulos, Odysseas; Flouty, Evangello; Stacey, Mark; Muscroft, Sam; Giataganas, Petros; Nehme, Jean; Chow, Andre; Stoyanov, Danail

2017-10-01

Computer-assisted interventions (CAI) aim to increase the effectiveness, precision and repeatability of procedures to improve surgical outcomes. The presence and motion of surgical tools is a key information input for CAI surgical phase recognition algorithms. Vision-based tool detection and recognition approaches are an attractive solution and can be designed to take advantage of the powerful deep learning paradigm that is rapidly advancing image recognition and classification. The challenge for such algorithms is the availability and quality of labelled data used for training. In this Letter, surgical simulation is used to train tool detection and segmentation based on deep convolutional neural networks and generative adversarial networks. The authors experiment with two network architectures for image segmentation in tool classes commonly encountered during cataract surgery. A commercially-available simulator is used to create a simulated cataract dataset for training models prior to performing transfer learning on real surgical data. To the best of authors' knowledge, this is the first attempt to train deep learning models for surgical instrument detection on simulated data while demonstrating promising results to generalise on real data. Results indicate that simulated data does have some potential for training advanced classification methods for CAI systems.

A generalized LSTM-like training algorithm for second-order recurrent neural networks

PubMed Central

Monner, Derek; Reggia, James A.

2011-01-01

The Long Short Term Memory (LSTM) is a second-order recurrent neural network architecture that excels at storing sequential short-term memories and retrieving them many time-steps later. LSTM’s original training algorithm provides the important properties of spatial and temporal locality, which are missing from other training approaches, at the cost of limiting it’s applicability to a small set of network architectures. Here we introduce the Generalized Long Short-Term Memory (LSTM-g) training algorithm, which provides LSTM-like locality while being applicable without modification to a much wider range of second-order network architectures. With LSTM-g, all units have an identical set of operating instructions for both activation and learning, subject only to the configuration of their local environment in the network; this is in contrast to the original LSTM training algorithm, where each type of unit has its own activation and training instructions. When applied to LSTM architectures with peephole connections, LSTM-g takes advantage of an additional source of back-propagated error which can enable better performance than the original algorithm. Enabled by the broad architectural applicability of LSTM-g, we demonstrate that training recurrent networks engineered for specific tasks can produce better results than single-layer networks. We conclude that LSTM-g has the potential to both improve the performance and broaden the applicability of spatially and temporally local gradient-based training algorithms for recurrent neural networks. PMID:21803542

Dreaming of Atmospheres

NASA Astrophysics Data System (ADS)

Waldmann, Ingo

2016-10-01

Radiative transfer retrievals have become the standard in modelling of exoplanetary transmission and emission spectra. Analysing currently available observations of exoplanetary atmospheres often invoke large and correlated parameter spaces that can be difficult to map or constrain.To address these issues, we have developed the Tau-REx (tau-retrieval of exoplanets) retrieval and the RobERt spectral recognition algorithms. Tau-REx is a bayesian atmospheric retrieval framework using Nested Sampling and cluster computing to fully map these large correlated parameter spaces. Nonetheless, data volumes can become prohibitively large and we must often select a subset of potential molecular/atomic absorbers in an atmosphere.In the era of open-source, automated and self-sufficient retrieval algorithms, such manual input should be avoided. User dependent input could, in worst case scenarios, lead to incomplete models and biases in the retrieval. The RobERt algorithm is build to address these issues. RobERt is a deep belief neural (DBN) networks trained to accurately recognise molecular signatures for a wide range of planets, atmospheric thermal profiles and compositions. Using these deep neural networks, we work towards retrieval algorithms that themselves understand the nature of the observed spectra, are able to learn from current and past data and make sensible qualitative preselections of atmospheric opacities to be used for the quantitative stage of the retrieval process.In this talk I will discuss how neural networks and Bayesian Nested Sampling can be used to solve highly degenerate spectral retrieval problems and what 'dreaming' neural networks can tell us about atmospheric characteristics.

Model-Free Adaptive Control for Unknown Nonlinear Zero-Sum Differential Game.

PubMed

Zhong, Xiangnan; He, Haibo; Wang, Ding; Ni, Zhen

2018-05-01

In this paper, we present a new model-free globalized dual heuristic dynamic programming (GDHP) approach for the discrete-time nonlinear zero-sum game problems. First, the online learning algorithm is proposed based on the GDHP method to solve the Hamilton-Jacobi-Isaacs equation associated with optimal regulation control problem. By setting backward one step of the definition of performance index, the requirement of system dynamics, or an identifier is relaxed in the proposed method. Then, three neural networks are established to approximate the optimal saddle point feedback control law, the disturbance law, and the performance index, respectively. The explicit updating rules for these three neural networks are provided based on the data generated during the online learning along the system trajectories. The stability analysis in terms of the neural network approximation errors is discussed based on the Lyapunov approach. Finally, two simulation examples are provided to show the effectiveness of the proposed method.

Deep Learning Methods for Improved Decoding of Linear Codes

NASA Astrophysics Data System (ADS)

Nachmani, Eliya; Marciano, Elad; Lugosch, Loren; Gross, Warren J.; Burshtein, David; Be'ery, Yair

2018-02-01

The problem of low complexity, close to optimal, channel decoding of linear codes with short to moderate block length is considered. It is shown that deep learning methods can be used to improve a standard belief propagation decoder, despite the large example space. Similar improvements are obtained for the min-sum algorithm. It is also shown that tying the parameters of the decoders across iterations, so as to form a recurrent neural network architecture, can be implemented with comparable results. The advantage is that significantly less parameters are required. We also introduce a recurrent neural decoder architecture based on the method of successive relaxation. Improvements over standard belief propagation are also observed on sparser Tanner graph representations of the codes. Furthermore, we demonstrate that the neural belief propagation decoder can be used to improve the performance, or alternatively reduce the computational complexity, of a close to optimal decoder of short BCH codes.

Classification of Parkinson's disease utilizing multi-edit nearest-neighbor and ensemble learning algorithms with speech samples.

PubMed

Zhang, He-Hua; Yang, Liuyang; Liu, Yuchuan; Wang, Pin; Yin, Jun; Li, Yongming; Qiu, Mingguo; Zhu, Xueru; Yan, Fang

2016-11-16

The use of speech based data in the classification of Parkinson disease (PD) has been shown to provide an effect, non-invasive mode of classification in recent years. Thus, there has been an increased interest in speech pattern analysis methods applicable to Parkinsonism for building predictive tele-diagnosis and tele-monitoring models. One of the obstacles in optimizing classifications is to reduce noise within the collected speech samples, thus ensuring better classification accuracy and stability. While the currently used methods are effect, the ability to invoke instance selection has been seldomly examined. In this study, a PD classification algorithm was proposed and examined that combines a multi-edit-nearest-neighbor (MENN) algorithm and an ensemble learning algorithm. First, the MENN algorithm is applied for selecting optimal training speech samples iteratively, thereby obtaining samples with high separability. Next, an ensemble learning algorithm, random forest (RF) or decorrelated neural network ensembles (DNNE), is used to generate trained samples from the collected training samples. Lastly, the trained ensemble learning algorithms are applied to the test samples for PD classification. This proposed method was examined using a more recently deposited public datasets and compared against other currently used algorithms for validation. Experimental results showed that the proposed algorithm obtained the highest degree of improved classification accuracy (29.44%) compared with the other algorithm that was examined. Furthermore, the MENN algorithm alone was found to improve classification accuracy by as much as 45.72%. Moreover, the proposed algorithm was found to exhibit a higher stability, particularly when combining the MENN and RF algorithms. This study showed that the proposed method could improve PD classification when using speech data and can be applied to future studies seeking to improve PD classification methods.

Teaching a Machine to Feel Postoperative Pain: Combining High-Dimensional Clinical Data with Machine Learning Algorithms to Forecast Acute Postoperative Pain

PubMed Central

Tighe, Patrick J.; Harle, Christopher A.; Hurley, Robert W.; Aytug, Haldun; Boezaart, Andre P.; Fillingim, Roger B.

2015-01-01

Background Given their ability to process highly dimensional datasets with hundreds of variables, machine learning algorithms may offer one solution to the vexing challenge of predicting postoperative pain. Methods Here, we report on the application of machine learning algorithms to predict postoperative pain outcomes in a retrospective cohort of 8071 surgical patients using 796 clinical variables. Five algorithms were compared in terms of their ability to forecast moderate to severe postoperative pain: Least Absolute Shrinkage and Selection Operator (LASSO), gradient-boosted decision tree, support vector machine, neural network, and k-nearest neighbor, with logistic regression included for baseline comparison. Results In forecasting moderate to severe postoperative pain for postoperative day (POD) 1, the LASSO algorithm, using all 796 variables, had the highest accuracy with an area under the receiver-operating curve (ROC) of 0.704. Next, the gradient-boosted decision tree had an ROC of 0.665 and the k-nearest neighbor algorithm had an ROC of 0.643. For POD 3, the LASSO algorithm, using all variables, again had the highest accuracy, with an ROC of 0.727. Logistic regression had a lower ROC of 0.5 for predicting pain outcomes on POD 1 and 3. Conclusions Machine learning algorithms, when combined with complex and heterogeneous data from electronic medical record systems, can forecast acute postoperative pain outcomes with accuracies similar to methods that rely only on variables specifically collected for pain outcome prediction. PMID:26031220

Variable complexity online sequential extreme learning machine, with applications to streamflow prediction

NASA Astrophysics Data System (ADS)

Lima, Aranildo R.; Hsieh, William W.; Cannon, Alex J.

2017-12-01

In situations where new data arrive continually, online learning algorithms are computationally much less costly than batch learning ones in maintaining the model up-to-date. The extreme learning machine (ELM), a single hidden layer artificial neural network with random weights in the hidden layer, is solved by linear least squares, and has an online learning version, the online sequential ELM (OSELM). As more data become available during online learning, information on the longer time scale becomes available, so ideally the model complexity should be allowed to change, but the number of hidden nodes (HN) remains fixed in OSELM. A variable complexity VC-OSELM algorithm is proposed to dynamically add or remove HN in the OSELM, allowing the model complexity to vary automatically as online learning proceeds. The performance of VC-OSELM was compared with OSELM in daily streamflow predictions at two hydrological stations in British Columbia, Canada, with VC-OSELM significantly outperforming OSELM in mean absolute error, root mean squared error and Nash-Sutcliffe efficiency at both stations.

Cerebellar supervised learning revisited: biophysical modeling and degrees-of-freedom control.

PubMed

Kawato, Mitsuo; Kuroda, Shinya; Schweighofer, Nicolas

2011-10-01

The biophysical models of spike-timing-dependent plasticity have explored dynamics with molecular basis for such computational concepts as coincidence detection, synaptic eligibility trace, and Hebbian learning. They overall support different learning algorithms in different brain areas, especially supervised learning in the cerebellum. Because a single spine is physically very small, chemical reactions at it are essentially stochastic, and thus sensitivity-longevity dilemma exists in the synaptic memory. Here, the cascade of excitable and bistable dynamics is proposed to overcome this difficulty. All kinds of learning algorithms in different brain regions confront with difficult generalization problems. For resolution of this issue, the control of the degrees-of-freedom can be realized by changing synchronicity of neural firing. Especially, for cerebellar supervised learning, the triangle closed-loop circuit consisting of Purkinje cells, the inferior olive nucleus, and the cerebellar nucleus is proposed as a circuit to optimally control synchronous firing and degrees-of-freedom in learning. Copyright © 2011 Elsevier Ltd. All rights reserved.

Proceedings of the Third International Workshop on Neural Networks and Fuzzy Logic, volume 1

NASA Technical Reports Server (NTRS)

Culbert, Christopher J. (Editor)

1993-01-01

Documented here are papers presented at the Neural Networks and Fuzzy Logic Workshop sponsored by the National Aeronautics and Space Administration and cosponsored by the University of Houston, Clear Lake. The workshop was held June 1-3, 1992 at the Lyndon B. Johnson Space Center in Houston, Texas. During the three days approximately 50 papers were presented. Technical topics addressed included adaptive systems; learning algorithms; network architectures; vision; robotics; neurobiological connections; speech recognition and synthesis; fuzzy set theory and application, control, and dynamics processing; space applications; fuzzy logic and neural network computers; approximate reasoning; and multiobject decision making.

SKYNET: an efficient and robust neural network training tool for machine learning in astronomy

NASA Astrophysics Data System (ADS)

Graff, Philip; Feroz, Farhan; Hobson, Michael P.; Lasenby, Anthony

2014-06-01

We present the first public release of our generic neural network training algorithm, called SKYNET. This efficient and robust machine learning tool is able to train large and deep feed-forward neural networks, including autoencoders, for use in a wide range of supervised and unsupervised learning applications, such as regression, classification, density estimation, clustering and dimensionality reduction. SKYNET uses a `pre-training' method to obtain a set of network parameters that has empirically been shown to be close to a good solution, followed by further optimization using a regularized variant of Newton's method, where the level of regularization is determined and adjusted automatically; the latter uses second-order derivative information to improve convergence, but without the need to evaluate or store the full Hessian matrix, by using a fast approximate method to calculate Hessian-vector products. This combination of methods allows for the training of complicated networks that are difficult to optimize using standard backpropagation techniques. SKYNET employs convergence criteria that naturally prevent overfitting, and also includes a fast algorithm for estimating the accuracy of network outputs. The utility and flexibility of SKYNET are demonstrated by application to a number of toy problems, and to astronomical problems focusing on the recovery of structure from blurred and noisy images, the identification of gamma-ray bursters, and the compression and denoising of galaxy images. The SKYNET software, which is implemented in standard ANSI C and fully parallelized using MPI, is available at http://www.mrao.cam.ac.uk/software/skynet/.

Event-driven contrastive divergence for spiking neuromorphic systems.

PubMed

Neftci, Emre; Das, Srinjoy; Pedroni, Bruno; Kreutz-Delgado, Kenneth; Cauwenberghs, Gert

2013-01-01

Restricted Boltzmann Machines (RBMs) and Deep Belief Networks have been demonstrated to perform efficiently in a variety of applications, such as dimensionality reduction, feature learning, and classification. Their implementation on neuromorphic hardware platforms emulating large-scale networks of spiking neurons can have significant advantages from the perspectives of scalability, power dissipation and real-time interfacing with the environment. However, the traditional RBM architecture and the commonly used training algorithm known as Contrastive Divergence (CD) are based on discrete updates and exact arithmetics which do not directly map onto a dynamical neural substrate. Here, we present an event-driven variation of CD to train a RBM constructed with Integrate & Fire (I&F) neurons, that is constrained by the limitations of existing and near future neuromorphic hardware platforms. Our strategy is based on neural sampling, which allows us to synthesize a spiking neural network that samples from a target Boltzmann distribution. The recurrent activity of the network replaces the discrete steps of the CD algorithm, while Spike Time Dependent Plasticity (STDP) carries out the weight updates in an online, asynchronous fashion. We demonstrate our approach by training an RBM composed of leaky I&F neurons with STDP synapses to learn a generative model of the MNIST hand-written digit dataset, and by testing it in recognition, generation and cue integration tasks. Our results contribute to a machine learning-driven approach for synthesizing networks of spiking neurons capable of carrying out practical, high-level functionality.

Event-driven contrastive divergence for spiking neuromorphic systems

PubMed Central

Neftci, Emre; Das, Srinjoy; Pedroni, Bruno; Kreutz-Delgado, Kenneth; Cauwenberghs, Gert

2014-01-01

Restricted Boltzmann Machines (RBMs) and Deep Belief Networks have been demonstrated to perform efficiently in a variety of applications, such as dimensionality reduction, feature learning, and classification. Their implementation on neuromorphic hardware platforms emulating large-scale networks of spiking neurons can have significant advantages from the perspectives of scalability, power dissipation and real-time interfacing with the environment. However, the traditional RBM architecture and the commonly used training algorithm known as Contrastive Divergence (CD) are based on discrete updates and exact arithmetics which do not directly map onto a dynamical neural substrate. Here, we present an event-driven variation of CD to train a RBM constructed with Integrate & Fire (I&F) neurons, that is constrained by the limitations of existing and near future neuromorphic hardware platforms. Our strategy is based on neural sampling, which allows us to synthesize a spiking neural network that samples from a target Boltzmann distribution. The recurrent activity of the network replaces the discrete steps of the CD algorithm, while Spike Time Dependent Plasticity (STDP) carries out the weight updates in an online, asynchronous fashion. We demonstrate our approach by training an RBM composed of leaky I&F neurons with STDP synapses to learn a generative model of the MNIST hand-written digit dataset, and by testing it in recognition, generation and cue integration tasks. Our results contribute to a machine learning-driven approach for synthesizing networks of spiking neurons capable of carrying out practical, high-level functionality. PMID:24574952

Neuroprosthetic Decoder Training as Imitation Learning

PubMed Central

Merel, Josh; Paninski, Liam; Cunningham, John P.

2016-01-01

Neuroprosthetic brain-computer interfaces function via an algorithm which decodes neural activity of the user into movements of an end effector, such as a cursor or robotic arm. In practice, the decoder is often learned by updating its parameters while the user performs a task. When the user’s intention is not directly observable, recent methods have demonstrated value in training the decoder against a surrogate for the user’s intended movement. Here we show that training a decoder in this way is a novel variant of an imitation learning problem, where an oracle or expert is employed for supervised training in lieu of direct observations, which are not available. Specifically, we describe how a generic imitation learning meta-algorithm, dataset aggregation (DAgger), can be adapted to train a generic brain-computer interface. By deriving existing learning algorithms for brain-computer interfaces in this framework, we provide a novel analysis of regret (an important metric of learning efficacy) for brain-computer interfaces. This analysis allows us to characterize the space of algorithmic variants and bounds on their regret rates. Existing approaches for decoder learning have been performed in the cursor control setting, but the available design principles for these decoders are such that it has been impossible to scale them to naturalistic settings. Leveraging our findings, we then offer an algorithm that combines imitation learning with optimal control, which should allow for training of arbitrary effectors for which optimal control can generate goal-oriented control. We demonstrate this novel and general BCI algorithm with simulated neuroprosthetic control of a 26 degree-of-freedom model of an arm, a sophisticated and realistic end effector. PMID:27191387

Segmentation of the hippocampus by transferring algorithmic knowledge for large cohort processing.

PubMed

Thyreau, Benjamin; Sato, Kazunori; Fukuda, Hiroshi; Taki, Yasuyuki

2018-01-01

The hippocampus is a particularly interesting target for neuroscience research studies due to its essential role within the human brain. In large human cohort studies, bilateral hippocampal structures are frequently identified and measured to gain insight into human behaviour or genomic variability in neuropsychiatric disorders of interest. Automatic segmentation is performed using various algorithms, with FreeSurfer being a popular option. In this manuscript, we present a method to segment the bilateral hippocampus using a deep-learned appearance model. Deep convolutional neural networks (ConvNets) have shown great success in recent years, due to their ability to learn meaningful features from a mass of training data. Our method relies on the following key novelties: (i) we use a wide and variable training set coming from multiple cohorts (ii) our training labels come in part from the output of the FreeSurfer algorithm, and (iii) we include synthetic data and use a powerful data augmentation scheme. Our method proves to be robust, and it has fast inference (<30s total per subject), with trained model available online (https://github.com/bthyreau/hippodeep). We depict illustrative results and show extensive qualitative and quantitative cohort-wide comparisons with FreeSurfer. Our work demonstrates that deep neural-network methods can easily encode, and even improve, existing anatomical knowledge, even when this knowledge exists in algorithmic form. Copyright © 2017 Elsevier B.V. All rights reserved.

Robust visual tracking based on deep convolutional neural networks and kernelized correlation filters

NASA Astrophysics Data System (ADS)

Yang, Hua; Zhong, Donghong; Liu, Chenyi; Song, Kaiyou; Yin, Zhouping

2018-03-01

Object tracking is still a challenging problem in computer vision, as it entails learning an effective model to account for appearance changes caused by occlusion, out of view, plane rotation, scale change, and background clutter. This paper proposes a robust visual tracking algorithm called deep convolutional neural network (DCNNCT) to simultaneously address these challenges. The proposed DCNNCT algorithm utilizes a DCNN to extract the image feature of a tracked target, and the full range of information regarding each convolutional layer is used to express the image feature. Subsequently, the kernelized correlation filters (CF) in each convolutional layer are adaptively learned, the correlation response maps of that are combined to estimate the location of the tracked target. To avoid the case of tracking failure, an online random ferns classifier is employed to redetect the tracked target, and a dual-threshold scheme is used to obtain the final target location by comparing the tracking result with the detection result. Finally, the change in scale of the target is determined by building scale pyramids and training a CF. Extensive experiments demonstrate that the proposed algorithm is effective at tracking, especially when evaluated using an index called the overlap rate. The DCNNCT algorithm is also highly competitive in terms of robustness with respect to state-of-the-art trackers in various challenging scenarios.

Airplane detection in remote sensing images using convolutional neural networks

NASA Astrophysics Data System (ADS)

Ouyang, Chao; Chen, Zhong; Zhang, Feng; Zhang, Yifei

2018-03-01

Airplane detection in remote sensing images remains a challenging problem and has also been taking a great interest to researchers. In this paper we propose an effective method to detect airplanes in remote sensing images using convolutional neural networks. Deep learning methods show greater advantages than the traditional methods with the rise of deep neural networks in target detection, and we give an explanation why this happens. To improve the performance on detection of airplane, we combine a region proposal algorithm with convolutional neural networks. And in the training phase, we divide the background into multi classes rather than one class, which can reduce false alarms. Our experimental results show that the proposed method is effective and robust in detecting airplane.

Machine-learning in astronomy

NASA Astrophysics Data System (ADS)

Hobson, Michael; Graff, Philip; Feroz, Farhan; Lasenby, Anthony

2014-05-01

Machine-learning methods may be used to perform many tasks required in the analysis of astronomical data, including: data description and interpretation, pattern recognition, prediction, classification, compression, inference and many more. An intuitive and well-established approach to machine learning is the use of artificial neural networks (NNs), which consist of a group of interconnected nodes, each of which processes information that it receives and then passes this product on to other nodes via weighted connections. In particular, I discuss the first public release of the generic neural network training algorithm, called SkyNet, and demonstrate its application to astronomical problems focusing on its use in the BAMBI package for accelerated Bayesian inference in cosmology, and the identification of gamma-ray bursters. The SkyNet and BAMBI packages, which are fully parallelised using MPI, are available at http://www.mrao.cam.ac.uk/software/.

Communications and control for electric power systems: Power flow classification for static security assessment

NASA Technical Reports Server (NTRS)

Niebur, D.; Germond, A.

1993-01-01

This report investigates the classification of power system states using an artificial neural network model, Kohonen's self-organizing feature map. The ultimate goal of this classification is to assess power system static security in real-time. Kohonen's self-organizing feature map is an unsupervised neural network which maps N-dimensional input vectors to an array of M neurons. After learning, the synaptic weight vectors exhibit a topological organization which represents the relationship between the vectors of the training set. This learning is unsupervised, which means that the number and size of the classes are not specified beforehand. In the application developed in this report, the input vectors used as the training set are generated by off-line load-flow simulations. The learning algorithm and the results of the organization are discussed.

Application of adaptive boosting to EP-derived multilayer feed-forward neural networks (MLFN) to improve benign/malignant breast cancer classification

NASA Astrophysics Data System (ADS)

Land, Walker H., Jr.; Masters, Timothy D.; Lo, Joseph Y.; McKee, Dan

2001-07-01

A new neural network technology was developed for improving the benign/malignant diagnosis of breast cancer using mammogram findings. A new paradigm, Adaptive Boosting (AB), uses a markedly different theory in solutioning Computational Intelligence (CI) problems. AB, a new machine learning paradigm, focuses on finding weak learning algorithm(s) that initially need to provide slightly better than random performance (i.e., approximately 55%) when processing a mammogram training set. Then, by successive development of additional architectures (using the mammogram training set), the adaptive boosting process improves the performance of the basic Evolutionary Programming derived neural network architectures. The results of these several EP-derived hybrid architectures are then intelligently combined and tested using a similar validation mammogram data set. Optimization focused on improving specificity and positive predictive value at very high sensitivities, where an analysis of the performance of the hybrid would be most meaningful. Using the DUKE mammogram database of 500 biopsy proven samples, on average this hybrid was able to achieve (under statistical 5-fold cross-validation) a specificity of 48.3% and a positive predictive value (PPV) of 51.8% while maintaining 100% sensitivity. At 97% sensitivity, a specificity of 56.6% and a PPV of 55.8% were obtained.

An immune-inspired semi-supervised algorithm for breast cancer diagnosis.

PubMed

Peng, Lingxi; Chen, Wenbin; Zhou, Wubai; Li, Fufang; Yang, Jin; Zhang, Jiandong

2016-10-01

Breast cancer is the most frequently and world widely diagnosed life-threatening cancer, which is the leading cause of cancer death among women. Early accurate diagnosis can be a big plus in treating breast cancer. Researchers have approached this problem using various data mining and machine learning techniques such as support vector machine, artificial neural network, etc. The computer immunology is also an intelligent method inspired by biological immune system, which has been successfully applied in pattern recognition, combination optimization, machine learning, etc. However, most of these diagnosis methods belong to a supervised diagnosis method. It is very expensive to obtain labeled data in biology and medicine. In this paper, we seamlessly integrate the state-of-the-art research on life science with artificial intelligence, and propose a semi-supervised learning algorithm to reduce the need for labeled data. We use two well-known benchmark breast cancer datasets in our study, which are acquired from the UCI machine learning repository. Extensive experiments are conducted and evaluated on those two datasets. Our experimental results demonstrate the effectiveness and efficiency of our proposed algorithm, which proves that our algorithm is a promising automatic diagnosis method for breast cancer. Copyright © 2016 Elsevier Ireland Ltd. All rights reserved.

Query construction, entropy, and generalization in neural-network models

NASA Astrophysics Data System (ADS)

Sollich, Peter

1994-05-01

We study query construction algorithms, which aim at improving the generalization ability of systems that learn from examples by choosing optimal, nonredundant training sets. We set up a general probabilistic framework for deriving such algorithms from the requirement of optimizing a suitable objective function; specifically, we consider the objective functions entropy (or information gain) and generalization error. For two learning scenarios, the high-low game and the linear perceptron, we evaluate the generalization performance obtained by applying the corresponding query construction algorithms and compare it to training on random examples. We find qualitative differences between the two scenarios due to the different structure of the underlying rules (nonlinear and ``noninvertible'' versus linear); in particular, for the linear perceptron, random examples lead to the same generalization ability as a sequence of queries in the limit of an infinite number of examples. We also investigate learning algorithms which are ill matched to the learning environment and find that, in this case, minimum entropy queries can in fact yield a lower generalization ability than random examples. Finally, we study the efficiency of single queries and its dependence on the learning history, i.e., on whether the previous training examples were generated randomly or by querying, and the difference between globally and locally optimal query construction.

Computer aided lung cancer diagnosis with deep learning algorithms

NASA Astrophysics Data System (ADS)

Sun, Wenqing; Zheng, Bin; Qian, Wei

2016-03-01

Deep learning is considered as a popular and powerful method in pattern recognition and classification. However, there are not many deep structured applications used in medical imaging diagnosis area, because large dataset is not always available for medical images. In this study we tested the feasibility of using deep learning algorithms for lung cancer diagnosis with the cases from Lung Image Database Consortium (LIDC) database. The nodules on each computed tomography (CT) slice were segmented according to marks provided by the radiologists. After down sampling and rotating we acquired 174412 samples with 52 by 52 pixel each and the corresponding truth files. Three deep learning algorithms were designed and implemented, including Convolutional Neural Network (CNN), Deep Belief Networks (DBNs), Stacked Denoising Autoencoder (SDAE). To compare the performance of deep learning algorithms with traditional computer aided diagnosis (CADx) system, we designed a scheme with 28 image features and support vector machine. The accuracies of CNN, DBNs, and SDAE are 0.7976, 0.8119, and 0.7929, respectively; the accuracy of our designed traditional CADx is 0.7940, which is slightly lower than CNN and DBNs. We also noticed that the mislabeled nodules using DBNs are 4% larger than using traditional CADx, this might be resulting from down sampling process lost some size information of the nodules.

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).

Engineering neural systems for high-level problem solving.

PubMed

Sylvester, Jared; Reggia, James

2016-07-01

There is a long-standing, sometimes contentious debate in AI concerning the relative merits of a symbolic, top-down approach vs. a neural, bottom-up approach to engineering intelligent machine behaviors. While neurocomputational methods excel at lower-level cognitive tasks (incremental learning for pattern classification, low-level sensorimotor control, fault tolerance and processing of noisy data, etc.), they are largely non-competitive with top-down symbolic methods for tasks involving high-level cognitive problem solving (goal-directed reasoning, metacognition, planning, etc.). Here we take a step towards addressing this limitation by developing a purely neural framework named galis. Our goal in this work is to integrate top-down (non-symbolic) control of a neural network system with more traditional bottom-up neural computations. galis is based on attractor networks that can be "programmed" with temporal sequences of hand-crafted instructions that control problem solving by gating the activity retention of, communication between, and learning done by other neural networks. We demonstrate the effectiveness of this approach by showing that it can be applied successfully to solve sequential card matching problems, using both human performance and a top-down symbolic algorithm as experimental controls. Solving this kind of problem makes use of top-down attention control and the binding together of visual features in ways that are easy for symbolic AI systems but not for neural networks to achieve. Our model can not only be instructed on how to solve card matching problems successfully, but its performance also qualitatively (and sometimes quantitatively) matches the performance of both human subjects that we had perform the same task and the top-down symbolic algorithm that we used as an experimental control. We conclude that the core principles underlying the galis framework provide a promising approach to engineering purely neurocomputational systems for problem-solving tasks that in people require higher-level cognitive functions. Copyright © 2016 Elsevier Ltd. All rights reserved.

Using neural networks and Dyna algorithm for integrated planning, reacting and learning in systems

NASA Technical Reports Server (NTRS)

Lima, Pedro; Beard, Randal

1992-01-01

The traditional AI answer to the decision making problem for a robot is planning. However, planning is usually CPU-time consuming, depending on the availability and accuracy of a world model. The Dyna system generally described in earlier work, uses trial and error to learn a world model which is simultaneously used to plan reactions resulting in optimal action sequences. It is an attempt to integrate planning, reactive, and learning systems. The architecture of Dyna is presented. The different blocks are described. There are three main components of the system. The first is the world model used by the robot for internal world representation. The input of the world model is the current state and the action taken in the current state. The output is the corresponding reward and resulting state. The second module in the system is the policy. The policy observes the current state and outputs the action to be executed by the robot. At the beginning of program execution, the policy is stochastic and through learning progressively becomes deterministic. The policy decides upon an action according to the output of an evaluation function, which is the third module of the system. The evaluation function takes the following as input: the current state of the system, the action taken in that state, the resulting state, and a reward generated by the world which is proportional to the current distance from the goal state. Originally, the work proposed was as follows: (1) to implement a simple 2-D world where a 'robot' is navigating around obstacles, to learn the path to a goal, by using lookup tables; (2) to substitute the world model and Q estimate function Q by neural networks; and (3) to apply the algorithm to a more complex world where the use of a neural network would be fully justified. In this paper, the system design and achieved results will be described. First we implement the world model with a neural network and leave Q implemented as a look up table. Next, we use a lookup table for the world model and implement the Q function with a neural net. Time limitations prevented the combination of these two approaches. The final section discusses the results and gives clues for future work.

Sensorimotor Learning Biases Choice Behavior: A Learning Neural Field Model for Decision Making

PubMed Central

Schöner, Gregor; Gail, Alexander

2012-01-01

According to a prominent view of sensorimotor processing in primates, selection and specification of possible actions are not sequential operations. Rather, a decision for an action emerges from competition between different movement plans, which are specified and selected in parallel. For action choices which are based on ambiguous sensory input, the frontoparietal sensorimotor areas are considered part of the common underlying neural substrate for selection and specification of action. These areas have been shown capable of encoding alternative spatial motor goals in parallel during movement planning, and show signatures of competitive value-based selection among these goals. Since the same network is also involved in learning sensorimotor associations, competitive action selection (decision making) should not only be driven by the sensory evidence and expected reward in favor of either action, but also by the subject's learning history of different sensorimotor associations. Previous computational models of competitive neural decision making used predefined associations between sensory input and corresponding motor output. Such hard-wiring does not allow modeling of how decisions are influenced by sensorimotor learning or by changing reward contingencies. We present a dynamic neural field model which learns arbitrary sensorimotor associations with a reward-driven Hebbian learning algorithm. We show that the model accurately simulates the dynamics of action selection with different reward contingencies, as observed in monkey cortical recordings, and that it correctly predicted the pattern of choice errors in a control experiment. With our adaptive model we demonstrate how network plasticity, which is required for association learning and adaptation to new reward contingencies, can influence choice behavior. The field model provides an integrated and dynamic account for the operations of sensorimotor integration, working memory and action selection required for decision making in ambiguous choice situations. PMID:23166483

Adaptive nodes enrich nonlinear cooperative learning beyond traditional adaptation by links.

PubMed

Sardi, Shira; Vardi, Roni; Goldental, Amir; Sheinin, Anton; Uzan, Herut; Kanter, Ido

2018-03-23

Physical models typically assume time-independent interactions, whereas neural networks and machine learning incorporate interactions that function as adjustable parameters. Here we demonstrate a new type of abundant cooperative nonlinear dynamics where learning is attributed solely to the nodes, instead of the network links which their number is significantly larger. The nodal, neuronal, fast adaptation follows its relative anisotropic (dendritic) input timings, as indicated experimentally, similarly to the slow learning mechanism currently attributed to the links, synapses. It represents a non-local learning rule, where effectively many incoming links to a node concurrently undergo the same adaptation. The network dynamics is now counterintuitively governed by the weak links, which previously were assumed to be insignificant. This cooperative nonlinear dynamic adaptation presents a self-controlled mechanism to prevent divergence or vanishing of the learning parameters, as opposed to learning by links, and also supports self-oscillations of the effective learning parameters. It hints on a hierarchical computational complexity of nodes, following their number of anisotropic inputs and opens new horizons for advanced deep learning algorithms and artificial intelligence based applications, as well as a new mechanism for enhanced and fast learning by neural networks.

Neural Network Based Intrusion Detection System for Critical Infrastructures

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

Todd Vollmer; Ondrej Linda; Milos Manic

2009-07-01

Resiliency and security in control systems such as SCADA and Nuclear plant’s in today’s world of hackers and malware are a relevant concern. Computer systems used within critical infrastructures to control physical functions are not immune to the threat of cyber attacks and may be potentially vulnerable. Tailoring an intrusion detection system to the specifics of critical infrastructures can significantly improve the security of such systems. The IDS-NNM – Intrusion Detection System using Neural Network based Modeling, is presented in this paper. The main contributions of this work are: 1) the use and analyses of real network data (data recordedmore » from an existing critical infrastructure); 2) the development of a specific window based feature extraction technique; 3) the construction of training dataset using randomly generated intrusion vectors; 4) the use of a combination of two neural network learning algorithms – the Error-Back Propagation and Levenberg-Marquardt, for normal behavior modeling. The presented algorithm was evaluated on previously unseen network data. The IDS-NNM algorithm proved to be capable of capturing all intrusion attempts presented in the network communication while not generating any false alerts.« less

Application of machine learning techniques to analyse the effects of physical exercise in ventricular fibrillation.

PubMed

Caravaca, Juan; Soria-Olivas, Emilio; Bataller, Manuel; Serrano, Antonio J; Such-Miquel, Luis; Vila-Francés, Joan; Guerrero, Juan F

2014-02-01

This work presents the application of machine learning techniques to analyse the influence of physical exercise in the physiological properties of the heart, during ventricular fibrillation. To this end, different kinds of classifiers (linear and neural models) are used to classify between trained and sedentary rabbit hearts. The use of those classifiers in combination with a wrapper feature selection algorithm allows to extract knowledge about the most relevant features in the problem. The obtained results show that neural models outperform linear classifiers (better performance indices and a better dimensionality reduction). The most relevant features to describe the benefits of physical exercise are those related to myocardial heterogeneity, mean activation rate and activation complexity. © 2013 Published by Elsevier Ltd.

"Who" is saying "what"? Brain-based decoding of human voice and speech.

PubMed

Formisano, Elia; De Martino, Federico; Bonte, Milene; Goebel, Rainer

2008-11-07

Can we decipher speech content ("what" is being said) and speaker identity ("who" is saying it) from observations of brain activity of a listener? Here, we combine functional magnetic resonance imaging with a data-mining algorithm and retrieve what and whom a person is listening to from the neural fingerprints that speech and voice signals elicit in the listener's auditory cortex. These cortical fingerprints are spatially distributed and insensitive to acoustic variations of the input so as to permit the brain-based recognition of learned speech from unknown speakers and of learned voices from previously unheard utterances. Our findings unravel the detailed cortical layout and computational properties of the neural populations at the basis of human speech recognition and speaker identification.

Optimizing Virtual Network Functions Placement in Virtual Data Center Infrastructure Using Machine Learning

NASA Astrophysics Data System (ADS)

Bolodurina, I. P.; Parfenov, D. I.

2018-01-01

We have elaborated a neural network model of virtual network flow identification based on the statistical properties of flows circulating in the network of the data center and characteristics that describe the content of packets transmitted through network objects. This enabled us to establish the optimal set of attributes to identify virtual network functions. We have established an algorithm for optimizing the placement of virtual data functions using the data obtained in our research. Our approach uses a hybrid method of visualization using virtual machines and containers, which enables to reduce the infrastructure load and the response time in the network of the virtual data center. The algorithmic solution is based on neural networks, which enables to scale it at any number of the network function copies.

Fast converging minimum probability of error neural network receivers for DS-CDMA communications.

PubMed

Matyjas, John D; Psaromiligkos, Ioannis N; Batalama, Stella N; Medley, Michael J

2004-03-01

We consider a multilayer perceptron neural network (NN) receiver architecture for the recovery of the information bits of a direct-sequence code-division-multiple-access (DS-CDMA) user. We develop a fast converging adaptive training algorithm that minimizes the bit-error rate (BER) at the output of the receiver. The adaptive algorithm has three key features: i) it incorporates the BER, i.e., the ultimate performance evaluation measure, directly into the learning process, ii) it utilizes constraints that are derived from the properties of the optimum single-user decision boundary for additive white Gaussian noise (AWGN) multiple-access channels, and iii) it embeds importance sampling (IS) principles directly into the receiver optimization process. Simulation studies illustrate the BER performance of the proposed scheme.

Software tool for data mining and its applications

NASA Astrophysics Data System (ADS)

Yang, Jie; Ye, Chenzhou; Chen, Nianyi

2002-03-01

A software tool for data mining is introduced, which integrates pattern recognition (PCA, Fisher, clustering, hyperenvelop, regression), artificial intelligence (knowledge representation, decision trees), statistical learning (rough set, support vector machine), computational intelligence (neural network, genetic algorithm, fuzzy systems). It consists of nine function models: pattern recognition, decision trees, association rule, fuzzy rule, neural network, genetic algorithm, Hyper Envelop, support vector machine, visualization. The principle and knowledge representation of some function models of data mining are described. The software tool of data mining is realized by Visual C++ under Windows 2000. Nonmonotony in data mining is dealt with by concept hierarchy and layered mining. The software tool of data mining has satisfactorily applied in the prediction of regularities of the formation of ternary intermetallic compounds in alloy systems, and diagnosis of brain glioma.

Deep neural networks to enable real-time multimessenger astrophysics

NASA Astrophysics Data System (ADS)

George, Daniel; Huerta, E. A.

2018-02-01

Gravitational wave astronomy has set in motion a scientific revolution. To further enhance the science reach of this emergent field of research, there is a pressing need to increase the depth and speed of the algorithms used to enable these ground-breaking discoveries. We introduce Deep Filtering—a new scalable machine learning method for end-to-end time-series signal processing. Deep Filtering is based on deep learning with two deep convolutional neural networks, which are designed for classification and regression, to detect gravitational wave signals in highly noisy time-series data streams and also estimate the parameters of their sources in real time. Acknowledging that some of the most sensitive algorithms for the detection of gravitational waves are based on implementations of matched filtering, and that a matched filter is the optimal linear filter in Gaussian noise, the application of Deep Filtering using whitened signals in Gaussian noise is investigated in this foundational article. The results indicate that Deep Filtering outperforms conventional machine learning techniques, achieves similar performance compared to matched filtering, while being several orders of magnitude faster, allowing real-time signal processing with minimal resources. Furthermore, we demonstrate that Deep Filtering can detect and characterize waveform signals emitted from new classes of eccentric or spin-precessing binary black holes, even when trained with data sets of only quasicircular binary black hole waveforms. The results presented in this article, and the recent use of deep neural networks for the identification of optical transients in telescope data, suggests that deep learning can facilitate real-time searches of gravitational wave sources and their electromagnetic and astroparticle counterparts. In the subsequent article, the framework introduced herein is directly applied to identify and characterize gravitational wave events in real LIGO data.

Self-Adaptive Prediction of Cloud Resource Demands Using Ensemble Model and Subtractive-Fuzzy Clustering Based Fuzzy Neural Network

PubMed Central

Chen, Zhijia; Zhu, Yuanchang; Di, Yanqiang; Feng, Shaochong

2015-01-01

In IaaS (infrastructure as a service) cloud environment, users are provisioned with virtual machines (VMs). To allocate resources for users dynamically and effectively, accurate resource demands predicting is essential. For this purpose, this paper proposes a self-adaptive prediction method using ensemble model and subtractive-fuzzy clustering based fuzzy neural network (ESFCFNN). We analyze the characters of user preferences and demands. Then the architecture of the prediction model is constructed. We adopt some base predictors to compose the ensemble model. Then the structure and learning algorithm of fuzzy neural network is researched. To obtain the number of fuzzy rules and the initial value of the premise and consequent parameters, this paper proposes the fuzzy c-means combined with subtractive clustering algorithm, that is, the subtractive-fuzzy clustering. Finally, we adopt different criteria to evaluate the proposed method. The experiment results show that the method is accurate and effective in predicting the resource demands. PMID:25691896

Models of Acetylcholine and Dopamine Signals Differentially Improve Neural Representations

PubMed Central

Holca-Lamarre, Raphaël; Lücke, Jörg; Obermayer, Klaus

2017-01-01

Biological and artificial neural networks (ANNs) represent input signals as patterns of neural activity. In biology, neuromodulators can trigger important reorganizations of these neural representations. For instance, pairing a stimulus with the release of either acetylcholine (ACh) or dopamine (DA) evokes long lasting increases in the responses of neurons to the paired stimulus. The functional roles of ACh and DA in rearranging representations remain largely unknown. Here, we address this question using a Hebbian-learning neural network model. Our aim is both to gain a functional understanding of ACh and DA transmission in shaping biological representations and to explore neuromodulator-inspired learning rules for ANNs. We model the effects of ACh and DA on synaptic plasticity and confirm that stimuli coinciding with greater neuromodulator activation are over represented in the network. We then simulate the physiological release schedules of ACh and DA. We measure the impact of neuromodulator release on the network's representation and on its performance on a classification task. We find that ACh and DA trigger distinct changes in neural representations that both improve performance. The putative ACh signal redistributes neural preferences so that more neurons encode stimulus classes that are challenging for the network. The putative DA signal adapts synaptic weights so that they better match the classes of the task at hand. Our model thus offers a functional explanation for the effects of ACh and DA on cortical representations. Additionally, our learning algorithm yields performances comparable to those of state-of-the-art optimisation methods in multi-layer perceptrons while requiring weaker supervision signals and interacting with synaptically-local weight updates. PMID:28690509

Machine learning methods in chemoinformatics

PubMed Central

Mitchell, John B O

2014-01-01

Machine learning algorithms are generally developed in computer science or adjacent disciplines and find their way into chemical modeling by a process of diffusion. Though particular machine learning methods are popular in chemoinformatics and quantitative structure–activity relationships (QSAR), many others exist in the technical literature. This discussion is methods-based and focused on some algorithms that chemoinformatics researchers frequently use. It makes no claim to be exhaustive. We concentrate on methods for supervised learning, predicting the unknown property values of a test set of instances, usually molecules, based on the known values for a training set. Particularly relevant approaches include Artificial Neural Networks, Random Forest, Support Vector Machine, k-Nearest Neighbors and naïve Bayes classifiers. WIREs Comput Mol Sci 2014, 4:468–481. How to cite this article: WIREs Comput Mol Sci 2014, 4:468–481. doi:10.1002/wcms.1183 PMID:25285160

Boosting Learning Algorithm for Stock Price Forecasting

NASA Astrophysics Data System (ADS)

Wang, Chengzhang; Bai, Xiaoming

2018-03-01

To tackle complexity and uncertainty of stock market behavior, more studies have introduced machine learning algorithms to forecast stock price. ANN (artificial neural network) is one of the most successful and promising applications. We propose a boosting-ANN model in this paper to predict the stock close price. On the basis of boosting theory, multiple weak predicting machines, i.e. ANNs, are assembled to build a stronger predictor, i.e. boosting-ANN model. New error criteria of the weak studying machine and rules of weights updating are adopted in this study. We select technical factors from financial markets as forecasting input variables. Final results demonstrate the boosting-ANN model works better than other ones for stock price forecasting.

Pileup Mitigation with Machine Learning (PUMML)

NASA Astrophysics Data System (ADS)

Komiske, Patrick T.; Metodiev, Eric M.; Nachman, Benjamin; Schwartz, Matthew D.

2017-12-01

Pileup involves the contamination of the energy distribution arising from the primary collision of interest (leading vertex) by radiation from soft collisions (pileup). We develop a new technique for removing this contamination using machine learning and convolutional neural networks. The network takes as input the energy distribution of charged leading vertex particles, charged pileup particles, and all neutral particles and outputs the energy distribution of particles coming from leading vertex alone. The PUMML algorithm performs remarkably well at eliminating pileup distortion on a wide range of simple and complex jet observables. We test the robustness of the algorithm in a number of ways and discuss how the network can be trained directly on data.

An assessment of support vector machines for land cover classification

USGS Publications Warehouse

Huang, C.; Davis, L.S.; Townshend, J.R.G.

2002-01-01

The support vector machine (SVM) is a group of theoretically superior machine learning algorithms. It was found competitive with the best available machine learning algorithms in classifying high-dimensional data sets. This paper gives an introduction to the theoretical development of the SVM and an experimental evaluation of its accuracy, stability and training speed in deriving land cover classifications from satellite images. The SVM was compared to three other popular classifiers, including the maximum likelihood classifier (MLC), neural network classifiers (NNC) and decision tree classifiers (DTC). The impacts of kernel configuration on the performance of the SVM and of the selection of training data and input variables on the four classifiers were also evaluated in this experiment.

ChemNet: A Transferable and Generalizable Deep Neural Network for Small-Molecule Property Prediction

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

Goh, Garrett B.; Siegel, Charles M.; Vishnu, Abhinav

With access to large datasets, deep neural networks through representation learning have been able to identify patterns from raw data, achieving human-level accuracy in image and speech recognition tasks. However, in chemistry, availability of large standardized and labelled datasets is scarce, and with a multitude of chemical properties of interest, chemical data is inherently small and fragmented. In this work, we explore transfer learning techniques in conjunction with the existing Chemception CNN model, to create a transferable and generalizable deep neural network for small-molecule property prediction. Our latest model, ChemNet learns in a semi-supervised manner from inexpensive labels computed frommore » the ChEMBL database. When fine-tuned to the Tox21, HIV and FreeSolv dataset, which are 3 separate chemical tasks that ChemNet was not originally trained on, we demonstrate that ChemNet exceeds the performance of existing Chemception models, contemporary MLP models that trains on molecular fingerprints, and it matches the performance of the ConvGraph algorithm, the current state-of-the-art. Furthermore, as ChemNet has been pre-trained on a large diverse chemical database, it can be used as a universal “plug-and-play” deep neural network, which accelerates the deployment of deep neural networks for the prediction of novel small-molecule chemical properties.« less

Deep 3D convolution neural network for CT brain hemorrhage classification

NASA Astrophysics Data System (ADS)

Jnawali, Kamal; Arbabshirani, Mohammad R.; Rao, Navalgund; Patel, Alpen A.

2018-02-01

Intracranial hemorrhage is a critical conditional with the high mortality rate that is typically diagnosed based on head computer tomography (CT) images. Deep learning algorithms, in particular, convolution neural networks (CNN), are becoming the methodology of choice in medical image analysis for a variety of applications such as computer-aided diagnosis, and segmentation. In this study, we propose a fully automated deep learning framework which learns to detect brain hemorrhage based on cross sectional CT images. The dataset for this work consists of 40,367 3D head CT studies (over 1.5 million 2D images) acquired retrospectively over a decade from multiple radiology facilities at Geisinger Health System. The proposed algorithm first extracts features using 3D CNN and then detects brain hemorrhage using the logistic function as the last layer of the network. Finally, we created an ensemble of three different 3D CNN architectures to improve the classification accuracy. The area under the curve (AUC) of the receiver operator characteristic (ROC) curve of the ensemble of three architectures was 0.87. Their results are very promising considering the fact that the head CT studies were not controlled for slice thickness, scanner type, study protocol or any other settings. Moreover, the proposed algorithm reliably detected various types of hemorrhage within the skull. This work is one of the first applications of 3D CNN trained on a large dataset of cross sectional medical images for detection of a critical radiological condition

A CNN based neurobiology inspired approach for retinal image quality assessment.

PubMed

Mahapatra, Dwarikanath; Roy, Pallab K; Sedai, Suman; Garnavi, Rahil

2016-08-01

Retinal image quality assessment (IQA) algorithms use different hand crafted features for training classifiers without considering the working of the human visual system (HVS) which plays an important role in IQA. We propose a convolutional neural network (CNN) based approach that determines image quality using the underlying principles behind the working of the HVS. CNNs provide a principled approach to feature learning and hence higher accuracy in decision making. Experimental results demonstrate the superior performance of our proposed algorithm over competing methods.

Evolutionary and Neural Computing Based Decision Support System for Disease Diagnosis from Clinical Data Sets in Medical Practice.

PubMed

Sudha, M

2017-09-27

As a recent trend, various computational intelligence and machine learning approaches have been used for mining inferences hidden in the large clinical databases to assist the clinician in strategic decision making. In any target data the irrelevant information may be detrimental, causing confusion for the mining algorithm and degrades the prediction outcome. To address this issue, this study attempts to identify an intelligent approach to assist disease diagnostic procedure using an optimal set of attributes instead of all attributes present in the clinical data set. In this proposed Application Specific Intelligent Computing (ASIC) decision support system, a rough set based genetic algorithm is employed in pre-processing phase and a back propagation neural network is applied in training and testing phase. ASIC has two phases, the first phase handles outliers, noisy data, and missing values to obtain a qualitative target data to generate appropriate attribute reduct sets from the input data using rough computing based genetic algorithm centred on a relative fitness function measure. The succeeding phase of this system involves both training and testing of back propagation neural network classifier on the selected reducts. The model performance is evaluated with widely adopted existing classifiers. The proposed ASIC system for clinical decision support has been tested with breast cancer, fertility diagnosis and heart disease data set from the University of California at Irvine (UCI) machine learning repository. The proposed system outperformed the existing approaches attaining the accuracy rate of 95.33%, 97.61%, and 93.04% for breast cancer, fertility issue and heart disease diagnosis.

Using Active Learning for Speeding up Calibration in Simulation Models.

PubMed

Cevik, Mucahit; Ergun, Mehmet Ali; Stout, Natasha K; Trentham-Dietz, Amy; Craven, Mark; Alagoz, Oguzhan

2016-07-01

Most cancer simulation models include unobservable parameters that determine disease onset and tumor growth. These parameters play an important role in matching key outcomes such as cancer incidence and mortality, and their values are typically estimated via a lengthy calibration procedure, which involves evaluating a large number of combinations of parameter values via simulation. The objective of this study is to demonstrate how machine learning approaches can be used to accelerate the calibration process by reducing the number of parameter combinations that are actually evaluated. Active learning is a popular machine learning method that enables a learning algorithm such as artificial neural networks to interactively choose which parameter combinations to evaluate. We developed an active learning algorithm to expedite the calibration process. Our algorithm determines the parameter combinations that are more likely to produce desired outputs and therefore reduces the number of simulation runs performed during calibration. We demonstrate our method using the previously developed University of Wisconsin breast cancer simulation model (UWBCS). In a recent study, calibration of the UWBCS required the evaluation of 378 000 input parameter combinations to build a race-specific model, and only 69 of these combinations produced results that closely matched observed data. By using the active learning algorithm in conjunction with standard calibration methods, we identify all 69 parameter combinations by evaluating only 5620 of the 378 000 combinations. Machine learning methods hold potential in guiding model developers in the selection of more promising parameter combinations and hence speeding up the calibration process. Applying our machine learning algorithm to one model shows that evaluating only 1.49% of all parameter combinations would be sufficient for the calibration. © The Author(s) 2015.

Using Active Learning for Speeding up Calibration in Simulation Models

PubMed Central

Cevik, Mucahit; Ali Ergun, Mehmet; Stout, Natasha K.; Trentham-Dietz, Amy; Craven, Mark; Alagoz, Oguzhan

2015-01-01

Background Most cancer simulation models include unobservable parameters that determine the disease onset and tumor growth. These parameters play an important role in matching key outcomes such as cancer incidence and mortality and their values are typically estimated via lengthy calibration procedure, which involves evaluating large number of combinations of parameter values via simulation. The objective of this study is to demonstrate how machine learning approaches can be used to accelerate the calibration process by reducing the number of parameter combinations that are actually evaluated. Methods Active learning is a popular machine learning method that enables a learning algorithm such as artificial neural networks to interactively choose which parameter combinations to evaluate. We develop an active learning algorithm to expedite the calibration process. Our algorithm determines the parameter combinations that are more likely to produce desired outputs, therefore reduces the number of simulation runs performed during calibration. We demonstrate our method using previously developed University of Wisconsin Breast Cancer Simulation Model (UWBCS). Results In a recent study, calibration of the UWBCS required the evaluation of 378,000 input parameter combinations to build a race-specific model and only 69 of these combinations produced results that closely matched observed data. By using the active learning algorithm in conjunction with standard calibration methods, we identify all 69 parameter combinations by evaluating only 5620 of the 378,000 combinations. Conclusion Machine learning methods hold potential in guiding model developers in the selection of more promising parameter combinations and hence speeding up the calibration process. Applying our machine learning algorithm to one model shows that evaluating only 1.49% of all parameter combinations would be sufficient for the calibration. PMID:26471190

Projection decomposition algorithm for dual-energy computed tomography via deep neural network.

PubMed

Xu, Yifu; Yan, Bin; Chen, Jian; Zeng, Lei; Li, Lei

2018-03-15

Dual-energy computed tomography (DECT) has been widely used to improve identification of substances from different spectral information. Decomposition of the mixed test samples into two materials relies on a well-calibrated material decomposition function. This work aims to establish and validate a data-driven algorithm for estimation of the decomposition function. A deep neural network (DNN) consisting of two sub-nets is proposed to solve the projection decomposition problem. The compressing sub-net, substantially a stack auto-encoder (SAE), learns a compact representation of energy spectrum. The decomposing sub-net with a two-layer structure fits the nonlinear transform between energy projection and basic material thickness. The proposed DNN not only delivers image with lower standard deviation and higher quality in both simulated and real data, and also yields the best performance in cases mixed with photon noise. Moreover, DNN costs only 0.4 s to generate a decomposition solution of 360 × 512 size scale, which is about 200 times faster than the competing algorithms. The DNN model is applicable to the decomposition tasks with different dual energies. Experimental results demonstrated the strong function fitting ability of DNN. Thus, the Deep learning paradigm provides a promising approach to solve the nonlinear problem in DECT.

An automated diagnosis system of liver disease using artificial immune and genetic algorithms.

PubMed

Liang, Chunlin; Peng, Lingxi

2013-04-01

The rise of health care cost is one of the world's most important problems. Disease prediction is also a vibrant research area. Researchers have approached this problem using various techniques such as support vector machine, artificial neural network, etc. This study typically exploits the immune system's characteristics of learning and memory to solve the problem of liver disease diagnosis. The proposed system applies a combination of two methods of artificial immune and genetic algorithm to diagnose the liver disease. The system architecture is based on artificial immune system. The learning procedure of system adopts genetic algorithm to interfere the evolution of antibody population. The experiments use two benchmark datasets in our study, which are acquired from the famous UCI machine learning repository. The obtained diagnosis accuracies are very promising with regard to the other diagnosis system in the literatures. These results suggest that this system may be a useful automatic diagnosis tool for liver disease.

Paradigms for machine learning

NASA Technical Reports Server (NTRS)

Schlimmer, Jeffrey C.; Langley, Pat

1991-01-01

Five paradigms are described for machine learning: connectionist (neural network) methods, genetic algorithms and classifier systems, empirical methods for inducing rules and decision trees, analytic learning methods, and case-based approaches. Some dimensions are considered along with these paradigms vary in their approach to learning, and the basic methods are reviewed that are used within each framework, together with open research issues. It is argued that the similarities among the paradigms are more important than their differences, and that future work should attempt to bridge the existing boundaries. Finally, some recent developments in the field of machine learning are discussed, and their impact on both research and applications is examined.

Application of a hybrid model of neural networks and genetic algorithms to evaluate landslide susceptibility

NASA Astrophysics Data System (ADS)

Wang, H. B.; Li, J. W.; Zhou, B.; Yuan, Z. Q.; Chen, Y. P.

2013-03-01

In the last few decades, the development of Geographical Information Systems (GIS) technology has provided a method for the evaluation of landslide susceptibility and hazard. Slope units were found to be appropriate for the fundamental morphological elements in landslide susceptibility evaluation. Following the DEM construction in a loess area susceptible to landslides, the direct-reverse DEM technology was employed to generate 216 slope units in the studied area. After a detailed investigation, the landslide inventory was mapped in which 39 landslides, including paleo-landslides, old landslides and recent landslides, were present. Of the 216 slope units, 123 involved landslides. To analyze the mechanism of these landslides, six environmental factors were selected to evaluate landslide occurrence: slope angle, aspect, the height and shape of the slope, distance to river and human activities. These factors were extracted in terms of the slope unit within the ArcGIS software. The spatial analysis demonstrates that most of the landslides are located on convex slopes at an elevation of 100-150 m with slope angles from 135°-225° and 40°-60°. Landslide occurrence was then checked according to these environmental factors using an artificial neural network with back propagation, optimized by genetic algorithms. A dataset of 120 slope units was chosen for training the neural network model, i.e., 80 units with landslide presence and 40 units without landslide presence. The parameters of genetic algorithms and neural networks were then set: population size of 100, crossover probability of 0.65, mutation probability of 0.01, momentum factor of 0.60, learning rate of 0.7, max learning number of 10 000, and target error of 0.000001. After training on the datasets, the susceptibility of landslides was mapped for the land-use plan and hazard mitigation. Comparing the susceptibility map with landslide inventory, it was noted that the prediction accuracy of landslide occurrence is 93.02%, whereas units without landslide occurrence are predicted with an accuracy of 81.13%. To sum up, the verification shows satisfactory agreement with an accuracy of 86.46% between the susceptibility map and the landslide locations. In the landslide susceptibility assessment, ten new slopes were predicted to show potential for failure, which can be confirmed by the engineering geological conditions of these slopes. It was also observed that some disadvantages could be overcome in the application of the neural networks with back propagation, for example, the low convergence rate and local minimum, after the network was optimized using genetic algorithms. To conclude, neural networks with back propagation that are optimized by genetic algorithms are an effective method to predict landslide susceptibility with high accuracy.

Automatic Classification of volcano-seismic events based on Deep Neural Networks.

NASA Astrophysics Data System (ADS)

Titos Luzón, M.; Bueno Rodriguez, A.; Garcia Martinez, L.; Benitez, C.; Ibáñez, J. M.

2017-12-01

Seismic monitoring of active volcanoes is a popular remote sensing technique to detect seismic activity, often associated to energy exchanges between the volcano and the environment. As a result, seismographs register a wide range of volcano-seismic signals that reflect the nature and underlying physics of volcanic processes. Machine learning and signal processing techniques provide an appropriate framework to analyze such data. In this research, we propose a new classification framework for seismic events based on deep neural networks. Deep neural networks are composed by multiple processing layers, and can discover intrinsic patterns from the data itself. Internal parameters can be initialized using a greedy unsupervised pre-training stage, leading to an efficient training of fully connected architectures. We aim to determine the robustness of these architectures as classifiers of seven different types of seismic events recorded at "Volcán de Fuego" (Colima, Mexico). Two deep neural networks with different pre-training strategies are studied: stacked denoising autoencoder and deep belief networks. Results are compared to existing machine learning algorithms (SVM, Random Forest, Multilayer Perceptron). We used 5 LPC coefficients over three non-overlapping segments as training features in order to characterize temporal evolution, avoid redundancy and encode the signal, regardless of its duration. Experimental results show that deep architectures can classify seismic events with higher accuracy than classical algorithms, attaining up to 92% recognition accuracy. Pre-training initialization helps these models to detect events that occur simultaneously in time (such explosions and rockfalls), increase robustness against noisy inputs, and provide better generalization. These results demonstrate deep neural networks are robust classifiers, and can be deployed in real-environments to monitor the seismicity of restless volcanoes.

The Next Era: Deep Learning in Pharmaceutical Research.

PubMed

Ekins, Sean

2016-11-01

Over the past decade we have witnessed the increasing sophistication of machine learning algorithms applied in daily use from internet searches, voice recognition, social network software to machine vision software in cameras, phones, robots and self-driving cars. Pharmaceutical research has also seen its fair share of machine learning developments. For example, applying such methods to mine the growing datasets that are created in drug discovery not only enables us to learn from the past but to predict a molecule's properties and behavior in future. The latest machine learning algorithm garnering significant attention is deep learning, which is an artificial neural network with multiple hidden layers. Publications over the last 3 years suggest that this algorithm may have advantages over previous machine learning methods and offer a slight but discernable edge in predictive performance. The time has come for a balanced review of this technique but also to apply machine learning methods such as deep learning across a wider array of endpoints relevant to pharmaceutical research for which the datasets are growing such as physicochemical property prediction, formulation prediction, absorption, distribution, metabolism, excretion and toxicity (ADME/Tox), target prediction and skin permeation, etc. We also show that there are many potential applications of deep learning beyond cheminformatics. It will be important to perform prospective testing (which has been carried out rarely to date) in order to convince skeptics that there will be benefits from investing in this technique.

A genetic algorithm for optimization of neural network capable of learning to search for food in a maze

NASA Astrophysics Data System (ADS)

Budilova, E. V.; Terekhin, A. T.; Chepurnov, S. A.

1994-09-01

A hypothetical neural scheme is proposed that ensures efficient decision making by an animal searching for food in a maze. Only the general structure of the network is fixed; its quantitative characteristics are found by numerical optimization that simulates the process of natural selection. Selection is aimed at maximization of the expected number of descendants, which is directly related to the energy stored during the reproductive cycle. The main parameters to be optimized are the increments of the interneuronal links and the working-memory constants.

Learning, epigenetics, and computation: An extension on Fitch's proposal. Comment on “Toward a computational framework for cognitive biology: Unifying approaches from cognitive neuroscience and comparative cognition” by W. Tecumseh Fitch

NASA Astrophysics Data System (ADS)

Okanoya, Kazuo

2014-09-01

The comparative computational approach of Fitch [1] attempts to renew the classical David Marr paradigm of computation, algorithm, and implementation, by introducing evolutionary view of the relationship between neural architecture and cognition. This comparative evolutionary view provides constraints useful in narrowing down the problem space for both cognition and neural mechanisms. I will provide two examples from our own studies that reinforce and extend Fitch's proposal.

Stable architectures for deep neural networks

NASA Astrophysics Data System (ADS)

Haber, Eldad; Ruthotto, Lars

2018-01-01

Deep neural networks have become invaluable tools for supervised machine learning, e.g. classification of text or images. While often offering superior results over traditional techniques and successfully expressing complicated patterns in data, deep architectures are known to be challenging to design and train such that they generalize well to new data. Critical issues with deep architectures are numerical instabilities in derivative-based learning algorithms commonly called exploding or vanishing gradients. In this paper, we propose new forward propagation techniques inspired by systems of ordinary differential equations (ODE) that overcome this challenge and lead to well-posed learning problems for arbitrarily deep networks. The backbone of our approach is our interpretation of deep learning as a parameter estimation problem of nonlinear dynamical systems. Given this formulation, we analyze stability and well-posedness of deep learning and use this new understanding to develop new network architectures. We relate the exploding and vanishing gradient phenomenon to the stability of the discrete ODE and present several strategies for stabilizing deep learning for very deep networks. While our new architectures restrict the solution space, several numerical experiments show their competitiveness with state-of-the-art networks.

Smoothing strategies combined with ARIMA and neural networks to improve the forecasting of traffic accidents.

PubMed

Barba, Lida; Rodríguez, Nibaldo; Montt, Cecilia

2014-01-01

Two smoothing strategies combined with autoregressive integrated moving average (ARIMA) and autoregressive neural networks (ANNs) models to improve the forecasting of time series are presented. The strategy of forecasting is implemented using two stages. In the first stage the time series is smoothed using either, 3-point moving average smoothing, or singular value Decomposition of the Hankel matrix (HSVD). In the second stage, an ARIMA model and two ANNs for one-step-ahead time series forecasting are used. The coefficients of the first ANN are estimated through the particle swarm optimization (PSO) learning algorithm, while the coefficients of the second ANN are estimated with the resilient backpropagation (RPROP) learning algorithm. The proposed models are evaluated using a weekly time series of traffic accidents of Valparaíso, Chilean region, from 2003 to 2012. The best result is given by the combination HSVD-ARIMA, with a MAPE of 0:26%, followed by MA-ARIMA with a MAPE of 1:12%; the worst result is given by the MA-ANN based on PSO with a MAPE of 15:51%.

Siamese convolutional networks for tracking the spine motion

NASA Astrophysics Data System (ADS)

Liu, Yuan; Sui, Xiubao; Sun, Yicheng; Liu, Chengwei; Hu, Yong

2017-09-01

Deep learning models have demonstrated great success in various computer vision tasks such as image classification and object tracking. However, tracking the lumbar spine by digitalized video fluoroscopic imaging (DVFI), which can quantitatively analyze the motion mode of spine to diagnose lumbar instability, has not yet been well developed due to the lack of steady and robust tracking method. In this paper, we propose a novel visual tracking algorithm of the lumbar vertebra motion based on a Siamese convolutional neural network (CNN) model. We train a full-convolutional neural network offline to learn generic image features. The network is trained to learn a similarity function that compares the labeled target in the first frame with the candidate patches in the current frame. The similarity function returns a high score if the two images depict the same object. Once learned, the similarity function is used to track a previously unseen object without any adapting online. In the current frame, our tracker is performed by evaluating the candidate rotated patches sampled around the previous frame target position and presents a rotated bounding box to locate the predicted target precisely. Results indicate that the proposed tracking method can detect the lumbar vertebra steadily and robustly. Especially for images with low contrast and cluttered background, the presented tracker can still achieve good tracking performance. Further, the proposed algorithm operates at high speed for real time tracking.

Reconstructing Genetic Regulatory Networks Using Two-Step Algorithms with the Differential Equation Models of Neural Networks.

PubMed

Chen, Chi-Kan

2017-07-26

The identification of genetic regulatory networks (GRNs) provides insights into complex cellular processes. A class of recurrent neural networks (RNNs) captures the dynamics of GRN. Algorithms combining the RNN and machine learning schemes were proposed to reconstruct small-scale GRNs using gene expression time series. We present new GRN reconstruction methods with neural networks. The RNN is extended to a class of recurrent multilayer perceptrons (RMLPs) with latent nodes. Our methods contain two steps: the edge rank assignment step and the network construction step. The former assigns ranks to all possible edges by a recursive procedure based on the estimated weights of wires of RNN/RMLP (RE RNN /RE RMLP ), and the latter constructs a network consisting of top-ranked edges under which the optimized RNN simulates the gene expression time series. The particle swarm optimization (PSO) is applied to optimize the parameters of RNNs and RMLPs in a two-step algorithm. The proposed RE RNN -RNN and RE RMLP -RNN algorithms are tested on synthetic and experimental gene expression time series of small GRNs of about 10 genes. The experimental time series are from the studies of yeast cell cycle regulated genes and E. coli DNA repair genes. The unstable estimation of RNN using experimental time series having limited data points can lead to fairly arbitrary predicted GRNs. Our methods incorporate RNN and RMLP into a two-step structure learning procedure. Results show that the RE RMLP using the RMLP with a suitable number of latent nodes to reduce the parameter dimension often result in more accurate edge ranks than the RE RNN using the regularized RNN on short simulated time series. Combining by a weighted majority voting rule the networks derived by the RE RMLP -RNN using different numbers of latent nodes in step one to infer the GRN, the method performs consistently and outperforms published algorithms for GRN reconstruction on most benchmark time series. The framework of two-step algorithms can potentially incorporate with different nonlinear differential equation models to reconstruct the GRN.

A hybrid fuzzy logic and extreme learning machine for improving efficiency of circulating water systems in power generation plant

NASA Astrophysics Data System (ADS)

Aziz, Nur Liyana Afiqah Abdul; Siah Yap, Keem; Afif Bunyamin, Muhammad

2013-06-01

This paper presents a new approach of the fault detection for improving efficiency of circulating water system (CWS) in a power generation plant using a hybrid Fuzzy Logic System (FLS) and Extreme Learning Machine (ELM) neural network. The FLS is a mathematical tool for calculating the uncertainties where precision and significance are applied in the real world. It is based on natural language which has the ability of "computing the word". The ELM is an extremely fast learning algorithm for neural network that can completed the training cycle in a very short time. By combining the FLS and ELM, new hybrid model, i.e., FLS-ELM is developed. The applicability of this proposed hybrid model is validated in fault detection in CWS which may help to improve overall efficiency of power generation plant, hence, consuming less natural recourses and producing less pollutions.

Deep Learning in Medical Imaging: General Overview

PubMed Central

Lee, June-Goo; Jun, Sanghoon; Cho, Young-Won; Lee, Hyunna; Kim, Guk Bae

2017-01-01

The artificial neural network (ANN)–a machine learning technique inspired by the human neuronal synapse system–was introduced in the 1950s. However, the ANN was previously limited in its ability to solve actual problems, due to the vanishing gradient and overfitting problems with training of deep architecture, lack of computing power, and primarily the absence of sufficient data to train the computer system. Interest in this concept has lately resurfaced, due to the availability of big data, enhanced computing power with the current graphics processing units, and novel algorithms to train the deep neural network. Recent studies on this technology suggest its potentially to perform better than humans in some visual and auditory recognition tasks, which may portend its applications in medicine and healthcare, especially in medical imaging, in the foreseeable future. This review article offers perspectives on the history, development, and applications of deep learning technology, particularly regarding its applications in medical imaging. PMID:28670152

Deep Learning in Medical Imaging: General Overview.

PubMed

Lee, June-Goo; Jun, Sanghoon; Cho, Young-Won; Lee, Hyunna; Kim, Guk Bae; Seo, Joon Beom; Kim, Namkug

2017-01-01

The artificial neural network (ANN)-a machine learning technique inspired by the human neuronal synapse system-was introduced in the 1950s. However, the ANN was previously limited in its ability to solve actual problems, due to the vanishing gradient and overfitting problems with training of deep architecture, lack of computing power, and primarily the absence of sufficient data to train the computer system. Interest in this concept has lately resurfaced, due to the availability of big data, enhanced computing power with the current graphics processing units, and novel algorithms to train the deep neural network. Recent studies on this technology suggest its potentially to perform better than humans in some visual and auditory recognition tasks, which may portend its applications in medicine and healthcare, especially in medical imaging, in the foreseeable future. This review article offers perspectives on the history, development, and applications of deep learning technology, particularly regarding its applications in medical imaging.

Developing Generic Image Search Strategies for Large Astronomical Data Sets and Archives using Convolutional Neural Networks and Transfer Learning

NASA Astrophysics Data System (ADS)

Peek, Joshua E. G.; Hargis, Jonathan R.; Jones, Craig K.

2018-01-01

Astronomical instruments produce petabytes of images every year, vastly more than can be inspected by a member of the astronomical community in search of a specific population of structures. Fortunately, the sky is mostly black and source extraction algorithms have been developed to provide searchable catalogs of unconfused sources like stars and galaxies. These tools often fail for studies of more diffuse structures like the interstellar medium and unresolved stellar structures in nearby galaxies, leaving astronomers interested in observations of photodissociation regions, stellar clusters, diffuse interstellar clouds without the crucial ability to search. In this work we present a new path forward for finding structures in large data sets similar to an input structure using convolutional neural networks, transfer learning, and machine learning clustering techniques. We show applications to archival data in the Mikulski Archive for Space Telescopes (MAST).

Neural networks for learning and prediction with applications to remote sensing and speech perception

NASA Astrophysics Data System (ADS)

Gjaja, Marin N.

1997-11-01

Neural networks for supervised and unsupervised learning are developed and applied to problems in remote sensing, continuous map learning, and speech perception. Adaptive Resonance Theory (ART) models are real-time neural networks for category learning, pattern recognition, and prediction. Unsupervised fuzzy ART networks synthesize fuzzy logic and neural networks, and supervised ARTMAP networks incorporate ART modules for prediction and classification. New ART and ARTMAP methods resulting from analyses of data structure, parameter specification, and category selection are developed. Architectural modifications providing flexibility for a variety of applications are also introduced and explored. A new methodology for automatic mapping from Landsat Thematic Mapper (TM) and terrain data, based on fuzzy ARTMAP, is developed. System capabilities are tested on a challenging remote sensing problem, prediction of vegetation classes in the Cleveland National Forest from spectral and terrain features. After training at the pixel level, performance is tested at the stand level, using sites not seen during training. Results are compared to those of maximum likelihood classifiers, back propagation neural networks, and K-nearest neighbor algorithms. Best performance is obtained using a hybrid system based on a convex combination of fuzzy ARTMAP and maximum likelihood predictions. This work forms the foundation for additional studies exploring fuzzy ARTMAP's capability to estimate class mixture composition for non-homogeneous sites. Exploratory simulations apply ARTMAP to the problem of learning continuous multidimensional mappings. A novel system architecture retains basic ARTMAP properties of incremental and fast learning in an on-line setting while adding components to solve this class of problems. The perceptual magnet effect is a language-specific phenomenon arising early in infant speech development that is characterized by a warping of speech sound perception. An unsupervised neural network model is proposed that embodies two principal hypotheses supported by experimental data--that sensory experience guides language-specific development of an auditory neural map and that a population vector can predict psychological phenomena based on map cell activities. Model simulations show how a nonuniform distribution of map cell firing preferences can develop from language-specific input and give rise to the magnet effect.

Neural networks for tracking of unknown SISO discrete-time nonlinear dynamic systems.

PubMed

Aftab, Muhammad Saleheen; Shafiq, Muhammad

2015-11-01

This article presents a Lyapunov function based neural network tracking (LNT) strategy for single-input, single-output (SISO) discrete-time nonlinear dynamic systems. The proposed LNT architecture is composed of two feedforward neural networks operating as controller and estimator. A Lyapunov function based back propagation learning algorithm is used for online adjustment of the controller and estimator parameters. The controller and estimator error convergence and closed-loop system stability analysis is performed by Lyapunov stability theory. Moreover, two simulation examples and one real-time experiment are investigated as case studies. The achieved results successfully validate the controller performance. Copyright © 2015 ISA. Published by Elsevier Ltd. All rights reserved.

An Automated Procedure for Evaluating Song Imitation

PubMed Central

Mandelblat-Cerf, Yael; Fee, Michale S.

2014-01-01

Songbirds have emerged as an excellent model system to understand the neural basis of vocal and motor learning. Like humans, songbirds learn to imitate the vocalizations of their parents or other conspecific “tutors.” Young songbirds learn by comparing their own vocalizations to the memory of their tutor song, slowly improving until over the course of several weeks they can achieve an excellent imitation of the tutor. Because of the slow progression of vocal learning, and the large amounts of singing generated, automated algorithms for quantifying vocal imitation have become increasingly important for studying the mechanisms underlying this process. However, methodologies for quantifying song imitation are complicated by the highly variable songs of either juvenile birds or those that learn poorly because of experimental manipulations. Here we present a method for the evaluation of song imitation that incorporates two innovations: First, an automated procedure for selecting pupil song segments, and, second, a new algorithm, implemented in Matlab, for computing both song acoustic and sequence similarity. We tested our procedure using zebra finch song and determined a set of acoustic features for which the algorithm optimally differentiates between similar and non-similar songs. PMID:24809510

A comparison of back propagation and Generalized Regression Neural Networks performance in neutron spectrometry.

PubMed

Martínez-Blanco, Ma Del Rosario; Ornelas-Vargas, Gerardo; Solís-Sánchez, Luis Octavio; Castañeda-Miranada, Rodrigo; Vega-Carrillo, Héctor René; Celaya-Padilla, José M; Garza-Veloz, Idalia; Martínez-Fierro, Margarita; Ortiz-Rodríguez, José Manuel

2016-11-01

The process of unfolding the neutron energy spectrum has been subject of research for many years. Monte Carlo, iterative methods, the bayesian theory, the principle of maximum entropy are some of the methods used. The drawbacks associated with traditional unfolding procedures have motivated the research of complementary approaches. Back Propagation Neural Networks (BPNN), have been applied with success in neutron spectrometry and dosimetry domains, however, the structure and learning parameters are factors that highly impact in the networks performance. In ANN domain, Generalized Regression Neural Network (GRNN) is one of the simplest neural networks in term of network architecture and learning algorithm. The learning is instantaneous, requiring no time for training. Opposite to BPNN, a GRNN would be formed instantly with just a 1-pass training on the development data. In the network development phase, the only hurdle is to optimize the hyper-parameter, which is known as sigma, governing the smoothness of the network. The aim of this work was to compare the performance of BPNN and GRNN in the solution of the neutron spectrometry problem. From results obtained it can be observed that despite the very similar results, GRNN performs better than BPNN. Copyright © 2016 Elsevier Ltd. All rights reserved.

A review and analysis of neural networks for classification of remotely sensed multispectral imagery

NASA Technical Reports Server (NTRS)

Paola, Justin D.; Schowengerdt, Robert A.

1993-01-01

A literature survey and analysis of the use of neural networks for the classification of remotely sensed multispectral imagery is presented. As part of a brief mathematical review, the backpropagation algorithm, which is the most common method of training multi-layer networks, is discussed with an emphasis on its application to pattern recognition. The analysis is divided into five aspects of neural network classification: (1) input data preprocessing, structure, and encoding; (2) output encoding and extraction of classes; (3) network architecture, (4) training algorithms; and (5) comparisons to conventional classifiers. The advantages of the neural network method over traditional classifiers are its non-parametric nature, arbitrary decision boundary capabilities, easy adaptation to different types of data and input structures, fuzzy output values that can enhance classification, and good generalization for use with multiple images. The disadvantages of the method are slow training time, inconsistent results due to random initial weights, and the requirement of obscure initialization values (e.g., learning rate and hidden layer size). Possible techniques for ameliorating these problems are discussed. It is concluded that, although the neural network method has several unique capabilities, it will become a useful tool in remote sensing only if it is made faster, more predictable, and easier to use.

Evaluation of machine learning algorithms for classification of primary biological aerosol using a new UV-LIF spectrometer

NASA Astrophysics Data System (ADS)

Ruske, Simon; Topping, David O.; Foot, Virginia E.; Kaye, Paul H.; Stanley, Warren R.; Crawford, Ian; Morse, Andrew P.; Gallagher, Martin W.

2017-03-01

Characterisation of bioaerosols has important implications within environment and public health sectors. Recent developments in ultraviolet light-induced fluorescence (UV-LIF) detectors such as the Wideband Integrated Bioaerosol Spectrometer (WIBS) and the newly introduced Multiparameter Bioaerosol Spectrometer (MBS) have allowed for the real-time collection of fluorescence, size and morphology measurements for the purpose of discriminating between bacteria, fungal spores and pollen.This new generation of instruments has enabled ever larger data sets to be compiled with the aim of studying more complex environments. In real world data sets, particularly those from an urban environment, the population may be dominated by non-biological fluorescent interferents, bringing into question the accuracy of measurements of quantities such as concentrations. It is therefore imperative that we validate the performance of different algorithms which can be used for the task of classification.For unsupervised learning we tested hierarchical agglomerative clustering with various different linkages. For supervised learning, 11 methods were tested, including decision trees, ensemble methods (random forests, gradient boosting and AdaBoost), two implementations for support vector machines (libsvm and liblinear) and Gaussian methods (Gaussian naïve Bayesian, quadratic and linear discriminant analysis, the k-nearest neighbours algorithm and artificial neural networks).The methods were applied to two different data sets produced using the new MBS, which provides multichannel UV-LIF fluorescence signatures for single airborne biological particles. The first data set contained mixed PSLs and the second contained a variety of laboratory-generated aerosol.Clustering in general performs slightly worse than the supervised learning methods, correctly classifying, at best, only 67. 6 and 91. 1 % for the two data sets respectively. For supervised learning the gradient boosting algorithm was found to be the most effective, on average correctly classifying 82. 8 and 98. 27 % of the testing data, respectively, across the two data sets.A possible alternative to gradient boosting is neural networks. We do however note that this method requires much more user input than the other methods, and we suggest that further research should be conducted using this method, especially using parallelised hardware such as the GPU, which would allow for larger networks to be trained, which could possibly yield better results.We also saw that some methods, such as clustering, failed to utilise the additional shape information provided by the instrument, whilst for others, such as the decision trees, ensemble methods and neural networks, improved performance could be attained with the inclusion of such information.

Using deep convolutional neural networks to identify and classify tumor-associated stroma in diagnostic breast biopsies.

PubMed

Ehteshami Bejnordi, Babak; Mullooly, Maeve; Pfeiffer, Ruth M; Fan, Shaoqi; Vacek, Pamela M; Weaver, Donald L; Herschorn, Sally; Brinton, Louise A; van Ginneken, Bram; Karssemeijer, Nico; Beck, Andrew H; Gierach, Gretchen L; van der Laak, Jeroen A W M; Sherman, Mark E

2018-06-13

The breast stromal microenvironment is a pivotal factor in breast cancer development, growth and metastases. Although pathologists often detect morphologic changes in stroma by light microscopy, visual classification of such changes is subjective and non-quantitative, limiting its diagnostic utility. To gain insights into stromal changes associated with breast cancer, we applied automated machine learning techniques to digital images of 2387 hematoxylin and eosin stained tissue sections of benign and malignant image-guided breast biopsies performed to investigate mammographic abnormalities among 882 patients, ages 40-65 years, that were enrolled in the Breast Radiology Evaluation and Study of Tissues (BREAST) Stamp Project. Using deep convolutional neural networks, we trained an algorithm to discriminate between stroma surrounding invasive cancer and stroma from benign biopsies. In test sets (928 whole-slide images from 330 patients), this algorithm could distinguish biopsies diagnosed as invasive cancer from benign biopsies solely based on the stromal characteristics (area under the receiver operator characteristics curve = 0.962). Furthermore, without being trained specifically using ductal carcinoma in situ as an outcome, the algorithm detected tumor-associated stroma in greater amounts and at larger distances from grade 3 versus grade 1 ductal carcinoma in situ. Collectively, these results suggest that algorithms based on deep convolutional neural networks that evaluate only stroma may prove useful to classify breast biopsies and aid in understanding and evaluating the biology of breast lesions.

BELM: Bayesian extreme learning machine.

PubMed

Soria-Olivas, Emilio; Gómez-Sanchis, Juan; Martín, José D; Vila-Francés, Joan; Martínez, Marcelino; Magdalena, José R; Serrano, Antonio J

2011-03-01

The theory of extreme learning machine (ELM) has become very popular on the last few years. ELM is a new approach for learning the parameters of the hidden layers of a multilayer neural network (as the multilayer perceptron or the radial basis function neural network). Its main advantage is the lower computational cost, which is especially relevant when dealing with many patterns defined in a high-dimensional space. This brief proposes a bayesian approach to ELM, which presents some advantages over other approaches: it allows the introduction of a priori knowledge; obtains the confidence intervals (CIs) without the need of applying methods that are computationally intensive, e.g., bootstrap; and presents high generalization capabilities. Bayesian ELM is benchmarked against classical ELM in several artificial and real datasets that are widely used for the evaluation of machine learning algorithms. Achieved results show that the proposed approach produces a competitive accuracy with some additional advantages, namely, automatic production of CIs, reduction of probability of model overfitting, and use of a priori knowledge.

A review of machine learning in obesity.

PubMed

DeGregory, K W; Kuiper, P; DeSilvio, T; Pleuss, J D; Miller, R; Roginski, J W; Fisher, C B; Harness, D; Viswanath, S; Heymsfield, S B; Dungan, I; Thomas, D M

2018-05-01

Rich sources of obesity-related data arising from sensors, smartphone apps, electronic medical health records and insurance data can bring new insights for understanding, preventing and treating obesity. For such large datasets, machine learning provides sophisticated and elegant tools to describe, classify and predict obesity-related risks and outcomes. Here, we review machine learning methods that predict and/or classify such as linear and logistic regression, artificial neural networks, deep learning and decision tree analysis. We also review methods that describe and characterize data such as cluster analysis, principal component analysis, network science and topological data analysis. We introduce each method with a high-level overview followed by examples of successful applications. The algorithms were then applied to National Health and Nutrition Examination Survey to demonstrate methodology, utility and outcomes. The strengths and limitations of each method were also evaluated. This summary of machine learning algorithms provides a unique overview of the state of data analysis applied specifically to obesity. © 2018 World Obesity Federation.

Logic Learning Machine and standard supervised methods for Hodgkin's lymphoma prognosis using gene expression data and clinical variables.

PubMed

Parodi, Stefano; Manneschi, Chiara; Verda, Damiano; Ferrari, Enrico; Muselli, Marco

2018-03-01

This study evaluates the performance of a set of machine learning techniques in predicting the prognosis of Hodgkin's lymphoma using clinical factors and gene expression data. Analysed samples from 130 Hodgkin's lymphoma patients included a small set of clinical variables and more than 54,000 gene features. Machine learning classifiers included three black-box algorithms ( k-nearest neighbour, Artificial Neural Network, and Support Vector Machine) and two methods based on intelligible rules (Decision Tree and the innovative Logic Learning Machine method). Support Vector Machine clearly outperformed any of the other methods. Among the two rule-based algorithms, Logic Learning Machine performed better and identified a set of simple intelligible rules based on a combination of clinical variables and gene expressions. Decision Tree identified a non-coding gene ( XIST) involved in the early phases of X chromosome inactivation that was overexpressed in females and in non-relapsed patients. XIST expression might be responsible for the better prognosis of female Hodgkin's lymphoma patients.

What does fault tolerant Deep Learning need from MPI?

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

Amatya, Vinay C.; Vishnu, Abhinav; Siegel, Charles M.

Deep Learning (DL) algorithms have become the {\\em de facto} Machine Learning (ML) algorithm for large scale data analysis. DL algorithms are computationally expensive -- even distributed DL implementations which use MPI require days of training (model learning) time on commonly studied datasets. Long running DL applications become susceptible to faults -- requiring development of a fault tolerant system infrastructure, in addition to fault tolerant DL algorithms. This raises an important question: {\\em What is needed from MPI for designing fault tolerant DL implementations?} In this paper, we address this problem for permanent faults. We motivate the need for amore » fault tolerant MPI specification by an in-depth consideration of recent innovations in DL algorithms and their properties, which drive the need for specific fault tolerance features. We present an in-depth discussion on the suitability of different parallelism types (model, data and hybrid); a need (or lack thereof) for check-pointing of any critical data structures; and most importantly, consideration for several fault tolerance proposals (user-level fault mitigation (ULFM), Reinit) in MPI and their applicability to fault tolerant DL implementations. We leverage a distributed memory implementation of Caffe, currently available under the Machine Learning Toolkit for Extreme Scale (MaTEx). We implement our approaches by extending MaTEx-Caffe for using ULFM-based implementation. Our evaluation using the ImageNet dataset and AlexNet neural network topology demonstrates the effectiveness of the proposed fault tolerant DL implementation using OpenMPI based ULFM.« less

A Universal 3D Voxel Descriptor for Solid-State Material Informatics with Deep Convolutional Neural Networks.

PubMed

Kajita, Seiji; Ohba, Nobuko; Jinnouchi, Ryosuke; Asahi, Ryoji

2017-12-05

Material informatics (MI) is a promising approach to liberate us from the time-consuming Edisonian (trial and error) process for material discoveries, driven by machine-learning algorithms. Several descriptors, which are encoded material features to feed computers, were proposed in the last few decades. Especially to solid systems, however, their insufficient representations of three dimensionality of field quantities such as electron distributions and local potentials have critically hindered broad and practical successes of the solid-state MI. We develop a simple, generic 3D voxel descriptor that compacts any field quantities, in such a suitable way to implement convolutional neural networks (CNNs). We examine the 3D voxel descriptor encoded from the electron distribution by a regression test with 680 oxides data. The present scheme outperforms other existing descriptors in the prediction of Hartree energies that are significantly relevant to the long-wavelength distribution of the valence electrons. The results indicate that this scheme can forecast any functionals of field quantities just by learning sufficient amount of data, if there is an explicit correlation between the target properties and field quantities. This 3D descriptor opens a way to import prominent CNNs-based algorithms of supervised, semi-supervised and reinforcement learnings into the solid-state MI.

Empirical modeling for intelligent, real-time manufacture control

NASA Technical Reports Server (NTRS)

Xu, Xiaoshu

1994-01-01

Artificial neural systems (ANS), also known as neural networks, are an attempt to develop computer systems that emulate the neural reasoning behavior of biological neural systems (e.g. the human brain). As such, they are loosely based on biological neural networks. The ANS consists of a series of nodes (neurons) and weighted connections (axons) that, when presented with a specific input pattern, can associate specific output patterns. It is essentially a highly complex, nonlinear, mathematical relationship or transform. These constructs have two significant properties that have proven useful to the authors in signal processing and process modeling: noise tolerance and complex pattern recognition. Specifically, the authors have developed a new network learning algorithm that has resulted in the successful application of ANS's to high speed signal processing and to developing models of highly complex processes. Two of the applications, the Weld Bead Geometry Control System and the Welding Penetration Monitoring System, are discussed in the body of this paper.

Effectiveness of feature and classifier algorithms in character recognition systems

NASA Astrophysics Data System (ADS)

Wilson, Charles L.

1993-04-01

At the first Census Optical Character Recognition Systems Conference, NIST generated accuracy data for more than character recognition systems. Most systems were tested on the recognition of isolated digits and upper and lower case alphabetic characters. The recognition experiments were performed on sample sizes of 58,000 digits, and 12,000 upper and lower case alphabetic characters. The algorithms used by the 26 conference participants included rule-based methods, image-based methods, statistical methods, and neural networks. The neural network methods included Multi-Layer Perceptron's, Learned Vector Quantitization, Neocognitrons, and cascaded neural networks. In this paper 11 different systems are compared using correlations between the answers of different systems, comparing the decrease in error rate as a function of confidence of recognition, and comparing the writer dependence of recognition. This comparison shows that methods that used different algorithms for feature extraction and recognition performed with very high levels of correlation. This is true for neural network systems, hybrid systems, and statistically based systems, and leads to the conclusion that neural networks have not yet demonstrated a clear superiority to more conventional statistical methods. Comparison of these results with the models of Vapnick (for estimation problems), MacKay (for Bayesian statistical models), Moody (for effective parameterization), and Boltzmann models (for information content) demonstrate that as the limits of training data variance are approached, all classifier systems have similar statistical properties. The limiting condition can only be approached for sufficiently rich feature sets because the accuracy limit is controlled by the available information content of the training set, which must pass through the feature extraction process prior to classification.

Improved relocatable over-the-horizon radar detection and tracking using the maximum likelihood adaptive neural system algorithm

NASA Astrophysics Data System (ADS)

Perlovsky, Leonid I.; Webb, Virgil H.; Bradley, Scott R.; Hansen, Christopher A.

1998-07-01

An advanced detection and tracking system is being developed for the U.S. Navy's Relocatable Over-the-Horizon Radar (ROTHR) to provide improved tracking performance against small aircraft typically used in drug-smuggling activities. The development is based on the Maximum Likelihood Adaptive Neural System (MLANS), a model-based neural network that combines advantages of neural network and model-based algorithmic approaches. The objective of the MLANS tracker development effort is to address user requirements for increased detection and tracking capability in clutter and improved track position, heading, and speed accuracy. The MLANS tracker is expected to outperform other approaches to detection and tracking for the following reasons. It incorporates adaptive internal models of target return signals, target tracks and maneuvers, and clutter signals, which leads to concurrent clutter suppression, detection, and tracking (track-before-detect). It is not combinatorial and thus does not require any thresholding or peak picking and can track in low signal-to-noise conditions. It incorporates superresolution spectrum estimation techniques exceeding the performance of conventional maximum likelihood and maximum entropy methods. The unique spectrum estimation method is based on the Einsteinian interpretation of the ROTHR received energy spectrum as a probability density of signal frequency. The MLANS neural architecture and learning mechanism are founded on spectrum models and maximization of the "Einsteinian" likelihood, allowing knowledge of the physical behavior of both targets and clutter to be injected into the tracker algorithms. The paper describes the addressed requirements and expected improvements, theoretical foundations, engineering methodology, and results of the development effort to date.

A novel and generalized approach in the inversion of geoelectrical resistivity data using Artificial Neural Networks (ANN)

NASA Astrophysics Data System (ADS)

Raj, A. Stanley; Srinivas, Y.; Oliver, D. Hudson; Muthuraj, D.

2014-03-01

The non-linear apparent resistivity problem in the subsurface study of the earth takes into account the model parameters in terms of resistivity and thickness of individual subsurface layers using the trained synthetic data by means of Artificial Neural Networks (ANN). Here we used a single layer feed-forward neural network with fast back propagation learning algorithm. So on proper training of back propagation networks it tends to give the resistivity and thickness of the subsurface layer model of the field resistivity data with reference to the synthetic data trained in the appropriate network. During training, the weights and biases of the network are iteratively adjusted to make network performance function level more efficient. On adequate training, errors are minimized and the best result is obtained using the artificial neural networks. The network is trained with more number of VES data and this trained network is demonstrated by the field data. The accuracy of inversion depends upon the number of data trained. In this novel and specially designed algorithm, the interpretation of the vertical electrical sounding has been done successfully with the more accurate layer model.

Adaptive enhanced sampling by force-biasing using neural networks

NASA Astrophysics Data System (ADS)

Guo, Ashley Z.; Sevgen, Emre; Sidky, Hythem; Whitmer, Jonathan K.; Hubbell, Jeffrey A.; de Pablo, Juan J.

2018-04-01

A machine learning assisted method is presented for molecular simulation of systems with rugged free energy landscapes. The method is general and can be combined with other advanced sampling techniques. In the particular implementation proposed here, it is illustrated in the context of an adaptive biasing force approach where, rather than relying on discrete force estimates, one can resort to a self-regularizing artificial neural network to generate continuous, estimated generalized forces. By doing so, the proposed approach addresses several shortcomings common to adaptive biasing force and other algorithms. Specifically, the neural network enables (1) smooth estimates of generalized forces in sparsely sampled regions, (2) force estimates in previously unexplored regions, and (3) continuous force estimates with which to bias the simulation, as opposed to biases generated at specific points of a discrete grid. The usefulness of the method is illustrated with three different examples, chosen to highlight the wide range of applicability of the underlying concepts. In all three cases, the new method is found to enhance considerably the underlying traditional adaptive biasing force approach. The method is also found to provide improvements over previous implementations of neural network assisted algorithms.

CNN-BLPred: a Convolutional neural network based predictor for β-Lactamases (BL) and their classes.

PubMed

White, Clarence; Ismail, Hamid D; Saigo, Hiroto; Kc, Dukka B

2017-12-28

The β-Lactamase (BL) enzyme family is an important class of enzymes that plays a key role in bacterial resistance to antibiotics. As the newly identified number of BL enzymes is increasing daily, it is imperative to develop a computational tool to classify the newly identified BL enzymes into one of its classes. There are two types of classification of BL enzymes: Molecular Classification and Functional Classification. Existing computational methods only address Molecular Classification and the performance of these existing methods is unsatisfactory. We addressed the unsatisfactory performance of the existing methods by implementing a Deep Learning approach called Convolutional Neural Network (CNN). We developed CNN-BLPred, an approach for the classification of BL proteins. The CNN-BLPred uses Gradient Boosted Feature Selection (GBFS) in order to select the ideal feature set for each BL classification. Based on the rigorous benchmarking of CCN-BLPred using both leave-one-out cross-validation and independent test sets, CCN-BLPred performed better than the other existing algorithms. Compared with other architectures of CNN, Recurrent Neural Network, and Random Forest, the simple CNN architecture with only one convolutional layer performs the best. After feature extraction, we were able to remove ~95% of the 10,912 features using Gradient Boosted Trees. During 10-fold cross validation, we increased the accuracy of the classic BL predictions by 7%. We also increased the accuracy of Class A, Class B, Class C, and Class D performance by an average of 25.64%. The independent test results followed a similar trend. We implemented a deep learning algorithm known as Convolutional Neural Network (CNN) to develop a classifier for BL classification. Combined with feature selection on an exhaustive feature set and using balancing method such as Random Oversampling (ROS), Random Undersampling (RUS) and Synthetic Minority Oversampling Technique (SMOTE), CNN-BLPred performs significantly better than existing algorithms for BL classification.

Modeling and control of magnetorheological fluid dampers using neural networks

NASA Astrophysics Data System (ADS)

Wang, D. H.; Liao, W. H.

2005-02-01

Due to the inherent nonlinear nature of magnetorheological (MR) fluid dampers, one of the challenging aspects for utilizing these devices to achieve high system performance is the development of accurate models and control algorithms that can take advantage of their unique characteristics. In this paper, the direct identification and inverse dynamic modeling for MR fluid dampers using feedforward and recurrent neural networks are studied. The trained direct identification neural network model can be used to predict the damping force of the MR fluid damper on line, on the basis of the dynamic responses across the MR fluid damper and the command voltage, and the inverse dynamic neural network model can be used to generate the command voltage according to the desired damping force through supervised learning. The architectures and the learning methods of the dynamic neural network models and inverse neural network models for MR fluid dampers are presented, and some simulation results are discussed. Finally, the trained neural network models are applied to predict and control the damping force of the MR fluid damper. Moreover, validation methods for the neural network models developed are proposed and used to evaluate their performance. Validation results with different data sets indicate that the proposed direct identification dynamic model using the recurrent neural network can be used to predict the damping force accurately and the inverse identification dynamic model using the recurrent neural network can act as a damper controller to generate the command voltage when the MR fluid damper is used in a semi-active mode.

An Interval Type-2 Neural Fuzzy System for Online System Identification and Feature Elimination.

PubMed

Lin, Chin-Teng; Pal, Nikhil R; Wu, Shang-Lin; Liu, Yu-Ting; Lin, Yang-Yin

2015-07-01

We propose an integrated mechanism for discarding derogatory features and extraction of fuzzy rules based on an interval type-2 neural fuzzy system (NFS)-in fact, it is a more general scheme that can discard bad features, irrelevant antecedent clauses, and even irrelevant rules. High-dimensional input variable and a large number of rules not only enhance the computational complexity of NFSs but also reduce their interpretability. Therefore, a mechanism for simultaneous extraction of fuzzy rules and reducing the impact of (or eliminating) the inferior features is necessary. The proposed approach, namely an interval type-2 Neural Fuzzy System for online System Identification and Feature Elimination (IT2NFS-SIFE), uses type-2 fuzzy sets to model uncertainties associated with information and data in designing the knowledge base. The consequent part of the IT2NFS-SIFE is of Takagi-Sugeno-Kang type with interval weights. The IT2NFS-SIFE possesses a self-evolving property that can automatically generate fuzzy rules. The poor features can be discarded through the concept of a membership modulator. The antecedent and modulator weights are learned using a gradient descent algorithm. The consequent part weights are tuned via the rule-ordered Kalman filter algorithm to enhance learning effectiveness. Simulation results show that IT2NFS-SIFE not only simplifies the system architecture by eliminating derogatory/irrelevant antecedent clauses, rules, and features but also maintains excellent performance.

A hybrid neural learning algorithm using evolutionary learning and derivative free local search method.

PubMed

Ghosh, Ranadhir; Yearwood, John; Ghosh, Moumita; Bagirov, Adil

2006-06-01

In this paper we investigate a hybrid model based on the Discrete Gradient method and an evolutionary strategy for determining the weights in a feed forward artificial neural network. Also we discuss different variants for hybrid models using the Discrete Gradient method and an evolutionary strategy for determining the weights in a feed forward artificial neural network. The Discrete Gradient method has the advantage of being able to jump over many local minima and find very deep local minima. However, earlier research has shown that a good starting point for the discrete gradient method can improve the quality of the solution point. Evolutionary algorithms are best suited for global optimisation problems. Nevertheless they are cursed with longer training times and often unsuitable for real world application. For optimisation problems such as weight optimisation for ANNs in real world applications the dimensions are large and time complexity is critical. Hence the idea of a hybrid model can be a suitable option. In this paper we propose different fusion strategies for hybrid models combining the evolutionary strategy with the discrete gradient method to obtain an optimal solution much quicker. Three different fusion strategies are discussed: a linear hybrid model, an iterative hybrid model and a restricted local search hybrid model. Comparative results on a range of standard datasets are provided for different fusion hybrid models.

High pressure air compressor valve fault diagnosis using feedforward neural networks

NASA Astrophysics Data System (ADS)

James Li, C.; Yu, Xueli

1995-09-01

Feedforward neural networks (FNNs) are developed and implemented to classify a four-stage high pressure air compressor into one of the following conditions: baseline, suction or exhaust valve faults. These FNNs are used for the compressor's automatic condition monitoring and fault diagnosis. Measurements of 39 variables are obtained under different baseline conditions and third-stage suction and exhaust valve faults. These variables include pressures and temperatures at all stages, voltage between phase aand phase b, voltage between phase band phase c, total three-phase real power, cooling water flow rate, etc. To reduce the number of variables, the amount of their discriminatory information is quantified by scattering matrices to identify statistical significant ones. Measurements of the selected variables are then used by a fully automatic structural and weight learning algorithm to construct three-layer FNNs to classify the compressor's condition. This learning algorithm requires neither guesses of initial weight values nor number of neurons in the hidden layer of an FNN. It takes an incremental approach in which a hidden neuron is trained by exemplars and then augmented to the existing network. These exemplars are then made orthogonal to the newly identified hidden neuron. They are subsequently used for the training of the next hidden neuron. The betterment continues until a desired accuracy is reached. After the neural networks are established, novel measurements from various conditions that haven't been previously seen by the FNNs are then used to evaluate their ability in fault diagnosis. The trained neural networks provide very accurate diagnosis for suction and discharge valve defects.

Interacting neural networks.

PubMed

Metzler, R; Kinzel, W; Kanter, I

2000-08-01

Several scenarios of interacting neural networks which are trained either in an identical or in a competitive way are solved analytically. In the case of identical training each perceptron receives the output of its neighbor. The symmetry of the stationary state as well as the sensitivity to the used training algorithm are investigated. Two competitive perceptrons trained on mutually exclusive learning aims and a perceptron which is trained on the opposite of its own output are examined analytically. An ensemble of competitive perceptrons is used as decision-making algorithms in a model of a closed market (El Farol Bar problem or the Minority Game. In this game, a set of agents who have to make a binary decision is considered.); each network is trained on the history of minority decisions. This ensemble of perceptrons relaxes to a stationary state whose performance can be better than random.

Interacting neural networks

NASA Astrophysics Data System (ADS)

Metzler, R.; Kinzel, W.; Kanter, I.

2000-08-01

Several scenarios of interacting neural networks which are trained either in an identical or in a competitive way are solved analytically. In the case of identical training each perceptron receives the output of its neighbor. The symmetry of the stationary state as well as the sensitivity to the used training algorithm are investigated. Two competitive perceptrons trained on mutually exclusive learning aims and a perceptron which is trained on the opposite of its own output are examined analytically. An ensemble of competitive perceptrons is used as decision-making algorithms in a model of a closed market (El Farol Bar problem or the Minority Game. In this game, a set of agents who have to make a binary decision is considered.); each network is trained on the history of minority decisions. This ensemble of perceptrons relaxes to a stationary state whose performance can be better than random.

Hyperspectral Image Classification With Markov Random Fields and a Convolutional Neural Network

NASA Astrophysics Data System (ADS)

Cao, Xiangyong; Zhou, Feng; Xu, Lin; Meng, Deyu; Xu, Zongben; Paisley, John

2018-05-01

This paper presents a new supervised classification algorithm for remotely sensed hyperspectral image (HSI) which integrates spectral and spatial information in a unified Bayesian framework. First, we formulate the HSI classification problem from a Bayesian perspective. Then, we adopt a convolutional neural network (CNN) to learn the posterior class distributions using a patch-wise training strategy to better use the spatial information. Next, spatial information is further considered by placing a spatial smoothness prior on the labels. Finally, we iteratively update the CNN parameters using stochastic gradient decent (SGD) and update the class labels of all pixel vectors using an alpha-expansion min-cut-based algorithm. Compared with other state-of-the-art methods, the proposed classification method achieves better performance on one synthetic dataset and two benchmark HSI datasets in a number of experimental settings.

Quantum Associative Neural Network with Nonlinear Search Algorithm

NASA Astrophysics Data System (ADS)

Zhou, Rigui; Wang, Huian; Wu, Qian; Shi, Yang

2012-03-01

Based on analysis on properties of quantum linear superposition, to overcome the complexity of existing quantum associative memory which was proposed by Ventura, a new storage method for multiply patterns is proposed in this paper by constructing the quantum array with the binary decision diagrams. Also, the adoption of the nonlinear search algorithm increases the pattern recalling speed of this model which has multiply patterns to O( {log2}^{2^{n -t}} ) = O( n - t ) time complexity, where n is the number of quantum bit and t is the quantum information of the t quantum bit. Results of case analysis show that the associative neural network model proposed in this paper based on quantum learning is much better and optimized than other researchers' counterparts both in terms of avoiding the additional qubits or extraordinary initial operators, storing pattern and improving the recalling speed.

Neural Network and Nearest Neighbor Algorithms for Enhancing Sampling of Molecular Dynamics.

PubMed

Galvelis, Raimondas; Sugita, Yuji

2017-06-13

The free energy calculations of complex chemical and biological systems with molecular dynamics (MD) are inefficient due to multiple local minima separated by high-energy barriers. The minima can be escaped using an enhanced sampling method such as metadynamics, which apply bias (i.e., importance sampling) along a set of collective variables (CV), but the maximum number of CVs (or dimensions) is severely limited. We propose a high-dimensional bias potential method (NN2B) based on two machine learning algorithms: the nearest neighbor density estimator (NNDE) and the artificial neural network (ANN) for the bias potential approximation. The bias potential is constructed iteratively from short biased MD simulations accounting for correlation among CVs. Our method is capable of achieving ergodic sampling and calculating free energy of polypeptides with up to 8-dimensional bias potential.