Hafnium isotope stratigraphy of ferromanganese crusts

PubMed

Lee; Halliday; Hein; Burton; Christensen; Gunther

1999-08-13

A Cenozoic record of hafnium isotopic compositions of central Pacific deep water has been obtained from two ferromanganese crusts. The crusts are separated by more than 3000 kilometers but display similar secular variations. Significant fluctuations in hafnium isotopic composition occurred in the Eocene and Oligocene, possibly related to direct advection from the Indian and Atlantic oceans. Hafnium isotopic compositions have remained approximately uniform for the past 20 million years, probably reflecting increased isolation of the central Pacific. The mechanisms responsible for the increase in (87)Sr/(86)Sr in seawater through the Cenozoic apparently had no effect on central Pacific deep-water hafnium.

Ab Initio Study of Chemical Reactions of Cold SrF and CaF Molecules with Alkali-Metal and Alkaline-Earth-Metal Atoms: The Implications for Sympathetic Cooling.

PubMed

Kosicki, Maciej Bartosz; Kędziera, Dariusz; Żuchowski, Piotr Szymon

2017-06-01

We investigate the energetics of the atom exchange reaction in the SrF + alkali-metal atom and CaF + alkali-metal atom systems. Such reactions are possible only for collisions of SrF and CaF with the lithium atoms, while they are energetically forbidden for other alkali-metal atoms. Specifically, we focus on SrF interacting with Li, Rb, and Sr atoms and use ab initio methods to demonstrate that the SrF + Li and SrF + Sr reactions are barrierless. We present potential energy surfaces for the interaction of the SrF molecule with the Li, Rb, and Sr atoms in their energetically lowest-lying electronic spin states. The obtained potential energy surfaces are deep and exhibit profound interaction anisotropies. We predict that the collisions of SrF molecules in the rotational or Zeeman excited states most likely have a strong inelastic character. We discuss the prospects for the sympathetic cooling of SrF and CaF molecules using ultracold alkali-metal atoms.

Typhoon impacts on chemical weathering source provenance of a High Standing Island watershed, Taiwan

NASA Astrophysics Data System (ADS)

Meyer, Kevin J.; Carey, Anne E.; You, Chen-Feng

2017-10-01

Chemical weathering source provenance changes associated with Typhoon Mindulle (2004) were identified for the Choshui River Watershed in west-central Taiwan using radiogenic Sr isotope (87Sr/86Sr) and major ion chemistry analysis of water samples collected before, during, and following the storm event. Storm water sampling over 72 h was conducted in 3 h intervals, allowing for novel insight into weathering regime changes in response to intense rainfall events. Chemical weathering sources were determined to be bulk silicate and disseminated carbonate minerals at the surface and silicate contributions from deep thermal waters. Loss on ignition analysis of collected rock samples indicate disseminated carbonate can compose over 25% by weight of surface mineralogy, but typically makes up ∼2-3% of watershed rock. 87Sr/86Sr and major element molar ratios indicate that Typhoon Mindulle caused a weathering regime switch from normal flow incorporating a deep thermal signature to that of a system dominated by surface weathering. The data suggest release of silicate solute rich soil pore waters during storm events, creating a greater relative contribution of silicate weathering to the solute load during periods of increased precipitation and runoff. Partial depletion of this soil solute reservoir and possible erosion enhanced carbonate weathering lead to increased importance of carbonates to the weathering regime as the storm continues. Major ion data indicate that complex mica weathering (muscovite, biotite, illite, chlorite) may represent an important silicate weathering pathway in the watershed. Deep thermal waters represent an important contribution to river solutes during normal non-storm flow conditions. Sulfuric acid sourced from pyrite weathering is likely a major weathering agent in the Choshui River watershed.

Assembling old tricks for new tasks: a neural model of instructional learning and control.

PubMed

Huang, Tsung-Ren; Hazy, Thomas E; Herd, Seth A; O'Reilly, Randall C

2013-06-01

We can learn from the wisdom of others to maximize success. However, it is unclear how humans take advice to flexibly adapt behavior. On the basis of data from neuroanatomy, neurophysiology, and neuroimaging, a biologically plausible model is developed to illustrate the neural mechanisms of learning from instructions. The model consists of two complementary learning pathways. The slow-learning parietal pathway carries out simple or habitual stimulus-response (S-R) mappings, whereas the fast-learning hippocampal pathway implements novel S-R rules. Specifically, the hippocampus can rapidly encode arbitrary S-R associations, and stimulus-cued responses are later recalled into the basal ganglia-gated pFC to bias response selection in the premotor and motor cortices. The interactions between the two model learning pathways explain how instructions can override habits and how automaticity can be achieved through motor consolidation.

Nitrided SrTiO3 as charge-trapping layer for nonvolatile memory applications

NASA Astrophysics Data System (ADS)

Huang, X. D.; Lai, P. T.; Liu, L.; Xu, J. P.

2011-06-01

Charge-trapping characteristics of SrTiO3 with and without nitrogen incorporation were investigated based on Al/Al2O3/SrTiO3/SiO2/Si (MONOS) capacitors. A Ti-silicate interlayer at the SrTiO3/SiO2 interface was confirmed by x-ray photoelectron spectroscopy and transmission electron microscopy. Compared with the MONOS capacitor with SrTiO3 as charge-trapping layer (CTL), the one with nitrided SrTiO3 showed a larger memory window (8.4 V at ±10 V sweeping voltage), higher P/E speeds (1.8 V at 1 ms +8 V) and better retention properties (charge loss of 38% after 104 s), due to the nitrided SrTiO3 film exhibiting higher dielectric constant, higher deep-level traps induced by nitrogen incorporation, and suppressed formation of Ti silicate between the CTL and SiO2 by nitrogen passivation.

Potential application of radiogenic isotopes and geophysical methods to understand the hydrothermal dystem of the Upper Geyser Basin, Yellowstone National Park

USGS Publications Warehouse

Paces, James B.; Long, Andrew J.; Koth, Karl R.

2015-01-01

Numerous geochemical and geophysical studies have been conducted at Yellowstone National Park to better understand the hydrogeologic processes supporting the thermal features of the Park. This report provides the first 87Sr/86Sr and 234U/238U data for thermal water from the Upper Geyser Basin (UGB) intended to evaluate whether heavy radiogenic isotopes might provide insight to sources of groundwater supply and how they interact over time and space. In addition, this report summarizes previous geophysical studies made at Yellowstone National Park and provides suggestions for applying non-invasive ground and airborne studies to better understand groundwater flow in the subsurface of the UGB. Multiple samples from Old Faithful, Aurum, Grand, Oblong, and Daisy geysers characterized previously for major-ion concentrations and isotopes of water (δ2H, δ18O, and 3H) were analyzed for Sr and U isotopes. Concentrations of dissolved Sr and U are low (4.3–128 ng g-1 Sr and 0.026–0.0008 ng g-1 U); consequently only 87Sr/86Sr data are reported for most samples. Values of 87Sr/86Sr for most geysers remained uniform between April and September 2007, but show large increases in all five geysers between late October 2007 and early April, 2008. By late summer of 2008, 87Sr/86Sr values returned to values similar to those observed a year earlier. Similar patterns are not present in major-ion data measured on the same samples. Furthermore, large geochemical differences documented between geysers are not observed in 87Sr/86Sr data, although smaller differences between sites may be present. Sr-isotope data are consistent with a stratified hydrologic system where water erupted in spring and summer of 2007 and summer of 2008 equilibrated with local intracaldera rhyolite flows at shallower depths. Water erupted between October 2007 and April 2008 includes greater amounts of groundwater that circulated deep enough to acquire a radiogenic 87Sr/86Sr, most likely from Archean basement rocks. Details of how the shallow and deep components interact and mechanisms causing these interactions remain unknown, but the data demonstrate the usefulness of obtaining Sr-isotope data from future sample campaigns. Geophysical methods that would be useful for characterization of the UGB subsurface properties and geothermal system include electromagnetic (EM), gravity, and ambient seismic. A suite of ground-based EM methods could be used in a synergistic combination together with airborne EM surveys to provide data for a range of spatial scales and resolutions. Existing thermal data for the shallow subsurface could be used to relate ground and airborne EM survey data to locations of geothermal fluids near the surface. Gravity surveys would be useful for mapping subsurface density anomalies and possibly monitoring changes in degree of saturation with groundwater. Ambient seismic surveys would be useful for estimating the thickness of unconsolidated deposits that contain the shallow groundwater system. A study that combines radiogenic isotope tracers with geophysical methods has the potential to better characterize the geothermal workings in the UGB. Insights gained could lead to a better understanding of the geothermal system and how Park infrastructure may cause perturbations. Measurements of radiogenic isotopes from multiple geysers and pools in localized areas within the UGB that are coupled with data from geophysical surveys would help refine conceptual models of mixing between deep- and shallow-derived subsurface fluids.

The effect of learning on bursting.

PubMed

Stegenga, Jan; Le Feber, Joost; Marani, Enrico; Rutten, Wim L C

2009-04-01

We have studied the effect that learning a new stimulus-response (SR) relationship had within a neuronal network cultured on a multielectrode array. For training, we applied repetitive focal electrical stimulation delivered at a low rate (<1/s). Stimulation was withdrawn when a desired SR success ratio was achieved. It has been shown elsewhere, and we verified that this training algorithm, named conditional repetitive stimulation (CRS), can be used to strengthen an initially weak SR. So far, it remained unclear what the role of the rest of the network during learning was. We therefore studied the effect of CRS on spontaneously occurring network bursts. To this end, we made profiles of the firing rates within network bursts. We have earlier shown that these profiles change shape on a time base of several hours during spontaneous development. We show here that profiles of summed activity, called burst profiles, changed shape at an increased rate during CRS. This suggests that the whole network was involved in making the changes necessary to incorporate the desired SR relationship. However, a local (path-specific) component to learning was also found by analyzing profiles of single-electrode-activity phase profiles. Phase profiles that were not part of the SR relationship changed far less during CRS than the phase profiles of the electrodes that were part of the SR relationship. Finally, the manner in which phase profiles changed shape varied and could not be linked to the SR relationship.

Reciprocal Effects of Self-Regulation, Semantic Knowledge, and Reading Comprehension in Early Elementary School

PubMed Central

Connor, Carol McDonald; Day, Stephanie L.; Phillips, Beth; Sparapani, Nicole; Ingebrand, Sarah W.; McLean, Leigh; Barrus, Angela; Kaschak, Michael P.

2016-01-01

Many assume that cognitive and linguistic processes, such as semantic knowledge (SK) and self-regulation (SR) subserve learned skills like reading. However, complex models of interacting and bootstrapping effects of SK, SR, instruction, and reading hypothesize reciprocal effects. Testing this “lattice” model with children (n = 852) followed from 1st–2nd grade (5.9–10.4 years-of-age), revealed reciprocal effects for reading and SR, and reading and SK, but not SR and SK. More effective literacy instruction reduced reading stability over time. Findings elucidate the synergistic and reciprocal effects of learning to read on other important linguistic, self-regulatory, and cognitive processes, the value of using complex models of development to inform intervention design, and how learned skills may influence development during middle childhood. PMID:27264645

Intra-accumbal blockade of endocannabinoid CB1 receptors impairs learning but not retention of conditioned relief.

PubMed

Bergado Acosta, Jorge R; Schneider, Miriam; Fendt, Markus

2017-10-01

Humans and animals are able to associate an environmental cue with the feeling of relief from an aversive event, a phenomenon called relief learning. Relief from an aversive event is rewarding and a relief-associated cue later induces an attenuation of the startle magnitude or approach behavior. Previous studies demonstrated that the nucleus accumbens is essential for relief learning. Here, we asked whether accumbal cannabinoid type 1 (CB1) receptors are involved in relief learning. In rats, we injected the CB1 receptor antagonist/inverse agonist SR141716A (rimonabant) directly into the nucleus accumbens at different time points during a relief learning experiment. SR141716A injections immediately before the conditioning inhibited relief learning. However, SR141716A injected immediately before the retention test was not effective when conditioning was without treatment. These findings indicate that accumbal CB1 receptors play an important role in the plasticity processes underlying relief learning. Copyright © 2017 Elsevier Inc. All rights reserved.

Reciprocal Effects of Self-Regulation, Semantic Knowledge, and Reading Comprehension in Early Elementary School.

PubMed

Connor, Carol McDonald; Day, Stephanie L; Phillips, Beth; Sparapani, Nicole; Ingebrand, Sarah W; McLean, Leigh; Barrus, Angela; Kaschak, Michael P

2016-11-01

Many assume that cognitive and linguistic processes, such as semantic knowledge (SK) and self-regulation (SR), subserve learned skills like reading. However, complex models of interacting and bootstrapping effects of SK, SR, instruction, and reading hypothesize reciprocal effects. Testing this "lattice" model with children (n = 852) followed from first to second grade (5.9-10.4 years of age) revealed reciprocal effects for reading and SR, and reading and SK, but not SR and SK. More effective literacy instruction reduced reading stability over time. Findings elucidate the synergistic and reciprocal effects of learning to read on other important linguistic, self-regulatory, and cognitive processes; the value of using complex models of development to inform intervention design; and how learned skills may influence development during middle childhood. © 2016 The Authors. Child Development © 2016 Society for Research in Child Development, Inc.

Retrieval practice improves memory in survivors of severe traumatic brain injury.

PubMed

Sumowski, James F; Coyne, Julia; Cohen, Amanda; Deluca, John

2014-02-01

To investigate whether retrieval practice (RP) improves delayed recall after short and long delays in survivors of severe traumatic brain injury (TBI) relative to massed restudy (MR) and spaced restudy (SR). 3(learning condition: MR, SR, RP)×2(delayed recall: 30min, 1wk) within-subject experiment. Nonprofit medical rehabilitation research center. Memory-impaired (<5th percentile) survivors of severe TBI (N=10). During RP, patients are quizzed on to-be-learned information shortly after it is presented, such that patients practice retrieval. MR consists of repeated restudy (ie, cramming). SR consists of restudy trials separated in time (ie, distributed learning). Forty-eight verbal paired associates (VPAs) were equally divided across 3 learning conditions (16 per condition). Delayed recall for one half of the VPAs was assessed after 30 minutes (8 per condition) and for the other half after 1 week (8 per condition). There was a large effect of learning condition after the short delay (P<.001, η(2)=.72), with much better recall of VPAs studied through RP (46.3%) relative to MR (12.5%) and SR (15.0%). This large effect of learning condition remained after the long delay (P=.001, η(2)=.56), as patients recalled 11.3% of the VPAs studied through RP, but nothing through MR (0.0%) and only 1.3% through SR. That is, RP was essentially the only learning condition to result in successful recall after 1 week, with most patients recalling at least 1 VPA. The robust effect of RP among TBI survivors with severe memory impairment engenders confidence that this strategy would work outside the laboratory to improve memory in real-life settings. Future randomized controlled trials of RP training are needed. Copyright © 2014 American Congress of Rehabilitation Medicine. Published by Elsevier Inc. All rights reserved.

Continental igneous rock composition: A major control of past global chemical weathering

PubMed Central

Bataille, Clément P.; Willis, Amy; Yang, Xiao; Liu, Xiao-Ming

2017-01-01

The composition of igneous rocks in the continental crust has changed throughout Earth’s history. However, the impact of these compositional variations on chemical weathering, and by extension on seawater and atmosphere evolution, is largely unknown. We use the strontium isotope ratio in seawater [(87Sr/86Sr)seawater] as a proxy for chemical weathering, and we test the sensitivity of (87Sr/86Sr)seawater variations to the strontium isotopic composition (87Sr/86Sr) in igneous rocks generated through time. We demonstrate that the 87Sr/86Sr ratio in igneous rocks is correlated to the epsilon hafnium (εHf) of their hosted zircon grains, and we use the detrital zircon record to reconstruct the evolution of the 87Sr/86Sr ratio in zircon-bearing igneous rocks. The reconstructed 87Sr/86Sr variations in igneous rocks are strongly correlated with the (87Sr/86Sr)seawater variations over the last 1000 million years, suggesting a direct control of the isotopic composition of silicic magmatism on (87Sr/86Sr)seawater variations. The correlation decreases during several time periods, likely reflecting changes in the chemical weathering rate associated with paleogeographic, climatic, or tectonic events. We argue that for most of the last 1000 million years, the (87Sr/86Sr)seawater variations are responding to changes in the isotopic composition of silicic magmatism rather than to changes in the global chemical weathering rate. We conclude that the (87Sr/86Sr)seawater variations are of limited utility to reconstruct changes in the global chemical weathering rate in deep times. PMID:28345044

High Sr/Y rocks are not all adakites!

NASA Astrophysics Data System (ADS)

Moyen, Jean-François

2010-05-01

The name of "adakite" is used to describe a far too large group of rocks, whose sole common feature is high Sr/Y and La/Yb ratios. Defining adakites only by this criterion is misleading, as the definition of this group of rocks does include many other criteria, including major elements. In itself, high (or commonly moderate!) Sr/Y ratios can be achieved via different processes: melting of a high Sr/Y (and La/Yb) source; deep melting, with abundant residual garnet; fractional crystallization or AFC; or interactions of felsic melts with the mantle, causing selective enrichment in LREE and Sr over HREE. A database of the compositions of "adakitic" rocks - including "high silica" and "low silica" adakites, "continental" adakites and Archaean adakites—was assembled. Geochemical modeling of the potential processes is used to interpret it, and reveals that (1) the genesis of high-silica adakites requires high pressure evolution (be it by melting or fractionation), in equilibrium with large amounts of garnet; (2) low-silica adakites are explained by garnet-present melting of an adakite-metasomatized mantle, i.e at depths greater than 2.5 GPa; (3) "Continental" adakites is a term encompassing a huge range of rocks, with a corresponding diversity of petrogenetic processes, and most of them are different from both low- and high- silica adakites; in fact in many cases it is a complete misnomer and the rocks studied are high-K calc-alkaline granitoids or even S-type granites; (4) Archaean adakites show a bimodal composition range, with some very high Sr/Y examples (similar to part of the TTG suite) reflecting deep melting (> 2.0 GPa) of a basaltic source with a relatively high Sr/Y, while lower Sr/Y rocks formed by shallower (1.0 GPa) melting of similar sources. Comparison with the Archaean TTG suite highlights the heterogeneity of the TTGs, whose composition spreads the whole combined range of HSA and Archaean adakites, pointing to a diversity of sources and processes contributing to the "TTG suite".

Positron annihilation studies on the behaviour of vacancies in LaAlO3/SrTiO3 heterostructures

NASA Astrophysics Data System (ADS)

Yuan, Guoliang; Li, Chen; Yin, Jiang; Liu, Zhiguo; Wu, Di; Uedono, Akira

2012-11-01

The formation and diffusion of vacancies are studied in LaAlO3/SrTiO3 heterostructures. Oxygen vacancies (VOS) appear easily in the SrTiO3 substrate during LaAlO3 film growth at 700 °C and 10-4 Pa oxygen pressure rather than at 10-3-10-1 Pa, thus the latter two-dimensional electron gas should come from the polarity discontinuity at the (LaO)+/(TiO2)0 interface. For SrTiO3-δ/LaAlO3/SrTiO3, high-density VOS of the SrTiO3-δ film can pass through the LaAlO3 film and then diffuse to 1.7 µm depth in the SrTiO3 substrate, suggesting that LaAlO3 has VOS at its middle-deep energy levels within the band gap. Moreover, high-density VOS may combine with a strontium/titanium vacancy (VSr/Ti) to form VSr/Ti-O complexes in the SrTiO3 substrate at 700 °C.

Students' Self-Regulation for Interaction with Others in Online Learning Environments

ERIC Educational Resources Information Center

Cho, Moon-Heum; Kim, B. Joon

2013-01-01

The purpose of this study was to explore variables explaining students' self-regulation (SR) for interaction with others, specifically peers and instructors, in online learning environments. A total of 407 students participated in the study. With hierarchical regression model (HRM), several variables were regressed on students' SR for interaction…

Learning "Social Responsibility" in the Workplace: Conjuring, Unsettling, and Folding Boundaries

ERIC Educational Resources Information Center

Fenwick, Tara

2011-01-01

This article proceeds from the argument that while the discourse of social responsibility (SR) is increasingly evident in pedagogies circulating through the workplace, its actual practices tend to be obscured beneath complex tensions and moral precepts presented as self-evident. Through an examination of individuals' learning of SR in the…

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

Rigali, Mark J.; Pye, Steven; Hardin, Ernest

This study considers the feasibility of large diameter deep boreholes for waste disposal. The conceptual approach considers examples of deep large diameter boreholes that have been successfully drilled, and also other deep borehole designs proposed in the literature. The objective for large diameter boreholes would be disposal of waste packages with diameters of 22 to 29 inches, which could enable disposal of waste forms such as existing vitrified high level waste. A large-diameter deep borehole design option would also be amenable to other waste forms including calcine waste, treated Na-bonded and Na-bearing waste, and Cs and Sr capsules.

Learning curve of speech recognition.

PubMed

Kauppinen, Tomi A; Kaipio, Johanna; Koivikko, Mika P

2013-12-01

Speech recognition (SR) speeds patient care processes by reducing report turnaround times. However, concerns have emerged about prolonged training and an added secretarial burden for radiologists. We assessed how much proofing radiologists who have years of experience with SR and radiologists new to SR must perform, and estimated how quickly the new users become as skilled as the experienced users. We studied SR log entries for 0.25 million reports from 154 radiologists and after careful exclusions, defined a group of 11 experienced radiologists and 71 radiologists new to SR (24,833 and 122,093 reports, respectively). Data were analyzed for sound file and report lengths, character-based error rates, and words unknown to the SR's dictionary. Experienced radiologists corrected 6 characters for each report and for new users, 11. Some users presented a very unfavorable learning curve, with error rates not declining as expected. New users' reports were longer, and data for the experienced users indicates that their reports, initially equally lengthy, shortened over a period of several years. For most radiologists, only minor corrections of dictated reports were necessary. While new users adopted SR quickly, with a subset outperforming experienced users from the start, identification of users struggling with SR will help facilitate troubleshooting and support.

Searching for prostate cancer by fully automated magnetic resonance imaging classification: deep learning versus non-deep learning.

PubMed

Wang, Xinggang; Yang, Wei; Weinreb, Jeffrey; Han, Juan; Li, Qiubai; Kong, Xiangchuang; Yan, Yongluan; Ke, Zan; Luo, Bo; Liu, Tao; Wang, Liang

2017-11-13

Prostate cancer (PCa) is a major cause of death since ancient time documented in Egyptian Ptolemaic mummy imaging. PCa detection is critical to personalized medicine and varies considerably under an MRI scan. 172 patients with 2,602 morphologic images (axial 2D T2-weighted imaging) of the prostate were obtained. A deep learning with deep convolutional neural network (DCNN) and a non-deep learning with SIFT image feature and bag-of-word (BoW), a representative method for image recognition and analysis, were used to distinguish pathologically confirmed PCa patients from prostate benign conditions (BCs) patients with prostatitis or prostate benign hyperplasia (BPH). In fully automated detection of PCa patients, deep learning had a statistically higher area under the receiver operating characteristics curve (AUC) than non-deep learning (P = 0.0007 < 0.001). The AUCs were 0.84 (95% CI 0.78-0.89) for deep learning method and 0.70 (95% CI 0.63-0.77) for non-deep learning method, respectively. Our results suggest that deep learning with DCNN is superior to non-deep learning with SIFT image feature and BoW model for fully automated PCa patients differentiation from prostate BCs patients. Our deep learning method is extensible to image modalities such as MR imaging, CT and PET of other organs.

A dissociation of dorso-lateral striatum and amygdala function on the same stimulus-response habit task.

PubMed

McDonald, R J; Hong, N S

2004-01-01

This experiment tested the idea that the amygdala-based learning and memory system covertly acquires a stimulus-reward (stimulus-outcome) association during acquisition of a stimulus-response (S-R) habit task developed for the eight-arm radial maze. Groups of rats were given dorso-lateral striatal or amygdala lesions and then trained on the S-R habit task on the eight-arm radial maze. Rats with neurotoxic damage to the dorso-lateral striatum were severely impaired on the acquisition of the S-R habit task but showed a conditioned-cue preference for the stimulus reinforced during S-R habit training. Rats with neurotoxic damage to the amygdala were able to acquire the S-R habit task but did not show a conditioned-cue preference for the stimulus reinforced during S-R habit training. This pattern of results represents a dissociation of learning and memory functions of the dorsal striatum and amygdala on the same task.

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.

Text feature extraction based on deep learning: a review.

PubMed

Liang, Hong; Sun, Xiao; Sun, Yunlei; Gao, Yuan

2017-01-01

Selection of text feature item is a basic and important matter for text mining and information retrieval. Traditional methods of feature extraction require handcrafted features. To hand-design, an effective feature is a lengthy process, but aiming at new applications, deep learning enables to acquire new effective feature representation from training data. As a new feature extraction method, deep learning has made achievements in text mining. The major difference between deep learning and conventional methods is that deep learning automatically learns features from big data, instead of adopting handcrafted features, which mainly depends on priori knowledge of designers and is highly impossible to take the advantage of big data. Deep learning can automatically learn feature representation from big data, including millions of parameters. This thesis outlines the common methods used in text feature extraction first, and then expands frequently used deep learning methods in text feature extraction and its applications, and forecasts the application of deep learning in feature extraction.

Effect of retrapping on the persistent luminescence in strontium silicate orange–yellow phosphor

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

Xu, Xuhui; Yu, Xue, E-mail: yuyu6593@126.com; Zhou, Dacheng

2013-10-15

The orange–yellow long persistent luminescence in Sr{sub 3}SiO{sub 5}:Eu{sup 2+}, Er{sup 3+} with the chromaticity coordination of (0.48, 0.49) can persist for over 20 h above the recognizable intensity level (≥0.32 mcd/m{sup 2}) because of retrapping carriers by the deep traps. The incorporation of Er{sup 3+} into Sr{sub 3}SiO{sub 5}:Eu{sup 2+} generates a large number of shallow traps responsible for the fast decay component as well as deep traps responsible for the decay tail of the LPL. It demonstrates that the retrapping of the carrier released from a trap plays an important role in the persistent luminescence process. - Graphicalmore » abstract: LPL decay curves of Sr{sub 3−x−y}SiO{sub 5}:xEu{sup 2+}, yEr{sup 3+} (x=0.0025, y=0, 0.0025). Inset: Orange–yellow emission images recorded using a classic Reflex digital camera with exposure times varying with the persistent luminescence times. Display Omitted - Highlights: • The persistence time of Sr{sub 3}SiO{sub 5}:Eu{sup 2+}, Er{sup 3+} lasts over 20 h above the recognizable intensity level. • The incorporation of Er{sup 3+} into Sr{sub 3}SiO{sub 5}:Eu{sup 2+} generates a large number of shallow traps. • The experimental results provide an evidence for the retrapping process in LPL processes.« less

Reframing the Principle of Specialisation in Legitimation Code Theory: A Blended Learning Perspective

ERIC Educational Resources Information Center

Owusu-Agyeman, Yaw; Larbi-Siaw, Otu

2017-01-01

This study argues that in developing a robust framework for students in a blended learning environment, Structural Alignment (SA) becomes the third principle of specialisation in addition to Epistemic Relation (ER) and Social Relation (SR). We provide an extended code: (ER+/-, SR+/-, SA+/-) that present strong classification and framing to the…

Overview of deep learning in medical imaging.

PubMed

Suzuki, Kenji

2017-09-01

The use of machine learning (ML) has been increasing rapidly in the medical imaging field, including computer-aided diagnosis (CAD), radiomics, and medical image analysis. Recently, an ML area called deep learning emerged in the computer vision field and became very popular in many fields. It started from an event in late 2012, when a deep-learning approach based on a convolutional neural network (CNN) won an overwhelming victory in the best-known worldwide computer vision competition, ImageNet Classification. Since then, researchers in virtually all fields, including medical imaging, have started actively participating in the explosively growing field of deep learning. In this paper, the area of deep learning in medical imaging is overviewed, including (1) what was changed in machine learning before and after the introduction of deep learning, (2) what is the source of the power of deep learning, (3) two major deep-learning models: a massive-training artificial neural network (MTANN) and a convolutional neural network (CNN), (4) similarities and differences between the two models, and (5) their applications to medical imaging. This review shows that ML with feature input (or feature-based ML) was dominant before the introduction of deep learning, and that the major and essential difference between ML before and after deep learning is the learning of image data directly without object segmentation or feature extraction; thus, it is the source of the power of deep learning, although the depth of the model is an important attribute. The class of ML with image input (or image-based ML) including deep learning has a long history, but recently gained popularity due to the use of the new terminology, deep learning. There are two major models in this class of ML in medical imaging, MTANN and CNN, which have similarities as well as several differences. In our experience, MTANNs were substantially more efficient in their development, had a higher performance, and required a lesser number of training cases than did CNNs. "Deep learning", or ML with image input, in medical imaging is an explosively growing, promising field. It is expected that ML with image input will be the mainstream area in the field of medical imaging in the next few decades.

Optimized multiple linear mappings for single image super-resolution

NASA Astrophysics Data System (ADS)

Zhang, Kaibing; Li, Jie; Xiong, Zenggang; Liu, Xiuping; Gao, Xinbo

2017-12-01

Learning piecewise linear regression has been recognized as an effective way for example learning-based single image super-resolution (SR) in literature. In this paper, we employ an expectation-maximization (EM) algorithm to further improve the SR performance of our previous multiple linear mappings (MLM) based SR method. In the training stage, the proposed method starts with a set of linear regressors obtained by the MLM-based method, and then jointly optimizes the clustering results and the low- and high-resolution subdictionary pairs for regression functions by using the metric of the reconstruction errors. In the test stage, we select the optimal regressor for SR reconstruction by accumulating the reconstruction errors of m-nearest neighbors in the training set. Thorough experimental results carried on six publicly available datasets demonstrate that the proposed SR method can yield high-quality images with finer details and sharper edges in terms of both quantitative and perceptual image quality assessments.

Deep Learning in Nuclear Medicine and Molecular Imaging: Current Perspectives and Future Directions.

PubMed

Choi, Hongyoon

2018-04-01

Recent advances in deep learning have impacted various scientific and industrial fields. Due to the rapid application of deep learning in biomedical data, molecular imaging has also started to adopt this technique. In this regard, it is expected that deep learning will potentially affect the roles of molecular imaging experts as well as clinical decision making. This review firstly offers a basic overview of deep learning particularly for image data analysis to give knowledge to nuclear medicine physicians and researchers. Because of the unique characteristics and distinctive aims of various types of molecular imaging, deep learning applications can be different from other fields. In this context, the review deals with current perspectives of deep learning in molecular imaging particularly in terms of development of biomarkers. Finally, future challenges of deep learning application for molecular imaging and future roles of experts in molecular imaging will be discussed.

Why large porphyry Cu deposits like high Sr/Y magmas?

PubMed Central

Chiaradia, Massimo; Ulianov, Alexey; Kouzmanov, Kalin; Beate, Bernardo

2012-01-01

Porphyry systems supply most copper and significant gold to our economy. Recent studies indicate that they are frequently associated with high Sr/Y magmatic rocks, but the meaning of this association remains elusive. Understanding the association between high Sr/Y magmatic rocks and porphyry-type deposits is essential to develop genetic models that can be used for exploration purposes. Here we present results on a Pleistocene volcano of Ecuador that highlight the behaviour of copper in magmas with variable (but generally high) Sr/Y values. We provide indirect evidence for Cu partitioning into a fluid phase exsolved at depths of ~15 km from high Sr/Y (>70) andesitic magmas before sulphide saturation. This lends support to the hypothesis that large amounts of Cu- and S-bearing fluids can be accumulated into and released from a long-lived high Sr/Y deep andesitic reservoir to a shallower magmatic-hydrothermal system with the potential of generating large porphyry-type deposits. PMID:23008750

Why large porphyry Cu deposits like high Sr/Y magmas?

PubMed

Chiaradia, Massimo; Ulianov, Alexey; Kouzmanov, Kalin; Beate, Bernardo

2012-01-01

Porphyry systems supply most copper and significant gold to our economy. Recent studies indicate that they are frequently associated with high Sr/Y magmatic rocks, but the meaning of this association remains elusive. Understanding the association between high Sr/Y magmatic rocks and porphyry-type deposits is essential to develop genetic models that can be used for exploration purposes. Here we present results on a Pleistocene volcano of Ecuador that highlight the behaviour of copper in magmas with variable (but generally high) Sr/Y values. We provide indirect evidence for Cu partitioning into a fluid phase exsolved at depths of ~15 km from high Sr/Y (>70) andesitic magmas before sulphide saturation. This lends support to the hypothesis that large amounts of Cu- and S-bearing fluids can be accumulated into and released from a long-lived high Sr/Y deep andesitic reservoir to a shallower magmatic-hydrothermal system with the potential of generating large porphyry-type deposits.

Using 87Sr/86Sr ratios to investigate changes in stream chemistry during snowmelt in the Provo River, Utah, USA

NASA Astrophysics Data System (ADS)

Hale, C. A.; Carling, G. T.; Fernandez, D. P.; Nelson, S.; Aanderud, Z.; Tingey, D. G.; Dastrup, D.

2017-12-01

Water chemistry in mountain streams is variable during spring snowmelt as shallow groundwater flow paths are activated in the watershed, introducing solutes derived from soil water. Sr isotopes and other tracers can be used to differentiate waters that have interacted with soils and dust (shallow groundwater) and bedrock (deep groundwater). To investigate processes controlling water chemistry during snowmelt, we analyzed 87Sr/86Sr ratios, Sr and other trace element concentrations in bulk snowpack, dust, soil, soil water, ephemeral channels, and river water during snowmelt runoff in the upper Provo River watershed in northern Utah, USA, over four years (2014-2017). Strontium concentrations in the river averaged 20 ppb during base flow and decreased to 10 ppb during snowmelt runoff. 87Sr/86Sr ratios were around 0.717 during base flow and decreased to 0.715 in 2014 and 0.713 in 2015 and 2016 during snowmelt, trending towards less radiogenic values of mineral dust inputs in the Uinta Mountain soils. Ephemeral channels, representing shallow flow paths with soil water inputs, had Sr concentrations between 7-20 ppb and 87Sr/86Sr ratios between 0.713-0.716. Snowpack Sr concentrations were generally <2 ppb with 87Sr/86Sr ratios between 0.710-711, similar to atmospheric dust inputs. The less radiogenic 87Sr/86Sr ratios and lower Sr concentrations in the river during snowmelt are likely a result of activating shallow groundwater flow paths, which allows melt water to interact with shallow soils that contain accumulated dust deposits with a less radiogenic 87Sr/86Sr ratio. These results suggest that flow paths and atmospheric dust are important to consider when investigating variable solute loads in mountain streams.

Virtual screening of inorganic materials synthesis parameters with deep learning

NASA Astrophysics Data System (ADS)

Kim, Edward; Huang, Kevin; Jegelka, Stefanie; Olivetti, Elsa

2017-12-01

Virtual materials screening approaches have proliferated in the past decade, driven by rapid advances in first-principles computational techniques, and machine-learning algorithms. By comparison, computationally driven materials synthesis screening is still in its infancy, and is mired by the challenges of data sparsity and data scarcity: Synthesis routes exist in a sparse, high-dimensional parameter space that is difficult to optimize over directly, and, for some materials of interest, only scarce volumes of literature-reported syntheses are available. In this article, we present a framework for suggesting quantitative synthesis parameters and potential driving factors for synthesis outcomes. We use a variational autoencoder to compress sparse synthesis representations into a lower dimensional space, which is found to improve the performance of machine-learning tasks. To realize this screening framework even in cases where there are few literature data, we devise a novel data augmentation methodology that incorporates literature synthesis data from related materials systems. We apply this variational autoencoder framework to generate potential SrTiO3 synthesis parameter sets, propose driving factors for brookite TiO2 formation, and identify correlations between alkali-ion intercalation and MnO2 polymorph selection.

Mg and Sr Incorporation in Foraminifer Shells: Patterns, Controls and Applications.

NASA Astrophysics Data System (ADS)

Lea, D. W.

2001-12-01

The incorporation of Mg and Sr in planktonic and benthic foraminifer shells is important for paleoceanographic research because of the potential to record physical and chemical changes in the oceanic environment. Pelagic shells are 99%+ CaCO3, and abundances of Mg and Sr are typically ~0.1%, requiring sensitive quantification methods such as ICP-MS or AES. Mg/Ca values range from 0.5 mmol/mol in cold planktics and benthics to ~5 mmol/mol in tropical planktics, with some species (Orbulina universa) having even higher values. The main control on Mg incorporation is temperature, but pH and salinity also exert small influences, presumably through calcification rate. The Mg/Ca content of the primary ontogenetic calcite can be altered by the addition of so-called gametogenic calcite, generally deposited in deep, colder waters. After deposition on the seafloor, dissolution becomes the main influence, with progressively lower Mg/Ca values in more dissolved samples. This loss appears to occur by preferential loss of the more Mg-rich portions of the shell, although the details remain unexplained. Sr/Ca values range from 0.9 in some benthic species (Uvigerina spp.) to 1.6 mmol/mol in some planktics. Culturing results suggest that temperature, salinity and pH all exert a weak control (i.e., 1% per ° C) on shell Sr, presumably through a kinetic effect. The main control appears to be related to environmental differences. For example, comparison of Sr/Ca in Neogloboquadrina pachyderma from plankton tows and cultures with core-top specimens indicates that the latter have significantly higher values, presumably due to deep crusting, perhaps added with a much higher calcification rate. This observation clearly demonstrates that Sr/Ca is not simply related to a single physical parameter such as temperature. Downcore records of shell Mg/Ca and Sr/Ca reveal substantial variability that can be correlated with known paleoceanographic change. For Mg/Ca, observed variations can largely be explained by climate-related variations in temperature. For Sr/Ca, it appears that observed variations related to secular changes in seawater Sr/Ca, but this cannot be fully substantiated without a more complete understanding of primary and post-depositional controls on shell composition.

Isotopic and geochemical characterization of fossil brines of the Cambrian Mt. Simon Sandstone and Ironton-Galesville Formation from the Illinois Basin, USA

NASA Astrophysics Data System (ADS)

Labotka, Dana M.; Panno, Samuel V.; Locke, Randall A.; Freiburg, Jared T.

2015-09-01

Geochemical and isotopic characteristics of deep-seated saline groundwater provide valuable insight into the origin and evolving composition, water-rock interaction, and mixing potential of fossil brines. Such information may yield insight into intra- and interbasinal brine movement and relationships between brine evolution and regional groundwater flow systems. This investigation reports on the δ18O and δD composition and activity values, 87Sr/86Sr ratios and Sr concentrations, and major ion concentrations of the Cambrian-hosted brines of the Mt. Simon Sandstone and Ironton-Galesville Formation and discusses the evolution of these brines as they relate to other intracontinental brines. Brines in the Illinois Basin are dominated by Na-Ca-Cl-type chemistry. The Mt. Simon and overlying Ironton-Galesville brines exhibit total dissolved solids concentrations of ∼195,000 mg/L and ∼66,270 mg/L, respectively. The δD of brine composition of the Mt. Simon ranges from -34‰ to -22‰ (V-SMOW), and the Ironton-Galesville is ∼-53.2‰ (V-SMOW). The δ18O composition of the Mt. Simon brine ranges from -5.0‰ to -2.8‰ (V-SMOW), and the Ironton-Galesville brine is ∼-6.9‰ (V-SMOW). The 87Sr/86Sr values in the Mt. Simon brine range from 0.7110 to 0.7116. The less radiogenic Ironton-Galesville brine has an average 87Sr/86Sr value of 0.7107. Evaluation of δ18O and δD composition and activities and 87Sr/86Sr ratios suggests that the Mt. Simon brine is likely connate seawater and recirculating deep-seated brines that have been diluted with meteoric water and influenced by the dissolution of evaporites with a minimal halite contribution based on Cl/Br ratios. The Ironton-Galesville brine is also likely originally connate seawater that mixed with other brines and meteoric waters, including possibly Pleistocene glacial recharge. The Ca-excess vs. Na-deficiency comparison with the Basinal Fluid Line suggests the Mt. Simon and Ironton-Galesville brines have been influenced by the effects of albitization and plot very close to the Basinal Fluid Line. These Cambrian-hosted brines appear to have a different albitization history than other regional basin brines and a strong component of seawater. The Ironton-Galesville brine appears more geochemically associated with other Illinois Basin brines than the Mt. Simon brine which appears more geochemically conservative. Comparisons with other extrabasinal North American brines suggest that the Michigan basin brines are geochemically most similar to the Mt. Simon brines with the exception of the influence from carbonates in the Michigan Basin. Analyses of 87Sr/86Sr values in the Mt. Simon brine suggest that brine Sr has isotopically equilibrated with clay minerals in the Lower Mt. Simon and underlying bedrock formations and not with whole rock suggesting the influence of recirculating brines from the crystalline basement. Overall, the geochemistry of these Cambrian-hosted brines suggests an evolution from original seawater-like compositions. This investigation shows that intracratonic basins do not behave as closed systems but can be strongly affected by water-rock interaction and regional groundwater flow systems that circulate deep crystalline basement brines and brines from nearby basins.

Large-scale Labeled Datasets to Fuel Earth Science Deep Learning Applications

NASA Astrophysics Data System (ADS)

Maskey, M.; Ramachandran, R.; Miller, J.

2017-12-01

Deep learning has revolutionized computer vision and natural language processing with various algorithms scaled using high-performance computing. However, generic large-scale labeled datasets such as the ImageNet are the fuel that drives the impressive accuracy of deep learning results. Large-scale labeled datasets already exist in domains such as medical science, but creating them in the Earth science domain is a challenge. While there are ways to apply deep learning using limited labeled datasets, there is a need in the Earth sciences for creating large-scale labeled datasets for benchmarking and scaling deep learning applications. At the NASA Marshall Space Flight Center, we are using deep learning for a variety of Earth science applications where we have encountered the need for large-scale labeled datasets. We will discuss our approaches for creating such datasets and why these datasets are just as valuable as deep learning algorithms. We will also describe successful usage of these large-scale labeled datasets with our deep learning based applications.

Document Analyses of Student Use of a Blogging-Mapping Tool to Explore Evidence of Deep and Reflective Learning

ERIC Educational Resources Information Center

Xie, Ying

2008-01-01

Theories about reflective thinking and deep-surface learning abound. In order to arrive at the definition for "reflective thinking toward deep learning," this study establishes that reflective thinking toward deep learning refers to a learner's purposeful and conscious activity of manipulating ideas toward meaningful learning and knowledge…

Electrophysiological correlates of reinforcement learning in young people with Tourette syndrome with and without co-occurring ADHD symptoms.

PubMed

Shephard, Elizabeth; Jackson, Georgina M; Groom, Madeleine J

2016-06-01

Altered reinforcement learning is implicated in the causes of Tourette syndrome (TS) and attention-deficit/hyperactivity disorder (ADHD). TS and ADHD frequently co-occur but how this affects reinforcement learning has not been investigated. We examined the ability of young people with TS (n=18), TS+ADHD (N=17), ADHD (n=13) and typically developing controls (n=20) to learn and reverse stimulus-response (S-R) associations based on positive and negative reinforcement feedback. We used a 2 (TS-yes, TS-no)×2 (ADHD-yes, ADHD-no) factorial design to assess the effects of TS, ADHD, and their interaction on behavioural (accuracy, RT) and event-related potential (stimulus-locked P3, feedback-locked P2, feedback-related negativity, FRN) indices of learning and reversing the S-R associations. TS was associated with intact learning and reversal performance and largely typical ERP amplitudes. ADHD was associated with lower accuracy during S-R learning and impaired reversal learning (significantly reduced accuracy and a trend for smaller P3 amplitude). The results indicate that co-occurring ADHD symptoms impair reversal learning in TS+ADHD. The implications of these findings for behavioural tic therapies are discussed. Copyright © 2016 ISDN. Published by Elsevier Ltd. All rights reserved.

Deep and surface learning in problem-based learning: a review of the literature.

PubMed

Dolmans, Diana H J M; Loyens, Sofie M M; Marcq, Hélène; Gijbels, David

2016-12-01

In problem-based learning (PBL), implemented worldwide, students learn by discussing professionally relevant problems enhancing application and integration of knowledge, which is assumed to encourage students towards a deep learning approach in which students are intrinsically interested and try to understand what is being studied. This review investigates: (1) the effects of PBL on students' deep and surface approaches to learning, (2) whether and why these effects do differ across (a) the context of the learning environment (single vs. curriculum wide implementation), and (b) study quality. Studies were searched dealing with PBL and students' approaches to learning. Twenty-one studies were included. The results indicate that PBL does enhance deep learning with a small positive average effect size of .11 and a positive effect in eleven of the 21 studies. Four studies show a decrease in deep learning and six studies show no effect. PBL does not seem to have an effect on surface learning as indicated by a very small average effect size (.08) and eleven studies showing no increase in the surface approach. Six studies demonstrate a decrease and four an increase in surface learning. It is concluded that PBL does seem to enhance deep learning and has little effect on surface learning, although more longitudinal research using high quality measurement instruments is needed to support this conclusion with stronger evidence. Differences cannot be explained by the study quality but a curriculum wide implementation of PBL has a more positive impact on the deep approach (effect size .18) compared to an implementation within a single course (effect size of -.05). PBL is assumed to enhance active learning and students' intrinsic motivation, which enhances deep learning. A high perceived workload and assessment that is perceived as not rewarding deep learning are assumed to enhance surface learning.

Training in summarizing notes: Effects of teaching students a self-regulation study strategy in science learning

NASA Astrophysics Data System (ADS)

Nebres, Michelle

The last two decades of national data assessments reveal that there has been a sharp decline in nationwide standardized test scores. International assessment data show that in 2012 a very low amount of American students were performing at proficiency or above in science literacy. Research in science literacy education suggests that students benefit most when they are self-regulated (SR) learners. Unfortunately, SR poses a challenge for many students because students lack these skills. The effects of having learned few SR strategies at an early age may lead to long term learning difficulties--preventing students from achieving academic success in college and beyond. As a result, some researchers have begun to investigate how to best support students' SR skills. In order for studying to be successful, students need to know which SR study strategies to implement. This can be tricky for struggling students because they need study strategies that are well defined. This needs to be addressed through effective classroom instruction, and should be addressed prior to entering high school in order for students to be prepared for higher level learning. In this study, students underwent a treatment in which they were taught a SR study strategy called summarizing notes. A crossover repeated measures design was employed to understand the effectiveness of the treatment. Results indicated a weak, but positive correlation between how well students summarized notes and how well they performed on science tests. Self-regulation skills are needed because these are the types of skills young adults will use as they enter the workforce. As young adults began working in a professional setting, they will be expected to know how to observe and become proficient on their own. This study is pertinent to the educational field because it is an opportunity for students to increase SR, which affords students with the skills needed to be a lifelong learner.

Life Cycle of Midlatitude Deep Convective Systems in a Lagrangian Framework

NASA Technical Reports Server (NTRS)

Feng, Zhe; Dong, Xiquan; Xie, Baike; McFarlane, Sally A.; Kennedy, Aaron; Lin, Bing; Minnis, Patrick

2012-01-01

Deep Convective Systems (DCSs) consist of intense convective cores (CC), large stratiform rain (SR) regions, and extensive non-precipitating anvil clouds (AC). This study focuses on the evolution of these three components and the factors that affect convective AC production. An automated satellite tracking method is used in conjunction with a recently developed multi-sensor hybrid classification to analyze the evolution of DCS structure in a Lagrangian framework over the central United States. Composite analysis from 4221 tracked DCSs during two warm seasons (May-August, 2010-2011) shows that maximum system size correlates with lifetime, and longer-lived DCSs have more extensive SR and AC. Maximum SR and AC area lag behind peak convective intensity and the lag increases linearly from approximately 1-hour for short-lived systems to more than 3-hours for long-lived ones. The increased lag, which depends on the convective environment, suggests that changes in the overall diabatic heating structure associated with the transition from CC to SR and AC could prolong the system lifetime by sustaining stratiform cloud development. Longer-lasting systems are associated with up to 60% higher mid-tropospheric relative humidity and up to 40% stronger middle to upper tropospheric wind shear. Regression analysis shows that the areal coverage of thick AC is strongly correlated with the size of CC, updraft strength, and SR area. Ambient upper tropospheric wind speed and wind shear also play an important role for convective AC production where for systems with large AC (radius greater than 120-km) they are 24% and 20% higher, respectively, than those with small AC (radius=20 km).

The Change in Black Sea Water Composition and Hydrology during Deglaciation from Multiproxy Reconstructions

NASA Astrophysics Data System (ADS)

Yanchilina, A.; Ryan, W. B. F.; McManus, J. F.

2014-12-01

This study presents a reconstruction of changes in the water column from the last glacial into the early Holocene using stable isotope, 87Sr/86Sr, 14C, and trace element ratios from mollusks from the shelf area and ostracods from the basin of the Black Sea. The stable isotope record is compared to a thoroughly U/Th dated terrestrial stable isotope record of a nearby cave, Sofular cave in northwestern Turkey. The combination of deep, surface, and terrestrial signals gives valuable insight towards the behavior of the lake water during the deglaciation in multiple dimensions, specifically the water column stratification and hydrological dynamics. The comparison of the stable isotope records of two independent proxies allows to make inferences on the changes in the 14C reservoir of the Black Sea-Lake. Results show that during the glacial period, water from the Black Sea-Lake was outflowing to the Sea of Marmara but the ventilation of the water column was weak as old 14C was not removed and allowed to accumulate giving the lake a large 14C reservoir age. A deglacial pulse of meltwater released from the Eurasian Fennoscandian pro-glacial lakes increased ventilation of the water column. This is seen in lighter δ18O and a spike in radiogenic 87Sr/86Sr in both deep and shallow parts of the water column. The dynamic ventilation and outflow of water into the Sea of Marmara continued until the onset of the Bolling/Allerod as the radiogenic 87Sr/86Sr was almost completely flushed out in a couple hundred years time. During the Bolling/Allerod and Preboreal warming, δ18O got heavier whereas the 87Sr/86Sr stayed constant and the 14C again accumulated and contributed to an older reservoir age. The Younger Dryas period, sandwiched in between the two warming periods, shows a return to glacial conditions in the δ13C and that the water outflowed to the Sea of Marmara as the δ18O only showed a slight change towards a more heavy value.

Deep imitation learning for 3D navigation tasks.

PubMed

Hussein, Ahmed; Elyan, Eyad; Gaber, Mohamed Medhat; Jayne, Chrisina

2018-01-01

Deep learning techniques have shown success in learning from raw high-dimensional data in various applications. While deep reinforcement learning is recently gaining popularity as a method to train intelligent agents, utilizing deep learning in imitation learning has been scarcely explored. Imitation learning can be an efficient method to teach intelligent agents by providing a set of demonstrations to learn from. However, generalizing to situations that are not represented in the demonstrations can be challenging, especially in 3D environments. In this paper, we propose a deep imitation learning method to learn navigation tasks from demonstrations in a 3D environment. The supervised policy is refined using active learning in order to generalize to unseen situations. This approach is compared to two popular deep reinforcement learning techniques: deep-Q-networks and Asynchronous actor-critic (A3C). The proposed method as well as the reinforcement learning methods employ deep convolutional neural networks and learn directly from raw visual input. Methods for combining learning from demonstrations and experience are also investigated. This combination aims to join the generalization ability of learning by experience with the efficiency of learning by imitation. The proposed methods are evaluated on 4 navigation tasks in a 3D simulated environment. Navigation tasks are a typical problem that is relevant to many real applications. They pose the challenge of requiring demonstrations of long trajectories to reach the target and only providing delayed rewards (usually terminal) to the agent. The experiments show that the proposed method can successfully learn navigation tasks from raw visual input while learning from experience methods fail to learn an effective policy. Moreover, it is shown that active learning can significantly improve the performance of the initially learned policy using a small number of active samples.

Deep learning with convolutional neural network in radiology.

PubMed

Yasaka, Koichiro; Akai, Hiroyuki; Kunimatsu, Akira; Kiryu, Shigeru; Abe, Osamu

2018-04-01

Deep learning with a convolutional neural network (CNN) is gaining attention recently for its high performance in image recognition. Images themselves can be utilized in a learning process with this technique, and feature extraction in advance of the learning process is not required. Important features can be automatically learned. Thanks to the development of hardware and software in addition to techniques regarding deep learning, application of this technique to radiological images for predicting clinically useful information, such as the detection and the evaluation of lesions, etc., are beginning to be investigated. This article illustrates basic technical knowledge regarding deep learning with CNNs along the actual course (collecting data, implementing CNNs, and training and testing phases). Pitfalls regarding this technique and how to manage them are also illustrated. We also described some advanced topics of deep learning, results of recent clinical studies, and the future directions of clinical application of deep learning techniques.

Recent machine learning advancements in sensor-based mobility analysis: Deep learning for Parkinson's disease assessment.

PubMed

Eskofier, Bjoern M; Lee, Sunghoon I; Daneault, Jean-Francois; Golabchi, Fatemeh N; Ferreira-Carvalho, Gabriela; Vergara-Diaz, Gloria; Sapienza, Stefano; Costante, Gianluca; Klucken, Jochen; Kautz, Thomas; Bonato, Paolo

2016-08-01

The development of wearable sensors has opened the door for long-term assessment of movement disorders. However, there is still a need for developing methods suitable to monitor motor symptoms in and outside the clinic. The purpose of this paper was to investigate deep learning as a method for this monitoring. Deep learning recently broke records in speech and image classification, but it has not been fully investigated as a potential approach to analyze wearable sensor data. We collected data from ten patients with idiopathic Parkinson's disease using inertial measurement units. Several motor tasks were expert-labeled and used for classification. We specifically focused on the detection of bradykinesia. For this, we compared standard machine learning pipelines with deep learning based on convolutional neural networks. Our results showed that deep learning outperformed other state-of-the-art machine learning algorithms by at least 4.6 % in terms of classification rate. We contribute a discussion of the advantages and disadvantages of deep learning for sensor-based movement assessment and conclude that deep learning is a promising method for this field.

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.

Deep Learning for Brain MRI Segmentation: State of the Art and Future Directions.

PubMed

Akkus, Zeynettin; Galimzianova, Alfiia; Hoogi, Assaf; Rubin, Daniel L; Erickson, Bradley J

2017-08-01

Quantitative analysis of brain MRI is routine for many neurological diseases and conditions and relies on accurate segmentation of structures of interest. Deep learning-based segmentation approaches for brain MRI are gaining interest due to their self-learning and generalization ability over large amounts of data. As the deep learning architectures are becoming more mature, they gradually outperform previous state-of-the-art classical machine learning algorithms. This review aims to provide an overview of current deep learning-based segmentation approaches for quantitative brain MRI. First we review the current deep learning architectures used for segmentation of anatomical brain structures and brain lesions. Next, the performance, speed, and properties of deep learning approaches are summarized and discussed. Finally, we provide a critical assessment of the current state and identify likely future developments and trends.

Learning and memory functions of the Basal Ganglia.

PubMed

Packard, Mark G; Knowlton, Barbara J

2002-01-01

Although the mammalian basal ganglia have long been implicated in motor behavior, it is generally recognized that the behavioral functions of this subcortical group of structures are not exclusively motoric in nature. Extensive evidence now indicates a role for the basal ganglia, in particular the dorsal striatum, in learning and memory. One prominent hypothesis is that this brain region mediates a form of learning in which stimulus-response (S-R) associations or habits are incrementally acquired. Support for this hypothesis is provided by numerous neurobehavioral studies in different mammalian species, including rats, monkeys, and humans. In rats and monkeys, localized brain lesion and pharmacological approaches have been used to examine the role of the basal ganglia in S-R learning. In humans, study of patients with neurodegenerative diseases that compromise the basal ganglia, as well as research using brain neuroimaging techniques, also provide evidence of a role for the basal ganglia in habit learning. Several of these studies have dissociated the role of the basal ganglia in S-R learning from those of a cognitive or declarative medial temporal lobe memory system that includes the hippocampus as a primary component. Evidence suggests that during learning, basal ganglia and medial temporal lobe memory systems are activated simultaneously and that in some learning situations competitive interference exists between these two systems.

Learning Sparse Feature Representations using Probabilistic Quadtrees and Deep Belief Nets

DTIC Science & Technology

2015-04-24

Feature Representations usingProbabilistic Quadtrees and Deep Belief Nets Learning sparse feature representations is a useful instru- ment for solving an...novel framework for the classifi cation of handwritten digits that learns sparse representations using probabilistic quadtrees and Deep Belief Nets... Learning Sparse Feature Representations usingProbabilistic Quadtrees and Deep Belief Nets Report Title Learning sparse feature representations is a useful

Longitudinal study of a cooperation-driven, socio-scientific issue intervention on promoting students' critical thinking and self-regulation in learning science

NASA Astrophysics Data System (ADS)

Wang, Hsin-Hui; Chen, Hsiang-Ting; Lin, Huann-shyang; Huang, Yu-Ning; Hong, Zuway-R.

2017-10-01

This longitudinal study explored the effects of a Cooperation-driven Socioscientific Issue (CDSSI) intervention on junior high school students' perceptions of critical thinking (CT) and self-regulation (SR) in Taiwan. Forty-nine grade 7 students were randomly selected as an experimental group (EG) to attend a 3-semester 72-hour intervention; while another 49 grade 7 students from the same school were randomly selected as the comparison group (CG). All participants completed a 4-wave student questionnaire to assess their perceptions of CT and SR. In addition, 8 target students from the EG with the lowest scores on either CT or SR were purposefully recruited for weekly observation. These target students and their teachers were interviewed one month after the intervention in each semester. Analyses of covariance and paired-wise t-tests revealed that the EG students' perceptions of CT and SR in learning science were improved during the study and were significantly better than their counterparts' at the end of the study. Systematic interview and classroom observation results were consistent with the quantitative findings. This study adds empirical evidence and provides insights into how CDSSI can be integrated into planning and implementing effective pedagogical strategies aimed at increasing students' perceptions of CT and SR in learning science.

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.

ShapeShop: Towards Understanding Deep Learning Representations via Interactive Experimentation.

PubMed

Hohman, Fred; Hodas, Nathan; Chau, Duen Horng

2017-05-01

Deep learning is the driving force behind many recent technologies; however, deep neural networks are often viewed as "black-boxes" due to their internal complexity that is hard to understand. Little research focuses on helping people explore and understand the relationship between a user's data and the learned representations in deep learning models. We present our ongoing work, ShapeShop, an interactive system for visualizing and understanding what semantics a neural network model has learned. Built using standard web technologies, ShapeShop allows users to experiment with and compare deep learning models to help explore the robustness of image classifiers.

Study of formation of deep trapping mechanism by UV, beta and gamma irradiated Eu(3+) activated SrY2O4 and Y4Al2O9 phosphors.

PubMed

Dubey, Vikas; Kaur, Jagjeet; Parganiha, Yogita; Suryanarayana, N S; Murthy, K V R

2016-04-01

This paper reports the thermoluminescence properties of Eu(3+) doped different host matrix phosphors (SrY2O4 and Y4Al2O9). The phosphor is prepared by high temperature solid state reaction method. The method is suitable for large scale production and fixed concentration of boric acid using as a flux. The prepared samples were characterized by X-ray diffraction technique and the crystallite size calculated by Scherer's formula. The prepared phosphor characterized by Scanning Electron Microscopic (SEM), Fourier Transform Infrared (FTIR), Energy Dispersive X-ray analysis (EDX), thermoluminescence (TL) and Transmission Electron Microscopic (TEM) techniques. The prepared phosphors for different concentration of Eu(3+) ions were examined by TL glow curve for UV, beta and gamma irradiation. The UV 254nm source used for UV irradiation, Sr(90) source was used for beta irradiation and Co(60) source used for gamma irradiation. SrY2O4:Eu(3+)and Y4Al2O9:Eu(3+) phosphors which shows both higher temperature peaks and lower temperature peaks for UV, beta and gamma irradiation. Here UV irradiated sample shows the formation of shallow trap (surface trapping) and the gamma irradiated sample shows the formation of deep trapping. The estimation of trap formation was evaluated by knowledge of trapping parameters. The trapping parameters such as activation energy, order of kinetics and frequency factor were calculated by peak shape method. Here most of the peak shows second order of kinetics. The effect of gamma, beta and UV exposure on TL studies was also examined and it shows linear response with dose which indicate that the samples may be useful for TL dosimetry. Formation of deep trapping mechanism by UV, beta and gamma irradiated Eu(3+) activated SrY2O4 and Y4Al2O9 phosphors is discussed in this paper. Copyright © 2015 Elsevier Ltd. All rights reserved.

Sr isotope variations in the Carnian-Norian succession at Pizzo Mondello, Sicani Mountains, Sicily

NASA Astrophysics Data System (ADS)

Onoue, T.; Yamashita, K.; Rigo, M.; Abate, B.

2017-12-01

The Norian stage in the Late Triassic is exceptionally long (23 Myr) and was subdivided into three substages: the Lacian, Alaunian, and Sevatian. In order to infer the Norian environmental changes in the western Tethys Ocean, the stratigraphic variations of 87Sr/86Sr in the Upper Triassic limestone succession in Sicily were examined. The Pizzo Mondello section studied here mainly consists of a pelagic carbonate sequence of the Scillato Formation, and ranges in age from Tuvalian (late Carnian) to Rhaetian. The Scillato Formation represents a deep-water pelagic facies deposited along the Sicanian Basin in the western Tethys Ocean. We selected fine-grained limestone samples from both the microfacies of lime-mudstone and wackestone to approximate the primary 87Sr/86Sr signature of the limestone beds. The 87Sr/86Sr values are relatively constant in the Tuvalian and Lacian (early Norian). However, the remarkable rise in 87Sr/86Sr occurred across the Lacian-Alaunian (early-middle Norian) transition. Variations in 87Sr/86Sr values show an increasing trend in 87Sr/86Sr from 0.7077 at the base of Lacian to 0.7080 in the Sevatian (late Norian). In the Sevatian, the 87Sr/86Sr ratios display a sudden negative excursion toward lower values and show a relatively quick recovery to pre-excursion 87Sr/86Sr ratios. Korte et al. (2003) suggested that the rise in the 87Sr/86Sr values from the middle Carnian to the late Norian coincide with the Cimmerian orogeny. Our new 87Sr/86Sr data from the Pizzo Mondello section reveal a comparable trend, with a sharp increase in 87Sr/86Sr within the Alaunian, suggesting the rapid uplift and erosion in the Cimmerian Mountains at this time. The cause of the 87Sr/86Sr excursion in the Sevatian remains uncertain. However, the biostratigraphic record of conodonts suggests that a morphological evolution towards platform-less elements occurred with the beginning of the Sr-isotope excursion.

A new Cu–cysteamine complex: structure and optical properties

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

Ma, Lun; Chen, Wei; Schatte, Gabriele

2014-06-07

Here we report the structure and optical properties of a new Cu–cysteamine complex (Cu–Cy) with a formula of Cu3Cl(SR)2 (R ¼ CH2CH2NH2). This Cu–Cy has a different structure from a previous Cu–Cy complex, in which both thio and amine groups from cysteamine bond with copper ions. Single-crystal X-ray diffraction and solid-state nuclear magnetic resonance results show that the oxidation state of copper in Cu3Cl(SR)2 is +1 rather than +2. Further, Cu3Cl(SR)2 has been observed to show intense photoluminescence and X-ray excited luminescence. More interesting is that Cu3Cl(SR)2 particles can produce singlet oxygen under irradiation by light or X-ray. This indicatesmore » that Cu3Cl(SR)2 is a new photosensitizer that can be used for deep cancer treatment as X-ray can penetrate soft tissues. All these findings mean that Cu3Cl(SR)2 is a new material with potential applications for lighting, radiation detection and cancer treatment.« less

Combined self-learning based single-image super-resolution and dual-tree complex wavelet transform denoising for medical images

NASA Astrophysics Data System (ADS)

Yang, Guang; Ye, Xujiong; Slabaugh, Greg; Keegan, Jennifer; Mohiaddin, Raad; Firmin, David

2016-03-01

In this paper, we propose a novel self-learning based single-image super-resolution (SR) method, which is coupled with dual-tree complex wavelet transform (DTCWT) based denoising to better recover high-resolution (HR) medical images. Unlike previous methods, this self-learning based SR approach enables us to reconstruct HR medical images from a single low-resolution (LR) image without extra training on HR image datasets in advance. The relationships between the given image and its scaled down versions are modeled using support vector regression with sparse coding and dictionary learning, without explicitly assuming reoccurrence or self-similarity across image scales. In addition, we perform DTCWT based denoising to initialize the HR images at each scale instead of simple bicubic interpolation. We evaluate our method on a variety of medical images. Both quantitative and qualitative results show that the proposed approach outperforms bicubic interpolation and state-of-the-art single-image SR methods while effectively removing noise.

An English Class with Emily.

ERIC Educational Resources Information Center

Bassett, Lawrence F.

1998-01-01

Presents a high school student's description in class of her deep connection to Hester Prynne in Nathaniel Hawthorne's "The Scarlet Letter," and how it offers a glimpse of the vast interior lives of women. (SR)

An adaptive deep Q-learning strategy for handwritten digit recognition.

PubMed

Qiao, Junfei; Wang, Gongming; Li, Wenjing; Chen, Min

2018-02-22

Handwritten digits recognition is a challenging problem in recent years. Although many deep learning-based classification algorithms are studied for handwritten digits recognition, the recognition accuracy and running time still need to be further improved. In this paper, an adaptive deep Q-learning strategy is proposed to improve accuracy and shorten running time for handwritten digit recognition. The adaptive deep Q-learning strategy combines the feature-extracting capability of deep learning and the decision-making of reinforcement learning to form an adaptive Q-learning deep belief network (Q-ADBN). First, Q-ADBN extracts the features of original images using an adaptive deep auto-encoder (ADAE), and the extracted features are considered as the current states of Q-learning algorithm. Second, Q-ADBN receives Q-function (reward signal) during recognition of the current states, and the final handwritten digits recognition is implemented by maximizing the Q-function using Q-learning algorithm. Finally, experimental results from the well-known MNIST dataset show that the proposed Q-ADBN has a superiority to other similar methods in terms of accuracy and running time. Copyright © 2018 Elsevier Ltd. All rights reserved.

Pb isotope compositions of modern deep sea turbidites

NASA Astrophysics Data System (ADS)

Hemming, S. R.; McLennan, S. M.

2001-01-01

Modern deep sea turbidite muds and sands collected from Lamont piston cores represent a large range in age of detrital sources as well as a spectrum of tectonic settings. Pb isotope compositions of all but three of the 66 samples lie to the right of the 4.56 Ga Geochron, and most also lie along a slope consistent with a time-integrated κ ( 232Th/ 238U) between 3.8 and 4.2. Modern deep sea turbidites show a predictable negative correlation between both Pb and Sr isotope ratios and ɛNd and ɛHf, clearly related to the age of continental sources. However, the consistency between Pb and Nd isotopes breaks down for samples with very old provenance ( ɛNd<-20) that are far less radiogenic than predicted by the negative correlation. The correlations among Sr, Nd and Hf isotopes also become more scattered in samples with very old provenance. The unradiogenic Pb isotopic character of modern sediments with Archean Nd model ages is consistent with a model where Th and U abundances of the Archean upper crust are significantly lower than the post-Archean upper crust.

Is the devil in the detail? Evidence for S-S learning after unconditional stimulus revaluation in human evaluative conditioning under a broader set of experimental conditions.

PubMed

Jensen-Fielding, Hannah; Luck, Camilla C; Lipp, Ottmar V

2017-11-28

Whether valence change during evaluative conditioning is mediated by a link between the conditional stimulus (CS) and the unconditional stimulus (US; S-S learning) or between the CS and the unconditional response (S-R learning) is a matter of continued debate. Changing the valence of the US after conditioning, known as US revaluation, can be used to dissociate these accounts. Changes in CS valence after US revaluation provide evidence for S-S learning but if CS valence does not change, evidence for S-R learning is found. Support for S-S learning has been provided by most past revaluation studies, but typically the CS and US have been from the same stimulus category, the task instructions have suggested that judgements of the CS should be based on the US, and USs have been mildly valenced stimuli. These factors may bias the results in favour of S-S learning. We examined whether S-R learning would be evident when CSs and USs were taken from different categories, the task instructions were removed, and more salient USs were used. US revaluation was found to influence explicit US evaluations and explicit and implicit CS evaluations, supporting an S-S learning account and suggesting that past results are stable across procedural changes.

ShapeShop: Towards Understanding Deep Learning Representations via Interactive Experimentation

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

Hohman, Frederick M.; Hodas, Nathan O.; Chau, Duen Horng

Deep learning is the driving force behind many recent technologies; however, deep neural networks are often viewed as “black-boxes” due to their internal complexity that is hard to understand. Little research focuses on helping people explore and understand the relationship between a user’s data and the learned representations in deep learning models. We present our ongoing work, ShapeShop, an interactive system for visualizing and understanding what semantics a neural network model has learned. Built using standard web technologies, ShapeShop allows users to experiment with and compare deep learning models to help explore the robustness of image classifiers.

An Energy-Efficient and Scalable Deep Learning/Inference Processor With Tetra-Parallel MIMD Architecture for Big Data Applications.

PubMed

Park, Seong-Wook; Park, Junyoung; Bong, Kyeongryeol; Shin, Dongjoo; Lee, Jinmook; Choi, Sungpill; Yoo, Hoi-Jun

2015-12-01

Deep Learning algorithm is widely used for various pattern recognition applications such as text recognition, object recognition and action recognition because of its best-in-class recognition accuracy compared to hand-crafted algorithm and shallow learning based algorithms. Long learning time caused by its complex structure, however, limits its usage only in high-cost servers or many-core GPU platforms so far. On the other hand, the demand on customized pattern recognition within personal devices will grow gradually as more deep learning applications will be developed. This paper presents a SoC implementation to enable deep learning applications to run with low cost platforms such as mobile or portable devices. Different from conventional works which have adopted massively-parallel architecture, this work adopts task-flexible architecture and exploits multiple parallelism to cover complex functions of convolutional deep belief network which is one of popular deep learning/inference algorithms. In this paper, we implement the most energy-efficient deep learning and inference processor for wearable system. The implemented 2.5 mm × 4.0 mm deep learning/inference processor is fabricated using 65 nm 8-metal CMOS technology for a battery-powered platform with real-time deep inference and deep learning operation. It consumes 185 mW average power, and 213.1 mW peak power at 200 MHz operating frequency and 1.2 V supply voltage. It achieves 411.3 GOPS peak performance and 1.93 TOPS/W energy efficiency, which is 2.07× higher than the state-of-the-art.

Model United Nations and Deep Learning: Theoretical and Professional Learning

ERIC Educational Resources Information Center

Engel, Susan; Pallas, Josh; Lambert, Sarah

2017-01-01

This article demonstrates that the purposeful subject design, incorporating a Model United Nations (MUN), facilitated deep learning and professional skills attainment in the field of International Relations. Deep learning was promoted in subject design by linking learning objectives to Anderson and Krathwohl's (2001) four levels of knowledge or…

Piezo-optic, photoelastic, and acousto-optic properties of SrB4O7 crystals.

PubMed

Mytsyk, Bohdan; Demyanyshyn, Natalia; Martynyuk-Lototska, Irina; Vlokh, Rostyslav

2011-07-20

On the basis of studies of the piezo-optic effect, it has been shown that SrB(4)O(7) crystals can be used as efficient acousto-optic materials in the vacuum ultraviolet spectral range. The full matrices of piezo-optic and photoelastic coefficients have been experimentally obtained for these crystals. The acousto-optic figure of merit and the diffraction efficiency have been estimated for both the visible and deep ultraviolet spectral ranges. © 2011 Optical Society of America

Deep learning in bioinformatics.

PubMed

Min, Seonwoo; Lee, Byunghan; Yoon, Sungroh

2017-09-01

In the era of big data, transformation of biomedical big data into valuable knowledge has been one of the most important challenges in bioinformatics. Deep learning has advanced rapidly since the early 2000s and now demonstrates state-of-the-art performance in various fields. Accordingly, application of deep learning in bioinformatics to gain insight from data has been emphasized in both academia and industry. Here, we review deep learning in bioinformatics, presenting examples of current research. To provide a useful and comprehensive perspective, we categorize research both by the bioinformatics domain (i.e. omics, biomedical imaging, biomedical signal processing) and deep learning architecture (i.e. deep neural networks, convolutional neural networks, recurrent neural networks, emergent architectures) and present brief descriptions of each study. Additionally, we discuss theoretical and practical issues of deep learning in bioinformatics and suggest future research directions. We believe that this review will provide valuable insights and serve as a starting point for researchers to apply deep learning approaches in their bioinformatics studies. © The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com.

Autonomous development and learning in artificial intelligence and robotics: Scaling up deep learning to human-like learning.

PubMed

Oudeyer, Pierre-Yves

2017-01-01

Autonomous lifelong development and learning are fundamental capabilities of humans, differentiating them from current deep learning systems. However, other branches of artificial intelligence have designed crucial ingredients towards autonomous learning: curiosity and intrinsic motivation, social learning and natural interaction with peers, and embodiment. These mechanisms guide exploration and autonomous choice of goals, and integrating them with deep learning opens stimulating perspectives.

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.

Deep learning methods for protein torsion angle prediction.

PubMed

Li, Haiou; Hou, Jie; Adhikari, Badri; Lyu, Qiang; Cheng, Jianlin

2017-09-18

Deep learning is one of the most powerful machine learning methods that has achieved the state-of-the-art performance in many domains. Since deep learning was introduced to the field of bioinformatics in 2012, it has achieved success in a number of areas such as protein residue-residue contact prediction, secondary structure prediction, and fold recognition. In this work, we developed deep learning methods to improve the prediction of torsion (dihedral) angles of proteins. We design four different deep learning architectures to predict protein torsion angles. The architectures including deep neural network (DNN) and deep restricted Boltzmann machine (DRBN), deep recurrent neural network (DRNN) and deep recurrent restricted Boltzmann machine (DReRBM) since the protein torsion angle prediction is a sequence related problem. In addition to existing protein features, two new features (predicted residue contact number and the error distribution of torsion angles extracted from sequence fragments) are used as input to each of the four deep learning architectures to predict phi and psi angles of protein backbone. The mean absolute error (MAE) of phi and psi angles predicted by DRNN, DReRBM, DRBM and DNN is about 20-21° and 29-30° on an independent dataset. The MAE of phi angle is comparable to the existing methods, but the MAE of psi angle is 29°, 2° lower than the existing methods. On the latest CASP12 targets, our methods also achieved the performance better than or comparable to a state-of-the art method. Our experiment demonstrates that deep learning is a valuable method for predicting protein torsion angles. The deep recurrent network architecture performs slightly better than deep feed-forward architecture, and the predicted residue contact number and the error distribution of torsion angles extracted from sequence fragments are useful features for improving prediction accuracy.

The deep layers of a Paleozoic arc: geochemistry of the Copley-Balaklala series, northern California

NASA Astrophysics Data System (ADS)

Brouxel, Marc; Lapierre, Henriette; Michard, Annie; Albarède, Francis

1987-10-01

REE, Zr, Nb concentrations and Sr, Nd isotope compositions have been measured in Copley basalts and andesites, Balaklala rhyolites, and Mule Mountain trondhjemites (northern California) which represent the deep layers of a well preserved intra-oceanic island arc of Siluro-Devonian age. 87Sr/ 86Sr is shifted towards high values (up to 0.707) whereas Ce is preferentially removed from rhyolites. A large proportion of the analyzed samples including some acidic rocks shows a pronounced depletion in light REE. The ɛ Nd(T) values of most Copley, Balaklala, and Mule Mountain rocks fall in the range +6 to +8 which suggests that they originated from a normal MORB-type source ( ɛ Nd(T) ≈ +9 ) contaminated with either sediments or an OIB-type component. In modern island arcs, only the shallow levels are accessible: comparison with the Copley-Balaklala-Mule Mountain Series suggests that, at depth, an immature island arc is likely to comprise thick layers of LILE-depleted tholeiites and rhyolites intensely altered by pervasive circulation of seawater. Least-square solutions of trace element models suggest that rhyolites and trondhjemites represent remelting of mafic volcanics from the arc basement rather than residual melts of basalt-andesite differentiation.

Postthrombotic Syndrome

MedlinePlus

... Rondina MT. Contemporary issues in the prevention and management of postthrombotic syndrome. Ann Pharmacother . 2009 ; 43 : 1824 –1835. OpenUrl CrossRef PubMed ↵ Kahn SR, Ginsberg JS. Relationship between deep venous thrombosis and the postthrombotic syndrome. ...

Towards deep learning with segregated dendrites

PubMed Central

Guerguiev, Jordan; Lillicrap, Timothy P

2017-01-01

Deep learning has led to significant advances in artificial intelligence, in part, by adopting strategies motivated by neurophysiology. However, it is unclear whether deep learning could occur in the real brain. Here, we show that a deep learning algorithm that utilizes multi-compartment neurons might help us to understand how the neocortex optimizes cost functions. Like neocortical pyramidal neurons, neurons in our model receive sensory information and higher-order feedback in electrotonically segregated compartments. Thanks to this segregation, neurons in different layers of the network can coordinate synaptic weight updates. As a result, the network learns to categorize images better than a single layer network. Furthermore, we show that our algorithm takes advantage of multilayer architectures to identify useful higher-order representations—the hallmark of deep learning. This work demonstrates that deep learning can be achieved using segregated dendritic compartments, which may help to explain the morphology of neocortical pyramidal neurons. PMID:29205151

Towards deep learning with segregated dendrites.

PubMed

Guerguiev, Jordan; Lillicrap, Timothy P; Richards, Blake A

2017-12-05

Deep learning has led to significant advances in artificial intelligence, in part, by adopting strategies motivated by neurophysiology. However, it is unclear whether deep learning could occur in the real brain. Here, we show that a deep learning algorithm that utilizes multi-compartment neurons might help us to understand how the neocortex optimizes cost functions. Like neocortical pyramidal neurons, neurons in our model receive sensory information and higher-order feedback in electrotonically segregated compartments. Thanks to this segregation, neurons in different layers of the network can coordinate synaptic weight updates. As a result, the network learns to categorize images better than a single layer network. Furthermore, we show that our algorithm takes advantage of multilayer architectures to identify useful higher-order representations-the hallmark of deep learning. This work demonstrates that deep learning can be achieved using segregated dendritic compartments, which may help to explain the morphology of neocortical pyramidal neurons.

Deep learning for neuroimaging: a validation study.

PubMed

Plis, Sergey M; Hjelm, Devon R; Salakhutdinov, Ruslan; Allen, Elena A; Bockholt, Henry J; Long, Jeffrey D; Johnson, Hans J; Paulsen, Jane S; Turner, Jessica A; Calhoun, Vince D

2014-01-01

Deep learning methods have recently made notable advances in the tasks of classification and representation learning. These tasks are important for brain imaging and neuroscience discovery, making the methods attractive for porting to a neuroimager's toolbox. Success of these methods is, in part, explained by the flexibility of deep learning models. However, this flexibility makes the process of porting to new areas a difficult parameter optimization problem. In this work we demonstrate our results (and feasible parameter ranges) in application of deep learning methods to structural and functional brain imaging data. These methods include deep belief networks and their building block the restricted Boltzmann machine. We also describe a novel constraint-based approach to visualizing high dimensional data. We use it to analyze the effect of parameter choices on data transformations. Our results show that deep learning methods are able to learn physiologically important representations and detect latent relations in neuroimaging data.

Deep Learning and Developmental Learning: Emergence of Fine-to-Coarse Conceptual Categories at Layers of Deep Belief Network.

PubMed

Sadeghi, Zahra

2016-09-01

In this paper, I investigate conceptual categories derived from developmental processing in a deep neural network. The similarity matrices of deep representation at each layer of neural network are computed and compared with their raw representation. While the clusters generated by raw representation stand at the basic level of abstraction, conceptual categories obtained from deep representation shows a bottom-up transition procedure. Results demonstrate a developmental course of learning from specific to general level of abstraction through learned layers of representations in a deep belief network. © The Author(s) 2016.

Sr isotope zoning in plagioclase from andesites at Cabo De Gata, Spain: Evidence for shallow and deep contamination

NASA Astrophysics Data System (ADS)

Waight, Tod E.; Tørnqvist, Jakob B.

2018-05-01

Plagioclase crystals in andesites from the Cabo De Gata region show generally radiogenic Sr isotope compositions and consistent core to rim increases in 87Sr/86Sr that are indicative of open system processes in the lithosphere and crustal contamination during crystallization. High-grade metamorphic rocks of the Alpujárride and Nevado-Filábride complexes represent the most likely crustal contaminants. The plagioclases are characterized by subtly zoned and resorbed calcic cores (An73-86). These cores also have radiogenic 87Sr/86Sr (0.7127-0.7129), although typically less radiogenic than plagioclase rims, groundmass plagioclase and whole rock compositions (up to 87Sr/86Sr = 0.7135). These cores are interpreted to represent early crystallization of plagioclase from hydrous melts emplaced into the lower crust. The parental melts to these andesites must therefore have already inherited their radiogenic Sr isotope compositions prior to entering the lower crust and before the onset of crystallization of plagioclase, which is inconsistent with previous models suggesting that the generally radiogenic nature of Sr in these volcanics reflects large amounts of crustal contamination. Instead, the isotope systematics are consistent with models invoked significant addition of a subducted sediment component to the mantle source. The high-An% plagioclase cores are characterized by resorption textures, which are consistent with dissolution during rapid decompression and/or devolatisation during magma migration from the lower crust into upper crustal magma chambers.

The Effects of Discipline on Deep Approaches to Student Learning and College Outcomes

ERIC Educational Resources Information Center

Nelson Laird, Thomas F.; Shoup, Rick; Kuh, George D.; Schwarz, Michael J.

2008-01-01

"Deep learning" represents student engagement in approaches to learning that emphasize integration, synthesis, and reflection. Because learning is a shared responsibility between students and faculty, it is important to determine whether faculty members emphasize deep approaches to learning and to assess how much students employ these approaches.…

SR&DB Cryogenic Research & Development for Space Applications

NASA Astrophysics Data System (ADS)

Bondarenko, S. I.; Arkhipov, V. T.; Logvinenko, S. P.; Solodovnik, L. L.; Rusanov, K. V.; Shcherbakova, N. S.

The Special Research and Development Bureau (SR&DB) for Cryogenic Technology of the B. Verkin Institute for Low Temperature Physics & Engineering was founded in 1971 and is located in Kharkov, Ukraine. Its primary focus has been in the area of applied r&d in the field of cryogenic technology for space applications. Within this field SR&DB has had many successful accomplishments, especially in the development of satellite based cryogenic cooling systems, mass spectrometer measurement devices, resistence thermometers, and cryogenically cooled optical systems. We have developed very advanced technology in the fields of fluids, heat transfer and hydrodynamics under micro-gravity conditions. Many of the SR&DB cryogenic products have been successfully implemented for former Soviet space applications, both near-earth and deep space. The SR&DB unique experience in many R&D areas can be and are being used for a new generation of space applications which have a requirement for planetary and deep-space missions. Systems we have developed have been proven to have a 5-year life in orbit. Recently we have focused much of our attention, as well, to the requirement low-weight and low-power systems which are mandatory requirements for outerspace missions. The funtionality of the exterior surfaces of a spacecraft are mainly dependent on the composition of its internally generated local atmosphere. In order to continually assess the content and concentration of components of this atmosphere we have developed space based mass spectrometric measuring devices. Devices which require such continual measurement are optical devices, emission receivers, solar cells, etc. A significant technology advance in the field of cryogenics is the application of cryoagents in systems of life support and spacecraft engine operation. We have studied and have an in-depth comprehension of unique phase-transition for these cryoagents such as oxygen, hydrogen, et al. under microgravity conditions. Currently SR&DB under contract to the National Space Agency of Ukraine has been developing an experimental apparatus for studying the continuous boiling off of cryogenic fluids under micro-gravity conditions.

The site occupation and valence of Mn ions in the crystal lattice of Sr{sub 4}Al{sub 14}O{sub 25} and its deep red emission for high color-rendering white light-emitting diodes

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

Chen, Lei, E-mail: shanggan2009@qq.com; Xue, Shaochan; Chen, Xiuling

2014-12-15

Highlights: • Different valences of Mn ions in Sr{sub 4}Al{sub 14}O{sub 25} were identified using XANES and EPR. • Red luminescence was attributed to Mn{sup 4+} occupying the center of AlO{sub 6} octahedron. • The Mn{sup 3+} incorporated in the center of AlO{sub 4} tetrahedron was non-luminescent. • The bond-valence theory was used to analyze the effective valences of cations. • A white LED device with CRI up to Ra 93.23 was packaged by using the red phosphor. - Abstract: The synthesis and component of red phosphor, Sr{sub 4}Al{sub 14}O{sub 25}: Mn, were optimized for application in white light-emitting diodes.more » The microstructure and morphology were investigated by the X-ray diffraction and scanning electron microscopy. Different valences of Mn ions in Sr{sub 4}Al{sub 14}O{sub 25} were discriminated using the electron paramagnetic resonance and X-ray absorption near-edge structure spectroscopy techniques. The bond-valence theory was used to analyze the effective valences of Sr{sup 2+} and Al{sup 3+} in Sr{sub 4}Al{sub 14}O{sub 25}. As a result, the strong covalence of Al{sup 3+} in the AlO{sub 4} tetrahedron other than in the AlO{sub 6} octahedron is disclosed. The deep red emission is attributed to Mn{sup 4+} occupying the center of AlO{sub 6} octahedron. The mechanism of energy transfer is mainly through dipole–dipole interaction, revealed by the analyses of critical distance and concentration quench. A high color rendering white LED prototype with color-rendering index up to Ra 93.23 packaged by using the red phosphor demonstrates its applicability.« less

Deep-Elaborative Learning of Introductory Management Accounting for Business Students

ERIC Educational Resources Information Center

Choo, Freddie; Tan, Kim B.

2005-01-01

Research by Choo and Tan (1990; 1995) suggests that accounting students, who engage in deep-elaborative learning, have a better understanding of the course materials. The purposes of this paper are: (1) to describe a deep-elaborative instructional approach (hereafter DEIA) that promotes deep-elaborative learning of introductory management…

Ramifications of codoping SrI2:Eu with isovalent and aliovalent impurities

NASA Astrophysics Data System (ADS)

Feng, Qingguo; Biswas, Koushik

2016-12-01

Eu2+ doped SrI2 is an important scintillator having applications in the field of radiation detection. Codoping techniques are often useful to improve the electronic response of such insulators. Using first-principles based approach, we report on the properties of SrI2:Eu and the influence of codoping with aliovalent (Na, Cs) and isovalent (Mg, Ca, Ba, and Sn) impurities. These codopants do not preferably bind with Eu and are expected to remain as isolated impurities in the SrI2 host. As isolated defects they display amphoteric behavior having, in most cases, significant ionization energies of the donor and acceptor levels. Furthermore, the acceptor states of Na, Cs, and Mg can bind with I-vacancy forming charge compensated donor-acceptor pairs. Such pairs may also bind additional holes or electrons similar to the isolated defects. Lack of deep-to-shallow behavior upon codoping and its ramifications will be discussed.

Hello World Deep Learning in Medical Imaging.

PubMed

Lakhani, Paras; Gray, Daniel L; Pett, Carl R; Nagy, Paul; Shih, George

2018-05-03

There is recent popularity in applying machine learning to medical imaging, notably deep learning, which has achieved state-of-the-art performance in image analysis and processing. The rapid adoption of deep learning may be attributed to the availability of machine learning frameworks and libraries to simplify their use. In this tutorial, we provide a high-level overview of how to build a deep neural network for medical image classification, and provide code that can help those new to the field begin their informatics projects.

Elemental and isotopic (C, O, Sr, Nd) compositions of Late Paleozoic carbonated eclogite and marble from the SW Tianshan UHP belt, NW China: Implications for deep carbon cycle

NASA Astrophysics Data System (ADS)

Zhu, Jianjiang; Zhang, Lifei; Lü, Zeng; Bader, Thomas

2018-03-01

Subduction zones are important for understanding of the global carbon cycle from the surface to deep part of the mantle. The processes involved the metamorphism of carbonate-bearing rocks largely control the fate of carbon and contribute to local carbon isotopic heterogeneities of the mantle. In this study, we present petrological and geochemical results for marbles and carbonated eclogites in the Southwestern Tianshan UHP belt, NW China. Marbles are interlayered with coesite-bearing pelitic schists, and have Sr-Nd isotopic values (εNd (T=320Ma) = -3.7 to -8.9, 87Sr/86Sr (i) = 0.7084-0.7089), typical of marine carbonates. The marbles have dispersed low δ18OVSMOW values (ranging from 14 to 29‰) and unaffected carbon isotope (δ13CVPDB = -0.2-3.6‰), possibly due to infiltration of external H2O-rich fluids. Recycling of these marbles into mantle may play a key role in the carbon budget and contributed to the mantle carbon isotope heterogeneity. The carbonated eclogites have high Sr isotopic compositions (87Sr/86Sr (i) = 0.7077-0.7082) and positive εNd (T = 320 Ma) values (from 7.6 to 8.2), indicative of strong seafloor alteration of their protolith. The carbonates in the carbonated eclogites are mainly dolomite (Fe# = 12-43, Fe# = Fe2+/(Fe2+ + Mg)) that were added into oceanic basalts during seafloor alteration and experienced calcite - dolomite - magnesite transformation during the subduction metamorphic process. The uniformly low δ18O values (∼11.44‰) of carbonates in the carbontaed eclogites can be explained by closed-system equilibrium between carbonate and silicate minerals. The low δ13C values (from -3.3 to -7.7‰) of the carbonated eclogites most likely reflect contribution from organic carbon. Recycling of these carbonated eclogites with C isotope similar to typical mantle reservoirs into mantle may have little effect on the mantle carbon isotope heterogeneity.

Deep learning predictions of survival based on MRI in amyotrophic lateral sclerosis.

PubMed

van der Burgh, Hannelore K; Schmidt, Ruben; Westeneng, Henk-Jan; de Reus, Marcel A; van den Berg, Leonard H; van den Heuvel, Martijn P

2017-01-01

Amyotrophic lateral sclerosis (ALS) is a progressive neuromuscular disease, with large variation in survival between patients. Currently, it remains rather difficult to predict survival based on clinical parameters alone. Here, we set out to use clinical characteristics in combination with MRI data to predict survival of ALS patients using deep learning, a machine learning technique highly effective in a broad range of big-data analyses. A group of 135 ALS patients was included from whom high-resolution diffusion-weighted and T1-weighted images were acquired at the first visit to the outpatient clinic. Next, each of the patients was monitored carefully and survival time to death was recorded. Patients were labeled as short, medium or long survivors, based on their recorded time to death as measured from the time of disease onset. In the deep learning procedure, the total group of 135 patients was split into a training set for deep learning (n = 83 patients), a validation set (n = 20) and an independent evaluation set (n = 32) to evaluate the performance of the obtained deep learning networks. Deep learning based on clinical characteristics predicted survival category correctly in 68.8% of the cases. Deep learning based on MRI predicted 62.5% correctly using structural connectivity and 62.5% using brain morphology data. Notably, when we combined the three sources of information, deep learning prediction accuracy increased to 84.4%. Taken together, our findings show the added value of MRI with respect to predicting survival in ALS, demonstrating the advantage of deep learning in disease prognostication.

Integrative and Deep Learning through a Learning Community: A Process View of Self

ERIC Educational Resources Information Center

Mahoney, Sandra; Schamber, Jon

2011-01-01

This study investigated deep learning produced in a community of general education courses. Student speeches on liberal education were analyzed for discovering a grounded theory of ideas about self. The study found that learning communities cultivate deep, integrative learning that makes the value of a liberal education relevant to students.…

Problem-Based Learning to Foster Deep Learning in Preservice Geography Teacher Education

ERIC Educational Resources Information Center

Golightly, Aubrey; Raath, Schalk

2015-01-01

In South Africa, geography education students' approach to deep learning has received little attention. Therefore the purpose of this one-shot experimental case study was to evaluate the extent to which first-year geography education students used deep or surface learning in an embedded problem-based learning (PBL) format. The researchers measured…

Teaching Decoding Strategies without Destroying Story.

ERIC Educational Resources Information Center

Kane, Sharon

1999-01-01

Argues that deep coding skills must and can be introduced, taught, practiced, and reinforced within contexts meaningful to students. Shows how teachers can provide these meaningful educational contexts within which decoding strategies make sense to emerging readers. (SR)

Pulmonary emphysema classification based on an improved texton learning model by sparse representation

NASA Astrophysics Data System (ADS)

Zhang, Min; Zhou, Xiangrong; Goshima, Satoshi; Chen, Huayue; Muramatsu, Chisako; Hara, Takeshi; Yokoyama, Ryujiro; Kanematsu, Masayuki; Fujita, Hiroshi

2013-03-01

In this paper, we present a texture classification method based on texton learned via sparse representation (SR) with new feature histogram maps in the classification of emphysema. First, an overcomplete dictionary of textons is learned via KSVD learning on every class image patches in the training dataset. In this stage, high-pass filter is introduced to exclude patches in smooth area to speed up the dictionary learning process. Second, 3D joint-SR coefficients and intensity histograms of the test images are used for characterizing regions of interest (ROIs) instead of conventional feature histograms constructed from SR coefficients of the test images over the dictionary. Classification is then performed using a classifier with distance as a histogram dissimilarity measure. Four hundreds and seventy annotated ROIs extracted from 14 test subjects, including 6 paraseptal emphysema (PSE) subjects, 5 centrilobular emphysema (CLE) subjects and 3 panlobular emphysema (PLE) subjects, are used to evaluate the effectiveness and robustness of the proposed method. The proposed method is tested on 167 PSE, 240 CLE and 63 PLE ROIs consisting of mild, moderate and severe pulmonary emphysema. The accuracy of the proposed system is around 74%, 88% and 89% for PSE, CLE and PLE, respectively.

New Rb-Sr mineral ages temporally link plume events with accretion at the margin of Gondwana

USGS Publications Warehouse

Flowerdew, M.J.; Daly, J.S.; Riley, T.R.

2007-01-01

Five of six Rb-Sr muscovite mineral isochron ages from the Scotia Metamorphic Complex of the South Orkney Islands, West Antarctica, average 190 ± 4 Ma. The muscovite ages are interpreted to date foliation-formation and thus also accretion and subduction at the Gondwana margin. Coincident picrite and ferropicrite magmatism, indicative of melts from deep-seated depleted mantle, permits a causative link between accretion and the arrival of the Karoo – Ferrar – Chon Aike mantle plume in the Early Jurassic. Three biotite Rb-Sr mineral isochron ages are consistently younger and average 176 ± 5 Ma. The biotite ages may record post-metamorphic cooling or more likely retrogressive metamorphic effects during uplift.

A deep hydrothermal fault zone in the lower oceanic crust, Samail ophiolite Oman

NASA Astrophysics Data System (ADS)

Zihlmann, B.; Mueller, S.; Koepke, J.; Teagle, D. A. H.

2017-12-01

Hydrothermal circulation is a key process for the exchange of chemical elements between the oceans and the solid Earth and for the extraction of heat from newly accreted crust at mid-ocean ridges. However, due to a dearth of samples from intact oceanic crust, or continuous samples from ophiolites, there remain major short comings in our understanding of hydrothermal circulation in the oceanic crust, especially in the deeper parts. In particular, it is unknown whether fluid recharge and discharge occurs pervasively or if it is mainly channeled within discrete zones such as faults. Here, we present a description of a hydrothermal fault zone that crops out in Wadi Gideah in the layered gabbro section of the Samail ophiolite of Oman. Field observations reveal a one meter thick chlorite-epidote normal fault with disseminated pyrite and chalcopyrite and heavily altered gabbro clasts at its core. In both, the hanging and the footwall the gabbro is altered and abundantly veined with amphibole, epidote, prehnite and zeolite. Whole rock mass balance calculations show enrichments in Fe, Mn, Sc, V, Co, Cu, Rb, Zr, Nb, Th and U and depletions of Si, Ca, Na, Cr, Zn, Sr, Ba and Pb concentrations in the fault rock compared to fresh layered gabbros. Gabbro clasts within the fault zone as well as altered rock from the hanging wall show enrichments in Na, Sc, V, Co, Rb, Zr, Nb and depletion of Cr, Ni, Cu, Zn, Sr and Pb. Strontium isotope whole rock data of the fault rock yield 87Sr/86Sr ratios of 0.7046, which is considerably more radiogenic than fresh layered gabbro from this locality (87Sr/86Sr = 0.7030 - 0.7034), and similar to black smoker hydrothermal signatures based on epidote, measured elsewhere in the ophiolite. Altered gabbro clasts within the fault zone show similar values with 87Sr/86Sr ratios of 0.7045 - 0.7050, whereas hanging wall and foot wall display values only slightly more radiogenic than fresh layered gabbro.The secondary mineral assemblages and strontium isotope compositions of the fault rock, clasts and hanging wall indicate interaction with a seawater-derived hydrothermal fluid during oceanic spreading at an ancient mid-ocean ridge. The considerable elemental mass changes in the fault rocks and surrounds compared to the primary layered gabbros suggests extensive hydrothermal fluid flow and exchange deep within the ocean crust.

Deep Direct Reinforcement Learning for Financial Signal Representation and Trading.

PubMed

Deng, Yue; Bao, Feng; Kong, Youyong; Ren, Zhiquan; Dai, Qionghai

2017-03-01

Can we train the computer to beat experienced traders for financial assert trading? In this paper, we try to address this challenge by introducing a recurrent deep neural network (NN) for real-time financial signal representation and trading. Our model is inspired by two biological-related learning concepts of deep learning (DL) and reinforcement learning (RL). In the framework, the DL part automatically senses the dynamic market condition for informative feature learning. Then, the RL module interacts with deep representations and makes trading decisions to accumulate the ultimate rewards in an unknown environment. The learning system is implemented in a complex NN that exhibits both the deep and recurrent structures. Hence, we propose a task-aware backpropagation through time method to cope with the gradient vanishing issue in deep training. The robustness of the neural system is verified on both the stock and the commodity future markets under broad testing conditions.

Deep Logic Networks: Inserting and Extracting Knowledge From Deep Belief Networks.

PubMed

Tran, Son N; d'Avila Garcez, Artur S

2018-02-01

Developments in deep learning have seen the use of layerwise unsupervised learning combined with supervised learning for fine-tuning. With this layerwise approach, a deep network can be seen as a more modular system that lends itself well to learning representations. In this paper, we investigate whether such modularity can be useful to the insertion of background knowledge into deep networks, whether it can improve learning performance when it is available, and to the extraction of knowledge from trained deep networks, and whether it can offer a better understanding of the representations learned by such networks. To this end, we use a simple symbolic language-a set of logical rules that we call confidence rules-and show that it is suitable for the representation of quantitative reasoning in deep networks. We show by knowledge extraction that confidence rules can offer a low-cost representation for layerwise networks (or restricted Boltzmann machines). We also show that layerwise extraction can produce an improvement in the accuracy of deep belief networks. Furthermore, the proposed symbolic characterization of deep networks provides a novel method for the insertion of prior knowledge and training of deep networks. With the use of this method, a deep neural-symbolic system is proposed and evaluated, with the experimental results indicating that modularity through the use of confidence rules and knowledge insertion can be beneficial to network performance.

Deep Learning for Computer Vision: A Brief Review

PubMed Central

Doulamis, Nikolaos; Doulamis, Anastasios; Protopapadakis, Eftychios

2018-01-01

Over the last years deep learning methods have been shown to outperform previous state-of-the-art machine learning techniques in several fields, with computer vision being one of the most prominent cases. This review paper provides a brief overview of some of the most significant deep learning schemes used in computer vision problems, that is, Convolutional Neural Networks, Deep Boltzmann Machines and Deep Belief Networks, and Stacked Denoising Autoencoders. A brief account of their history, structure, advantages, and limitations is given, followed by a description of their applications in various computer vision tasks, such as object detection, face recognition, action and activity recognition, and human pose estimation. Finally, a brief overview is given of future directions in designing deep learning schemes for computer vision problems and the challenges involved therein. PMID:29487619

Deep Belief Networks for Electroencephalography: A Review of Recent Contributions and Future Outlooks.

PubMed

Movahedi, Faezeh; Coyle, James L; Sejdic, Ervin

2018-05-01

Deep learning, a relatively new branch of machine learning, has been investigated for use in a variety of biomedical applications. Deep learning algorithms have been used to analyze different physiological signals and gain a better understanding of human physiology for automated diagnosis of abnormal conditions. In this paper, we provide an overview of deep learning approaches with a focus on deep belief networks in electroencephalography applications. We investigate the state-of-the-art algorithms for deep belief networks and then cover the application of these algorithms and their performances in electroencephalographic applications. We covered various applications of electroencephalography in medicine, including emotion recognition, sleep stage classification, and seizure detection, in order to understand how deep learning algorithms could be modified to better suit the tasks desired. This review is intended to provide researchers with a broad overview of the currently existing deep belief network methodology for electroencephalography signals, as well as to highlight potential challenges for future research.

Incorporating deep learning with convolutional neural networks and position specific scoring matrices for identifying electron transport proteins.

PubMed

Le, Nguyen-Quoc-Khanh; Ho, Quang-Thai; Ou, Yu-Yen

2017-09-05

In several years, deep learning is a modern machine learning technique using in a variety of fields with state-of-the-art performance. Therefore, utilization of deep learning to enhance performance is also an important solution for current bioinformatics field. In this study, we try to use deep learning via convolutional neural networks and position specific scoring matrices to identify electron transport proteins, which is an important molecular function in transmembrane proteins. Our deep learning method can approach a precise model for identifying of electron transport proteins with achieved sensitivity of 80.3%, specificity of 94.4%, and accuracy of 92.3%, with MCC of 0.71 for independent dataset. The proposed technique can serve as a powerful tool for identifying electron transport proteins and can help biologists understand the function of the electron transport proteins. Moreover, this study provides a basis for further research that can enrich a field of applying deep learning in bioinformatics. © 2017 Wiley Periodicals, Inc. © 2017 Wiley Periodicals, Inc.

Ion microprobe measurement of strontium isotopes in calcium carbonate with application to salmon otoliths

USGS Publications Warehouse

Weber, P.K.; Bacon, C.R.; Hutcheon, I.D.; Ingram, B.L.; Wooden, J.L.

2005-01-01

The ion microprobe has the capability to generate high resolution, high precision isotopic measurements, but analysis of the isotopic composition of strontium, as measured by the 87Sr/ 86Sr ratio, has been hindered by isobaric interferences. Here we report the first high precision measurements of 87Sr/ 86Sr by ion microprobe in calcium carbonate samples with moderate Sr concentrations. We use the high mass resolving power (7000 to 9000 M.R.P.) of the SHRIMP-RG ion microprobe in combination with its high transmission to reduce the number of interfering species while maintaining sufficiently high count rates for precise isotopic measurements. The isobaric interferences are characterized by peak modeling and repeated analyses of standards. We demonstrate that by sample-standard bracketing, 87Sr/86Sr ratios can be measured in inorganic and biogenic carbonates with Sr concentrations between 400 and 1500 ppm with ???2??? external precision (2??) for a single analysis, and subpermil external precision with repeated analyses. Explicit correction for isobaric interferences (peak-stripping) is found to be less accurate and precise than sample-standard bracketing. Spatial resolution is ???25 ??m laterally and 2 ??m deep for a single analysis, consuming on the order of 2 ng of material. The method is tested on otoliths from salmon to demonstrate its accuracy and utility. In these growth-banded aragonitic structures, one-week temporal resolution can be achieved. The analytical method should be applicable to other calcium carbonate samples with similar Sr concentrations. Copyright ?? 2005 Elsevier Ltd.

Deep Learning as an Individual, Conditional, and Contextual Influence on First-Year Student Outcomes

ERIC Educational Resources Information Center

Reason, Robert D.; Cox, Bradley E.; McIntosh, Kadian; Terenzini, Patrick T.

2010-01-01

For years, educators have drawn a distinction between deep cognitive processing and surface-level cognitive processing, with the former resulting in greater learning. In recent years, researchers at NSSE have created DEEP Learning scales, which consist of items related to students' experiences which are believed to encourage deep processing. In…

Enhanced Experience Replay for Deep Reinforcement Learning

DTIC Science & Technology

2015-11-01

ARL-TR-7538 ● NOV 2015 US Army Research Laboratory Enhanced Experience Replay for Deep Reinforcement Learning by David Doria...Experience Replay for Deep Reinforcement Learning by David Doria, Bryan Dawson, and Manuel Vindiola Computational and Information Sciences Directorate...

Deep learning of unsteady laminar flow over a cylinder

NASA Astrophysics Data System (ADS)

Lee, Sangseung; You, Donghyun

2017-11-01

Unsteady flow over a circular cylinder is reconstructed using deep learning with a particular emphasis on elucidating the potential of learning the solution of the Navier-Stokes equations. A deep neural network (DNN) is employed for deep learning, while numerical simulations are conducted to produce training database. Instantaneous and mean flow fields which are reconstructed by deep learning are compared with the simulation results. Fourier transform of flow variables has been conducted to validate the ability of DNN to capture both amplitudes and frequencies of flow motions. Basis decomposition of learned flow is performed to understand the underlying mechanisms of learning flow through DNN. The present study suggests that a deep learning technique can be utilized for reconstruction and, potentially, for prediction of fluid flow instead of solving the Navier-Stokes equations. This work was supported by the National Research Foundation of Korea(NRF) Grant funded by the Korea government(Ministry of Science, ICT and Future Planning) (No. 2014R1A2A1A11049599, No. 2015R1A2A1A15056086, No. 2016R1E1A2A01939553).

Negative mood reverses devaluation of goal-directed drug-seeking favouring an incentive learning account of drug dependence.

PubMed

Hogarth, Lee; He, Zhimin; Chase, Henry W; Wills, Andy J; Troisi, Joseph; Leventhal, Adam M; Mathew, Amanda R; Hitsman, Brian

2015-09-01

Two theories explain how negative mood primes smoking behaviour. The stimulus-response (S-R) account argues that in the negative mood state, smoking is experienced as more reinforcing, establishing a direct (automatic) association between the negative mood state and smoking behaviour. By contrast, the incentive learning account argues that in the negative mood state smoking is expected to be more reinforcing, which integrates with instrumental knowledge of the response required to produce that outcome. One differential prediction is that whereas the incentive learning account anticipates that negative mood induction could augment a novel tobacco-seeking response in an extinction test, the S-R account could not explain this effect because the extinction test prevents S-R learning by omitting experience of the reinforcer. To test this, overnight-deprived daily smokers (n = 44) acquired two instrumental responses for tobacco and chocolate points, respectively, before smoking to satiety. Half then received negative mood induction to raise the expected value of tobacco, opposing satiety, whilst the remainder received positive mood induction. Finally, a choice between tobacco and chocolate was measured in extinction to test whether negative mood could augment tobacco choice, opposing satiety, in the absence of direct experience of tobacco reinforcement. Negative mood induction not only abolished the devaluation of tobacco choice, but participants with a significant increase in negative mood increased their tobacco choice in extinction, despite satiety. These findings suggest that negative mood augments drug-seeking by raising the expected value of the drug through incentive learning, rather than through automatic S-R control.

Boosting compound-protein interaction prediction by deep learning.

PubMed

Tian, Kai; Shao, Mingyu; Wang, Yang; Guan, Jihong; Zhou, Shuigeng

2016-11-01

The identification of interactions between compounds and proteins plays an important role in network pharmacology and drug discovery. However, experimentally identifying compound-protein interactions (CPIs) is generally expensive and time-consuming, computational approaches are thus introduced. Among these, machine-learning based methods have achieved a considerable success. However, due to the nonlinear and imbalanced nature of biological data, many machine learning approaches have their own limitations. Recently, deep learning techniques show advantages over many state-of-the-art machine learning methods in some applications. In this study, we aim at improving the performance of CPI prediction based on deep learning, and propose a method called DL-CPI (the abbreviation of Deep Learning for Compound-Protein Interactions prediction), which employs deep neural network (DNN) to effectively learn the representations of compound-protein pairs. Extensive experiments show that DL-CPI can learn useful features of compound-protein pairs by a layerwise abstraction, and thus achieves better prediction performance than existing methods on both balanced and imbalanced datasets. Copyright © 2016 Elsevier Inc. All rights reserved.

Self-regulation: from goal orientation to job performance.

PubMed

Porath, Christine L; Bateman, Thomas S

2006-01-01

The authors investigated the effects on job performance of 3 forms of goal orientation and 4 self-regulation (SR) tactics. In a longitudinal field study with salespeople, learning and performance-prove goal orientation predicted subsequent sales performance, whereas performance-avoid goal orientation negatively predicted sales performance. The SR tactics functioned as mediating variables between learning and performance-prove goal orientations and performance. Social competence and proactive behavior directly and positively predicted sales performance, and emotional control negatively predicted performance. (c) 2006 APA, all rights reserved.

Performance evaluation of 2D and 3D deep learning approaches for automatic segmentation of multiple organs on CT images

NASA Astrophysics Data System (ADS)

Zhou, Xiangrong; Yamada, Kazuma; Kojima, Takuya; Takayama, Ryosuke; Wang, Song; Zhou, Xinxin; Hara, Takeshi; Fujita, Hiroshi

2018-02-01

The purpose of this study is to evaluate and compare the performance of modern deep learning techniques for automatically recognizing and segmenting multiple organ regions on 3D CT images. CT image segmentation is one of the important task in medical image analysis and is still very challenging. Deep learning approaches have demonstrated the capability of scene recognition and semantic segmentation on nature images and have been used to address segmentation problems of medical images. Although several works showed promising results of CT image segmentation by using deep learning approaches, there is no comprehensive evaluation of segmentation performance of the deep learning on segmenting multiple organs on different portions of CT scans. In this paper, we evaluated and compared the segmentation performance of two different deep learning approaches that used 2D- and 3D deep convolutional neural networks (CNN) without- and with a pre-processing step. A conventional approach that presents the state-of-the-art performance of CT image segmentation without deep learning was also used for comparison. A dataset that includes 240 CT images scanned on different portions of human bodies was used for performance evaluation. The maximum number of 17 types of organ regions in each CT scan were segmented automatically and compared to the human annotations by using ratio of intersection over union (IU) as the criterion. The experimental results demonstrated the IUs of the segmentation results had a mean value of 79% and 67% by averaging 17 types of organs that segmented by a 3D- and 2D deep CNN, respectively. All the results of the deep learning approaches showed a better accuracy and robustness than the conventional segmentation method that used probabilistic atlas and graph-cut methods. The effectiveness and the usefulness of deep learning approaches were demonstrated for solving multiple organs segmentation problem on 3D CT images.

Deep Learning: A Primer for Radiologists.

PubMed

Chartrand, Gabriel; Cheng, Phillip M; Vorontsov, Eugene; Drozdzal, Michal; Turcotte, Simon; Pal, Christopher J; Kadoury, Samuel; Tang, An

2017-01-01

Deep learning is a class of machine learning methods that are gaining success and attracting interest in many domains, including computer vision, speech recognition, natural language processing, and playing games. Deep learning methods produce a mapping from raw inputs to desired outputs (eg, image classes). Unlike traditional machine learning methods, which require hand-engineered feature extraction from inputs, deep learning methods learn these features directly from data. With the advent of large datasets and increased computing power, these methods can produce models with exceptional performance. These models are multilayer artificial neural networks, loosely inspired by biologic neural systems. Weighted connections between nodes (neurons) in the network are iteratively adjusted based on example pairs of inputs and target outputs by back-propagating a corrective error signal through the network. For computer vision tasks, convolutional neural networks (CNNs) have proven to be effective. Recently, several clinical applications of CNNs have been proposed and studied in radiology for classification, detection, and segmentation tasks. This article reviews the key concepts of deep learning for clinical radiologists, discusses technical requirements, describes emerging applications in clinical radiology, and outlines limitations and future directions in this field. Radiologists should become familiar with the principles and potential applications of deep learning in medical imaging. © RSNA, 2017.

Cognitive versus stimulus-response theories of learning

PubMed Central

Holland, Peter C.

2010-01-01

In his 1948 address to the Division of Theoretical-Experimental Psychology of the American Psychological Association, Kenneth W. Spence discussed six distinctions between cognitive and stimulus-response (S-R) theories of learning. In this article I first review these six distinctions, and then focus on two of them in the context of my own research. This research concerns the specification of stimulus-stimulus associations in associative learning, and the characterization of neural systems underlying those associations. In the course of describing Spence's views and my research, I hope to communicate some of the richness of Spence's S-R psychology, and its currency within modern scientific analyses of behavior. PMID:18683467

Using Nd and Sr isotopes to trace dust and volcanic inputs to soils on French Guadeloupe Island

NASA Astrophysics Data System (ADS)

Guo, J.; Pereyra, Y.; Ma, L.; Gaillardet, J.; Sak, P. B.; Bouchez, J.

2017-12-01

Soil is at the central part of the Critical Zone for its important roles in sustaining ecosystems and agriculture. At French Guadeloupe, a tropical humid volcanic island, previous studies have shown that the mineral nutrient elements such as K, Na, Ca, and Mg are highly depleted in the surface soil. And mineral nutrients introduced by dusts are an important mineral nutrient source for vegetation growth in this area. It is important to understand and quantify the sources of the mineral dust added to surface soils. Nd isotope ratios, due to their distinct signatures between two unique end-members in soils for this area: the young volcanic areas like Guadeloupe and the dust source region from the old continental shields like Sahara Desert, can be a robust tracer to understand this critical process. Nevertheless, Sr isotope ratios can trace the inputs of marine aerosols. Here we present a new Nd isotope study on Guadeloupe soil depth profiles, combined with previous Sr isotope data, to fingerprint the sources of dust and volcanic inputs into soils. Soil samples from three surface profiles (0 - 1000cm deep) at different locations of the Guadeloupe Island were systematically analyzed. The results show distinct depth variations for Nd isotope signature along profiles. For all profiles, deep soils are relatively consisted with bedrock value (ɛNd: 5.05). But in surface soils (0-600cm), unlike Sr isotope ratios that are significantly modified by marine aerosol input, Nd isotope ratios show similar decrease (to ɛNd:-10) and frequent fluctuations toward the surface, suggesting dust is the dominant source of Nd in these soils. This conclusion is further supported by REE and other trace element data. Thus, with a simplified two end-member model, Sahara dust contributes the Nd percentages in soils varying from 10.7% at the deepest profiles to 69.5% on surface, showing a significant amount of Nd on the surface soil came from dust source. The deep soil profiles are also characterized by the presence of Nd isotope spikes with negative values, suggesting dust signatures at depth. Such a feature could be related to the presence of a paleo-soil surface at the spike depth that was buried by later volcanic eruption. Both Nd and Sr isotopes hence show dust and volcanic inputs are important factors for soil developments on French Guadeloupe Island.

Distributed deep learning networks among institutions for medical imaging.

PubMed

Chang, Ken; Balachandar, Niranjan; Lam, Carson; Yi, Darvin; Brown, James; Beers, Andrew; Rosen, Bruce; Rubin, Daniel L; Kalpathy-Cramer, Jayashree

2018-03-29

Deep learning has become a promising approach for automated support for clinical diagnosis. When medical data samples are limited, collaboration among multiple institutions is necessary to achieve high algorithm performance. However, sharing patient data often has limitations due to technical, legal, or ethical concerns. In this study, we propose methods of distributing deep learning models as an attractive alternative to sharing patient data. We simulate the distribution of deep learning models across 4 institutions using various training heuristics and compare the results with a deep learning model trained on centrally hosted patient data. The training heuristics investigated include ensembling single institution models, single weight transfer, and cyclical weight transfer. We evaluated these approaches for image classification in 3 independent image collections (retinal fundus photos, mammography, and ImageNet). We find that cyclical weight transfer resulted in a performance that was comparable to that of centrally hosted patient data. We also found that there is an improvement in the performance of cyclical weight transfer heuristic with a high frequency of weight transfer. We show that distributing deep learning models is an effective alternative to sharing patient data. This finding has implications for any collaborative deep learning study.

Strategies for Effective Faculty Involvement in Online Activities Aimed at Promoting Critical Thinking and Deep Learning

ERIC Educational Resources Information Center

Abdul Razzak, Nina

2016-01-01

Highly-traditional education systems that mainly offer what is known as "direct instruction" usually result in graduates with a surface approach to learning rather than a deep one. What is meant by deep-learning is learning that involves critical analysis, the linking of ideas and concepts, creative problem solving, and application…

Deep learning applications in ophthalmology.

PubMed

Rahimy, Ehsan

2018-05-01

To describe the emerging applications of deep learning in ophthalmology. Recent studies have shown that various deep learning models are capable of detecting and diagnosing various diseases afflicting the posterior segment of the eye with high accuracy. Most of the initial studies have centered around detection of referable diabetic retinopathy, age-related macular degeneration, and glaucoma. Deep learning has shown promising results in automated image analysis of fundus photographs and optical coherence tomography images. Additional testing and research is required to clinically validate this technology.

DeepInfer: open-source deep learning deployment toolkit for image-guided therapy

NASA Astrophysics Data System (ADS)

Mehrtash, Alireza; Pesteie, Mehran; Hetherington, Jorden; Behringer, Peter A.; Kapur, Tina; Wells, William M.; Rohling, Robert; Fedorov, Andriy; Abolmaesumi, Purang

2017-03-01

Deep learning models have outperformed some of the previous state-of-the-art approaches in medical image analysis. Instead of using hand-engineered features, deep models attempt to automatically extract hierarchical representations at multiple levels of abstraction from the data. Therefore, deep models are usually considered to be more flexible and robust solutions for image analysis problems compared to conventional computer vision models. They have demonstrated significant improvements in computer-aided diagnosis and automatic medical image analysis applied to such tasks as image segmentation, classification and registration. However, deploying deep learning models often has a steep learning curve and requires detailed knowledge of various software packages. Thus, many deep models have not been integrated into the clinical research work ows causing a gap between the state-of-the-art machine learning in medical applications and evaluation in clinical research procedures. In this paper, we propose "DeepInfer" - an open-source toolkit for developing and deploying deep learning models within the 3D Slicer medical image analysis platform. Utilizing a repository of task-specific models, DeepInfer allows clinical researchers and biomedical engineers to deploy a trained model selected from the public registry, and apply it to new data without the need for software development or configuration. As two practical use cases, we demonstrate the application of DeepInfer in prostate segmentation for targeted MRI-guided biopsy and identification of the target plane in 3D ultrasound for spinal injections.

DeepInfer: Open-Source Deep Learning Deployment Toolkit for Image-Guided Therapy.

PubMed

Mehrtash, Alireza; Pesteie, Mehran; Hetherington, Jorden; Behringer, Peter A; Kapur, Tina; Wells, William M; Rohling, Robert; Fedorov, Andriy; Abolmaesumi, Purang

2017-02-11

Deep learning models have outperformed some of the previous state-of-the-art approaches in medical image analysis. Instead of using hand-engineered features, deep models attempt to automatically extract hierarchical representations at multiple levels of abstraction from the data. Therefore, deep models are usually considered to be more flexible and robust solutions for image analysis problems compared to conventional computer vision models. They have demonstrated significant improvements in computer-aided diagnosis and automatic medical image analysis applied to such tasks as image segmentation, classification and registration. However, deploying deep learning models often has a steep learning curve and requires detailed knowledge of various software packages. Thus, many deep models have not been integrated into the clinical research workflows causing a gap between the state-of-the-art machine learning in medical applications and evaluation in clinical research procedures. In this paper, we propose "DeepInfer" - an open-source toolkit for developing and deploying deep learning models within the 3D Slicer medical image analysis platform. Utilizing a repository of task-specific models, DeepInfer allows clinical researchers and biomedical engineers to deploy a trained model selected from the public registry, and apply it to new data without the need for software development or configuration. As two practical use cases, we demonstrate the application of DeepInfer in prostate segmentation for targeted MRI-guided biopsy and identification of the target plane in 3D ultrasound for spinal injections.

DeepInfer: Open-Source Deep Learning Deployment Toolkit for Image-Guided Therapy

PubMed Central

Mehrtash, Alireza; Pesteie, Mehran; Hetherington, Jorden; Behringer, Peter A.; Kapur, Tina; Wells, William M.; Rohling, Robert; Fedorov, Andriy; Abolmaesumi, Purang

2017-01-01

Deep learning models have outperformed some of the previous state-of-the-art approaches in medical image analysis. Instead of using hand-engineered features, deep models attempt to automatically extract hierarchical representations at multiple levels of abstraction from the data. Therefore, deep models are usually considered to be more flexible and robust solutions for image analysis problems compared to conventional computer vision models. They have demonstrated significant improvements in computer-aided diagnosis and automatic medical image analysis applied to such tasks as image segmentation, classification and registration. However, deploying deep learning models often has a steep learning curve and requires detailed knowledge of various software packages. Thus, many deep models have not been integrated into the clinical research workflows causing a gap between the state-of-the-art machine learning in medical applications and evaluation in clinical research procedures. In this paper, we propose “DeepInfer” – an open-source toolkit for developing and deploying deep learning models within the 3D Slicer medical image analysis platform. Utilizing a repository of task-specific models, DeepInfer allows clinical researchers and biomedical engineers to deploy a trained model selected from the public registry, and apply it to new data without the need for software development or configuration. As two practical use cases, we demonstrate the application of DeepInfer in prostate segmentation for targeted MRI-guided biopsy and identification of the target plane in 3D ultrasound for spinal injections. PMID:28615794

Landcover Classification Using Deep Fully Convolutional Neural Networks

NASA Astrophysics Data System (ADS)

Wang, J.; Li, X.; Zhou, S.; Tang, J.

2017-12-01

Land cover classification has always been an essential application in remote sensing. Certain image features are needed for land cover classification whether it is based on pixel or object-based methods. Different from other machine learning methods, deep learning model not only extracts useful information from multiple bands/attributes, but also learns spatial characteristics. In recent years, deep learning methods have been developed rapidly and widely applied in image recognition, semantic understanding, and other application domains. However, there are limited studies applying deep learning methods in land cover classification. In this research, we used fully convolutional networks (FCN) as the deep learning model to classify land covers. The National Land Cover Database (NLCD) within the state of Kansas was used as training dataset and Landsat images were classified using the trained FCN model. We also applied an image segmentation method to improve the original results from the FCN model. In addition, the pros and cons between deep learning and several machine learning methods were compared and explored. Our research indicates: (1) FCN is an effective classification model with an overall accuracy of 75%; (2) image segmentation improves the classification results with better match of spatial patterns; (3) FCN has an excellent ability of learning which can attains higher accuracy and better spatial patterns compared with several machine learning methods.

Deep kernel learning method for SAR image target recognition

NASA Astrophysics Data System (ADS)

Chen, Xiuyuan; Peng, Xiyuan; Duan, Ran; Li, Junbao

2017-10-01

With the development of deep learning, research on image target recognition has made great progress in recent years. Remote sensing detection urgently requires target recognition for military, geographic, and other scientific research. This paper aims to solve the synthetic aperture radar image target recognition problem by combining deep and kernel learning. The model, which has a multilayer multiple kernel structure, is optimized layer by layer with the parameters of Support Vector Machine and a gradient descent algorithm. This new deep kernel learning method improves accuracy and achieves competitive recognition results compared with other learning methods.

Interfacial coupling and polarization of perovskite ABO3 heterostructures

NASA Astrophysics Data System (ADS)

Wu, Lijun; Wang, Zhen; Zhang, Bangmin; Yu, Liping; Chow, G. M.; Tao, Jing; Han, Myung-Geun; Guo, Hangwen; Chen, Lina; Plummer, E. W.; Zhang, Jiandi; Zhu, Yimei

2017-02-01

Interfaces with subtle difference in atomic and electronic structures in perovskite ABO3 heterostructures often yield intriguingly different properties, yet their exact roles remain elusive. In this article, we report an integrated study of unusual transport, magnetic, and structural properties of Pr0.67Sr0.33MnO3 (PSMO) films and La0.67Sr0.33MnO3 (LSMO) films of various thicknesses on SrTiO3 (STO) substrate. In particular, using atomically resolved imaging and electron energy-loss spectroscopy (EELS), we measured interface related local lattice distortion, BO6 octahedral rotation and cation-anion displacement induced polarization. In the very thin PSMO film, an unexpected interface-induced ferromagnetic polaronic insulator phase was observed during the cubic-to-tetragonal phase transition of the substrate STO, due to the enhanced electron-phonon interaction and atomic disorder in the film. On the other hand, for the very thin LSMO films we observed a remarkably deep polarization in non-ferroelectric STO substrate near the interface. Combining the experimental results with first principles calculations, we propose that the observed deep polarization is induced by an electric field originating from oxygen vacancies that extend beyond a dozen unit-cells from the interface, thus providing important evidence of the role of defects in the emergent interface properties of transition metal oxides.

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

Rapid and accurate intraoperative pathological diagnosis by artificial intelligence with deep learning technology.

PubMed

Zhang, Jing; Song, Yanlin; Xia, Fan; Zhu, Chenjing; Zhang, Yingying; Song, Wenpeng; Xu, Jianguo; Ma, Xuelei

2017-09-01

Frozen section is widely used for intraoperative pathological diagnosis (IOPD), which is essential for intraoperative decision making. However, frozen section suffers from some drawbacks, such as time consuming and high misdiagnosis rate. Recently, artificial intelligence (AI) with deep learning technology has shown bright future in medicine. We hypothesize that AI with deep learning technology could help IOPD, with a computer trained by a dataset of intraoperative lesion images. Evidences supporting our hypothesis included the successful use of AI with deep learning technology in diagnosing skin cancer, and the developed method of deep-learning algorithm. Large size of the training dataset is critical to increase the diagnostic accuracy. The performance of the trained machine could be tested by new images before clinical use. Real-time diagnosis, easy to use and potential high accuracy were the advantages of AI for IOPD. In sum, AI with deep learning technology is a promising method to help rapid and accurate IOPD. Copyright © 2017 Elsevier Ltd. All rights reserved.

Automated analysis of high-content microscopy data with deep learning.

PubMed

Kraus, Oren Z; Grys, Ben T; Ba, Jimmy; Chong, Yolanda; Frey, Brendan J; Boone, Charles; Andrews, Brenda J

2017-04-18

Existing computational pipelines for quantitative analysis of high-content microscopy data rely on traditional machine learning approaches that fail to accurately classify more than a single dataset without substantial tuning and training, requiring extensive analysis. Here, we demonstrate that the application of deep learning to biological image data can overcome the pitfalls associated with conventional machine learning classifiers. Using a deep convolutional neural network (DeepLoc) to analyze yeast cell images, we show improved performance over traditional approaches in the automated classification of protein subcellular localization. We also demonstrate the ability of DeepLoc to classify highly divergent image sets, including images of pheromone-arrested cells with abnormal cellular morphology, as well as images generated in different genetic backgrounds and in different laboratories. We offer an open-source implementation that enables updating DeepLoc on new microscopy datasets. This study highlights deep learning as an important tool for the expedited analysis of high-content microscopy data. © 2017 The Authors. Published under the terms of the CC BY 4.0 license.

Benchmarking Deep Learning Models on Large Healthcare Datasets.

PubMed

Purushotham, Sanjay; Meng, Chuizheng; Che, Zhengping; Liu, Yan

2018-06-04

Deep learning models (aka Deep Neural Networks) have revolutionized many fields including computer vision, natural language processing, speech recognition, and is being increasingly used in clinical healthcare applications. However, few works exist which have benchmarked the performance of the deep learning models with respect to the state-of-the-art machine learning models and prognostic scoring systems on publicly available healthcare datasets. In this paper, we present the benchmarking results for several clinical prediction tasks such as mortality prediction, length of stay prediction, and ICD-9 code group prediction using Deep Learning models, ensemble of machine learning models (Super Learner algorithm), SAPS II and SOFA scores. We used the Medical Information Mart for Intensive Care III (MIMIC-III) (v1.4) publicly available dataset, which includes all patients admitted to an ICU at the Beth Israel Deaconess Medical Center from 2001 to 2012, for the benchmarking tasks. Our results show that deep learning models consistently outperform all the other approaches especially when the 'raw' clinical time series data is used as input features to the models. Copyright © 2018 Elsevier Inc. All rights reserved.

A Robust Deep Model for Improved Classification of AD/MCI Patients

PubMed Central

Li, Feng; Tran, Loc; Thung, Kim-Han; Ji, Shuiwang; Shen, Dinggang; Li, Jiang

2015-01-01

Accurate classification of Alzheimer’s Disease (AD) and its prodromal stage, Mild Cognitive Impairment (MCI), plays a critical role in possibly preventing progression of memory impairment and improving quality of life for AD patients. Among many research tasks, it is of particular interest to identify noninvasive imaging biomarkers for AD diagnosis. In this paper, we present a robust deep learning system to identify different progression stages of AD patients based on MRI and PET scans. We utilized the dropout technique to improve classical deep learning by preventing its weight co-adaptation, which is a typical cause of over-fitting in deep learning. In addition, we incorporated stability selection, an adaptive learning factor, and a multi-task learning strategy into the deep learning framework. We applied the proposed method to the ADNI data set and conducted experiments for AD and MCI conversion diagnosis. Experimental results showed that the dropout technique is very effective in AD diagnosis, improving the classification accuracies by 5.9% on average as compared to the classical deep learning methods. PMID:25955998

Self-Paced Prioritized Curriculum Learning With Coverage Penalty in Deep Reinforcement Learning.

PubMed

Ren, Zhipeng; Dong, Daoyi; Li, Huaxiong; Chen, Chunlin; Zhipeng Ren; Daoyi Dong; Huaxiong Li; Chunlin Chen; Dong, Daoyi; Li, Huaxiong; Chen, Chunlin; Ren, Zhipeng

2018-06-01

In this paper, a new training paradigm is proposed for deep reinforcement learning using self-paced prioritized curriculum learning with coverage penalty. The proposed deep curriculum reinforcement learning (DCRL) takes the most advantage of experience replay by adaptively selecting appropriate transitions from replay memory based on the complexity of each transition. The criteria of complexity in DCRL consist of self-paced priority as well as coverage penalty. The self-paced priority reflects the relationship between the temporal-difference error and the difficulty of the current curriculum for sample efficiency. The coverage penalty is taken into account for sample diversity. With comparison to deep Q network (DQN) and prioritized experience replay (PER) methods, the DCRL algorithm is evaluated on Atari 2600 games, and the experimental results show that DCRL outperforms DQN and PER on most of these games. More results further show that the proposed curriculum training paradigm of DCRL is also applicable and effective for other memory-based deep reinforcement learning approaches, such as double DQN and dueling network. All the experimental results demonstrate that DCRL can achieve improved training efficiency and robustness for deep reinforcement learning.

Opportunities and obstacles for deep learning in biology and medicine.

PubMed

Ching, Travers; Himmelstein, Daniel S; Beaulieu-Jones, Brett K; Kalinin, Alexandr A; Do, Brian T; Way, Gregory P; Ferrero, Enrico; Agapow, Paul-Michael; Zietz, Michael; Hoffman, Michael M; Xie, Wei; Rosen, Gail L; Lengerich, Benjamin J; Israeli, Johnny; Lanchantin, Jack; Woloszynek, Stephen; Carpenter, Anne E; Shrikumar, Avanti; Xu, Jinbo; Cofer, Evan M; Lavender, Christopher A; Turaga, Srinivas C; Alexandari, Amr M; Lu, Zhiyong; Harris, David J; DeCaprio, Dave; Qi, Yanjun; Kundaje, Anshul; Peng, Yifan; Wiley, Laura K; Segler, Marwin H S; Boca, Simina M; Swamidass, S Joshua; Huang, Austin; Gitter, Anthony; Greene, Casey S

2018-04-01

Deep learning describes a class of machine learning algorithms that are capable of combining raw inputs into layers of intermediate features. These algorithms have recently shown impressive results across a variety of domains. Biology and medicine are data-rich disciplines, but the data are complex and often ill-understood. Hence, deep learning techniques may be particularly well suited to solve problems of these fields. We examine applications of deep learning to a variety of biomedical problems-patient classification, fundamental biological processes and treatment of patients-and discuss whether deep learning will be able to transform these tasks or if the biomedical sphere poses unique challenges. Following from an extensive literature review, we find that deep learning has yet to revolutionize biomedicine or definitively resolve any of the most pressing challenges in the field, but promising advances have been made on the prior state of the art. Even though improvements over previous baselines have been modest in general, the recent progress indicates that deep learning methods will provide valuable means for speeding up or aiding human investigation. Though progress has been made linking a specific neural network's prediction to input features, understanding how users should interpret these models to make testable hypotheses about the system under study remains an open challenge. Furthermore, the limited amount of labelled data for training presents problems in some domains, as do legal and privacy constraints on work with sensitive health records. Nonetheless, we foresee deep learning enabling changes at both bench and bedside with the potential to transform several areas of biology and medicine. © 2018 The Authors.

Opportunities and obstacles for deep learning in biology and medicine

PubMed Central

2018-01-01

Deep learning describes a class of machine learning algorithms that are capable of combining raw inputs into layers of intermediate features. These algorithms have recently shown impressive results across a variety of domains. Biology and medicine are data-rich disciplines, but the data are complex and often ill-understood. Hence, deep learning techniques may be particularly well suited to solve problems of these fields. We examine applications of deep learning to a variety of biomedical problems—patient classification, fundamental biological processes and treatment of patients—and discuss whether deep learning will be able to transform these tasks or if the biomedical sphere poses unique challenges. Following from an extensive literature review, we find that deep learning has yet to revolutionize biomedicine or definitively resolve any of the most pressing challenges in the field, but promising advances have been made on the prior state of the art. Even though improvements over previous baselines have been modest in general, the recent progress indicates that deep learning methods will provide valuable means for speeding up or aiding human investigation. Though progress has been made linking a specific neural network's prediction to input features, understanding how users should interpret these models to make testable hypotheses about the system under study remains an open challenge. Furthermore, the limited amount of labelled data for training presents problems in some domains, as do legal and privacy constraints on work with sensitive health records. Nonetheless, we foresee deep learning enabling changes at both bench and bedside with the potential to transform several areas of biology and medicine. PMID:29618526

Unsupervised deep learning reveals prognostically relevant subtypes of glioblastoma.

PubMed

Young, Jonathan D; Cai, Chunhui; Lu, Xinghua

2017-10-03

One approach to improving the personalized treatment of cancer is to understand the cellular signaling transduction pathways that cause cancer at the level of the individual patient. In this study, we used unsupervised deep learning to learn the hierarchical structure within cancer gene expression data. Deep learning is a group of machine learning algorithms that use multiple layers of hidden units to capture hierarchically related, alternative representations of the input data. We hypothesize that this hierarchical structure learned by deep learning will be related to the cellular signaling system. Robust deep learning model selection identified a network architecture that is biologically plausible. Our model selection results indicated that the 1st hidden layer of our deep learning model should contain about 1300 hidden units to most effectively capture the covariance structure of the input data. This agrees with the estimated number of human transcription factors, which is approximately 1400. This result lends support to our hypothesis that the 1st hidden layer of a deep learning model trained on gene expression data may represent signals related to transcription factor activation. Using the 3rd hidden layer representation of each tumor as learned by our unsupervised deep learning model, we performed consensus clustering on all tumor samples-leading to the discovery of clusters of glioblastoma multiforme with differential survival. One of these clusters contained all of the glioblastoma samples with G-CIMP, a known methylation phenotype driven by the IDH1 mutation and associated with favorable prognosis, suggesting that the hidden units in the 3rd hidden layer representations captured a methylation signal without explicitly using methylation data as input. We also found differentially expressed genes and well-known mutations (NF1, IDH1, EGFR) that were uniquely correlated with each of these clusters. Exploring these unique genes and mutations will allow us to further investigate the disease mechanisms underlying each of these clusters. In summary, we show that a deep learning model can be trained to represent biologically and clinically meaningful abstractions of cancer gene expression data. Understanding what additional relationships these hidden layer abstractions have with the cancer cellular signaling system could have a significant impact on the understanding and treatment of cancer.

The effects of teacher read-alouds and student silent reading on predominantly bilingual high school seniors’ learning and retention of social studies content

PubMed Central

Reed, Deborah K.; Swanson, Elizabeth; Petscher, Yaacov; Vaughn, Sharon

2015-01-01

Teacher read-alouds (TRA) are common in middle and high school content area classes. Because the practice of reading the textbook out loud to students is often used out of concern about students’ ability to understand and learn from text when reading silently (SR), this randomized controlled trial was designed to experimentally manipulate text reading while blocking on all other instructional elements to determine the relative effects on learning content. Predominantly Spanish–English bilingual twelfth-graders (n = 123) were randomly assigned to either a TRA or SR condition and provided 1 week of high quality instruction in US history. Daily lessons included teaching key terms in the passage, previewing text headings, and conducting comprehension checks. Results of immediate, 1-week delayed, and 1-month delayed assessments of content learning revealed no significant differences between the two groups. Students were also asked to rate the method of reading they believed best helped them understand and remember information. Students in the SR condition more consistently agreed that reading silently was beneficial. Findings suggest low performing adolescents of different linguistic backgrounds can learn content as well when reading appropriately challenging text silently as when the teacher reads the text aloud to them. PMID:26346215

Opportunities and challenges in developing deep learning models using electronic health records data: a systematic review.

PubMed

Xiao, Cao; Choi, Edward; Sun, Jimeng

2018-06-08

To conduct a systematic review of deep learning models for electronic health record (EHR) data, and illustrate various deep learning architectures for analyzing different data sources and their target applications. We also highlight ongoing research and identify open challenges in building deep learning models of EHRs. We searched PubMed and Google Scholar for papers on deep learning studies using EHR data published between January 1, 2010, and January 31, 2018. We summarize them according to these axes: types of analytics tasks, types of deep learning model architectures, special challenges arising from health data and tasks and their potential solutions, as well as evaluation strategies. We surveyed and analyzed multiple aspects of the 98 articles we found and identified the following analytics tasks: disease detection/classification, sequential prediction of clinical events, concept embedding, data augmentation, and EHR data privacy. We then studied how deep architectures were applied to these tasks. We also discussed some special challenges arising from modeling EHR data and reviewed a few popular approaches. Finally, we summarized how performance evaluations were conducted for each task. Despite the early success in using deep learning for health analytics applications, there still exist a number of issues to be addressed. We discuss them in detail including data and label availability, the interpretability and transparency of the model, and ease of deployment.

Learning Deep Representations for Ground to Aerial Geolocalization (Open Access)

DTIC Science & Technology

2015-10-15

proposed approach, Where-CNN, is inspired by deep learning success in face verification and achieves significant improvements over tra- ditional hand...crafted features and existing deep features learned from other large-scale databases. We show the ef- fectiveness of Where-CNN in finding matches

Do Students Develop towards More Deep Approaches to Learning during Studies? A Systematic Review on the Development of Students' Deep and Surface Approaches to Learning in Higher Education

ERIC Educational Resources Information Center

Asikainen, Henna; Gijbels, David

2017-01-01

The focus of the present paper is on the contribution of the research in the student approaches to learning tradition. Several studies in this field have started from the assumption that students' approaches to learning develop towards more deep approaches to learning in higher education. This paper reports on a systematic review of longitudinal…

3D Deep Learning Angiography (3D-DLA) from C-arm Conebeam CT.

PubMed

Montoya, J C; Li, Y; Strother, C; Chen, G-H

2018-05-01

Deep learning is a branch of artificial intelligence that has demonstrated unprecedented performance in many medical imaging applications. Our purpose was to develop a deep learning angiography method to generate 3D cerebral angiograms from a single contrast-enhanced C-arm conebeam CT acquisition in order to reduce image artifacts and radiation dose. A set of 105 3D rotational angiography examinations were randomly selected from an internal data base. All were acquired using a clinical system in conjunction with a standard injection protocol. More than 150 million labeled voxels from 35 subjects were used for training. A deep convolutional neural network was trained to classify each image voxel into 3 tissue types (vasculature, bone, and soft tissue). The trained deep learning angiography model was then applied for tissue classification into a validation cohort of 8 subjects and a final testing cohort of the remaining 62 subjects. The final vasculature tissue class was used to generate the 3D deep learning angiography images. To quantify the generalization error of the trained model, we calculated the accuracy, sensitivity, precision, and Dice similarity coefficients for vasculature classification in relevant anatomy. The 3D deep learning angiography and clinical 3D rotational angiography images were subjected to a qualitative assessment for the presence of intersweep motion artifacts. Vasculature classification accuracy and 95% CI in the testing dataset were 98.7% (98.3%-99.1%). No residual signal from osseous structures was observed for any 3D deep learning angiography testing cases except for small regions in the otic capsule and nasal cavity compared with 37% (23/62) of the 3D rotational angiographies. Deep learning angiography accurately recreated the vascular anatomy of the 3D rotational angiography reconstructions without a mask. Deep learning angiography reduced misregistration artifacts induced by intersweep motion, and it reduced radiation exposure required to obtain clinically useful 3D rotational angiography. © 2018 by American Journal of Neuroradiology.

Machine Learning, deep learning and optimization in computer vision

NASA Astrophysics Data System (ADS)

Canu, Stéphane

2017-03-01

As quoted in the Large Scale Computer Vision Systems NIPS workshop, computer vision is a mature field with a long tradition of research, but recent advances in machine learning, deep learning, representation learning and optimization have provided models with new capabilities to better understand visual content. The presentation will go through these new developments in machine learning covering basic motivations, ideas, models and optimization in deep learning for computer vision, identifying challenges and opportunities. It will focus on issues related with large scale learning that is: high dimensional features, large variety of visual classes, and large number of examples.

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.

Distribution of rubidium, strontium, and zirconium in tuff from two deep coreholes at Yucca Mountain, Nevada

USGS Publications Warehouse

Spengler, Richard W.; Peterman, Zell E.; ,

1991-01-01

Variations in concentrations of trace elements Rb, Sr, and Zr within the sequence of high-silica tuff and dacitic lava beneath Yucca Mountain reflect both primary composition and secondary alteration. Rb and K concentrations have parallel trends. Rb concentrations are significantly lower within intervals containing zeolitic nonwelded to partially welded and bedded tuffs and are higher in thick moderately to densely welded zones. Sr concentrations increase with depth from about 30 ppm in the Topopah Spring Member of the Paintbrush Tuff to almost 300 ppm in the older tuffs. Zr concentrations are about 100 ppm in the Topopah Spring Member and also increase with depth to about 150 ppm in the Lithic Ridge Tuff and upper part of the older tuffs. Conspicuous local high concentrations of Sr in the lower part of the Tram Member, in the dacite lava, and in unit c of the older tuffs in USW G-1, and in the densely welded zone of the Bullfrog Member in USW GU-3/G-3 closely correlate with high concentrations of less-mobile Zr and may reflect either primary composition or elemental redistribution resulting largely from smectitic alteration. Initial 87Sr/86Sr values from composite samples increase upward in units above the Bullfrog Member of the Crater Flat Tuff. The progressive tenfold increase in Sr with depth coupled with the similarity of initial 87Sr/86Sr values within the Bullfrog Member and older units to those of Paleozoic marine carbonates are consistent with a massive influx of Sr from water derived from a Paleozoic carbonate aquifer.

The sRNAome mining revealed existence of unique signature small RNAs derived from 5.8SrRNA from Piper nigrum and other plant lineages.

PubMed

Asha, Srinivasan; Soniya, E V

2017-02-01

Small RNAs derived from ribosomal RNAs (srRNAs) are rarely explored in the high-throughput data of plant systems. Here, we analyzed srRNAs from the deep-sequenced small RNA libraries of Piper nigrum, a unique magnoliid plant. The 5' end of the putative long form of 5.8S rRNA (5.8S L rRNA) was identified as the site for biogenesis of highly abundant srRNAs that are unique among the Piperaceae family of plants. A subsequent comparative analysis of the ninety-seven sRNAomes of diverse plants successfully uncovered the abundant existence and precise cleavage of unique rRF signature small RNAs upstream of a novel 5' consensus sequence of the 5.8S rRNA. The major cleavage process mapped identically among the different tissues of the same plant. The differential expression and cleavage of 5'5.8S srRNAs in Phytophthora capsici infected P. nigrum tissues indicated the critical biological functions of these srRNAs during stress response. The non-canonical short hairpin precursor structure, the association with Argonaute proteins, and the potential targets of 5'5.8S srRNAs reinforced their regulatory role in the RNAi pathway in plants. In addition, this novel lineage specific small RNAs may have tremendous biological potential in the taxonomic profiling of plants.

The sRNAome mining revealed existence of unique signature small RNAs derived from 5.8SrRNA from Piper nigrum and other plant lineages

PubMed Central

Asha, Srinivasan; Soniya, E. V.

2017-01-01

Small RNAs derived from ribosomal RNAs (srRNAs) are rarely explored in the high-throughput data of plant systems. Here, we analyzed srRNAs from the deep-sequenced small RNA libraries of Piper nigrum, a unique magnoliid plant. The 5′ end of the putative long form of 5.8S rRNA (5.8SLrRNA) was identified as the site for biogenesis of highly abundant srRNAs that are unique among the Piperaceae family of plants. A subsequent comparative analysis of the ninety-seven sRNAomes of diverse plants successfully uncovered the abundant existence and precise cleavage of unique rRF signature small RNAs upstream of a novel 5′ consensus sequence of the 5.8S rRNA. The major cleavage process mapped identically among the different tissues of the same plant. The differential expression and cleavage of 5′5.8S srRNAs in Phytophthora capsici infected P. nigrum tissues indicated the critical biological functions of these srRNAs during stress response. The non-canonical short hairpin precursor structure, the association with Argonaute proteins, and the potential targets of 5′5.8S srRNAs reinforced their regulatory role in the RNAi pathway in plants. In addition, this novel lineage specific small RNAs may have tremendous biological potential in the taxonomic profiling of plants. PMID:28145468

Sr-Nd isotopes constrain on the deposit history of the basins in the Gulf of Mexico

NASA Astrophysics Data System (ADS)

Li, Y.; Jiang, S.

2015-12-01

The Brazos-Trinity Basin IV and Ursa Basin are situated on the northern slope of the Gulf of Mexico. The Ursa basin lies in the center of late Pleistocene Mississippi River deposition, received the sediment deposition during Marine Isotope Stage (MIS) 2- 4. The Brazos-Trinity Basin IV belongs to a part of the Brazos-Trinity fan, it recorded the turbidite deposition and hemiplegic deposition during MIS1- 5. The Sr and Nd isotopic composition of the detrital composition of the sediment in both basins indicates the change of the sediment provenance during the basin-filled process. In the Ursa basin, The difference of 87Sr/86Sr ratio and ɛNd of the detrital component between MIS1,2 (87Sr/86Sr ~ 0.7219 - 0.7321, ɛNd ~ -12 - -13.4) and MIS3,4(87Sr/86Sr ~ 0.7310 - 0.7354, ɛNd ~ -16 - -17.9) is suggested to be related with the provenance change of the detrital particles since LGM. The addition of detrital particle from Appalachians with less radiogenic 87Sr/86Sr and positive ɛNd altered the character of the sediment of the Mississippi River during the last glaciation and deglaciation. In the Brazos-Trinity Basin IV, the narrow range of 87Sr/86Sr and ɛNd indicate that the sediment source of Brazos-Trinity Basin IV had not changed obviously during MIS5e to MIS2, mostly from coastal rivers such as Brazos River, Trinity River and Sabine River. The pre-fan with 87Sr/86Sr ~0.735 and ɛNd ~ -14.5 to -16.9, which is very similar to the deep sediment in the Ursa Basin with 87Sr/86Sr ~0.733 to 0.735 and ɛNd ~ -16 to -18. It is suggested that sediments of the pre-fan of the Brazos-Trinity Basin IV were supplied from the ancestral Mississippi River Delta during the low sea level (MIS 6). During the MIS5, the discharge of Mississippi River is thought switched to its present course, ~300 km to the east.

A Deep Learning Method to Automatically Identify Reports of Scientifically Rigorous Clinical Research from the Biomedical Literature: Comparative Analytic Study.

PubMed

Del Fiol, Guilherme; Michelson, Matthew; Iorio, Alfonso; Cotoi, Chris; Haynes, R Brian

2018-06-25

A major barrier to the practice of evidence-based medicine is efficiently finding scientifically sound studies on a given clinical topic. To investigate a deep learning approach to retrieve scientifically sound treatment studies from the biomedical literature. We trained a Convolutional Neural Network using a noisy dataset of 403,216 PubMed citations with title and abstract as features. The deep learning model was compared with state-of-the-art search filters, such as PubMed's Clinical Query Broad treatment filter, McMaster's textword search strategy (no Medical Subject Heading, MeSH, terms), and Clinical Query Balanced treatment filter. A previously annotated dataset (Clinical Hedges) was used as the gold standard. The deep learning model obtained significantly lower recall than the Clinical Queries Broad treatment filter (96.9% vs 98.4%; P<.001); and equivalent recall to McMaster's textword search (96.9% vs 97.1%; P=.57) and Clinical Queries Balanced filter (96.9% vs 97.0%; P=.63). Deep learning obtained significantly higher precision than the Clinical Queries Broad filter (34.6% vs 22.4%; P<.001) and McMaster's textword search (34.6% vs 11.8%; P<.001), but was significantly lower than the Clinical Queries Balanced filter (34.6% vs 40.9%; P<.001). Deep learning performed well compared to state-of-the-art search filters, especially when citations were not indexed. Unlike previous machine learning approaches, the proposed deep learning model does not require feature engineering, or time-sensitive or proprietary features, such as MeSH terms and bibliometrics. Deep learning is a promising approach to identifying reports of scientifically rigorous clinical research. Further work is needed to optimize the deep learning model and to assess generalizability to other areas, such as diagnosis, etiology, and prognosis. ©Guilherme Del Fiol, Matthew Michelson, Alfonso Iorio, Chris Cotoi, R Brian Haynes. Originally published in the Journal of Medical Internet Research (http://www.jmir.org), 25.06.2018.

A Constructivist View of Music Education: Perspectives for Deep Learning

ERIC Educational Resources Information Center

Scott, Sheila

2006-01-01

The article analyzes a constructivist view of music education. A constructivist music classroom exemplifies deep learning when students formulate questions, acquire new knowledge by developing and implementing plans for investigating these questions, and reflect on the results. A context for deep learning requires that teachers and students work…

Sleep Deprivation Alters Rat Ventral Prostate Morphology, Leading to Glandular Atrophy: A Microscopic Study Contrasted with the Hormonal Assays

PubMed Central

Venâncio, Daniel P.; Andersen, Monica L.; Vilamaior, Patricia S. L.; Santos, Fernanda C.; Zager, Adriano; Tufik, Sérgio; Taboga, Sebastião R.; De Mello, Marco T.

2012-01-01

We investigated the effect of 96 h paradoxical sleep deprivation (PSD) and 21-day sleep restriction (SR) on prostate morphology using stereological assays in male rats. After euthanasia, the rat ventral prostate was removed, weighed, and prepared for conventional light microscopy. Microscopic analysis of the prostate reveals that morphology of this gland was altered after 96 h of PSD and 21 days of SR, with the most important alterations occurring in the epithelium and stroma in the course of both procedures compared with the control group. Both 96 h PSD and 21-day SR rats showed lower serum testosterone and higher corticosterone levels than control rats. The significance of our result referring to the sleep deprivation was responsible for deep morphological alterations in ventral prostate tissue, like to castration microscopic modifications. This result is due to the marked alterations in hormonal status caused by PSD and SR. PMID:22927719

Active semi-supervised learning method with hybrid deep belief networks.

PubMed

Zhou, Shusen; Chen, Qingcai; Wang, Xiaolong

2014-01-01

In this paper, we develop a novel semi-supervised learning algorithm called active hybrid deep belief networks (AHD), to address the semi-supervised sentiment classification problem with deep learning. First, we construct the previous several hidden layers using restricted Boltzmann machines (RBM), which can reduce the dimension and abstract the information of the reviews quickly. Second, we construct the following hidden layers using convolutional restricted Boltzmann machines (CRBM), which can abstract the information of reviews effectively. Third, the constructed deep architecture is fine-tuned by gradient-descent based supervised learning with an exponential loss function. Finally, active learning method is combined based on the proposed deep architecture. We did several experiments on five sentiment classification datasets, and show that AHD is competitive with previous semi-supervised learning algorithm. Experiments are also conducted to verify the effectiveness of our proposed method with different number of labeled reviews and unlabeled reviews respectively.

Theory of Self- vs. Externally-Regulated LearningTM: Fundamentals, Evidence, and Applicability.

PubMed

de la Fuente-Arias, Jesús

2017-01-01

The Theory of Self- vs. Externally-Regulated Learning TM has integrated the variables of SRL theory, the DEDEPRO model, and the 3P model. This new Theory has proposed: (a) in general, the importance of the cyclical model of individual self-regulation (SR) and of external regulation stemming from the context (ER), as two different and complementary variables, both in combination and in interaction; (b) specifically, in the teaching-learning context, the relevance of different types of combinations between levels of self-regulation (SR) and of external regulation (ER) in the prediction of self-regulated learning (SRL), and of cognitive-emotional achievement. This review analyzes the assumptions, conceptual elements, empirical evidence, benefits and limitations of SRL vs. ERL Theory . Finally, professional fields of application and future lines of research are suggested.

Deep dissection: motivating students beyond rote learning in veterinary anatomy.

PubMed

Cake, Martin A

2006-01-01

The profusion of descriptive, factual information in veterinary anatomy inevitably creates pressure on students to employ surface learning approaches and "rote learning." This phenomenon may contribute to negative perceptions of the relevance of anatomy as a discipline. Thus, encouraging deep learning outcomes will not only lead to greater satisfaction for both instructors and learners but may have the added effect of raising the profile of and respect for the discipline. Consideration of the literature reveals the broad scope of interventions required to motivate students to go beyond rote learning. While many of these are common to all disciplines (e.g., promoting active learning, making higher-order goals explicit, reducing content in favor of concepts, aligning assessment with outcomes), other factors are peculiar to anatomy, such as the benefits of incorporating clinical tidbits, "living anatomy," the anatomy museum, and dissection classes into a "learning context" that fosters deep approaches. Surprisingly, the 10 interventions discussed focus more on factors contributing to student perceptions of the course than on drastic changes to the anatomy course itself. This is because many traditional anatomy practices, such as dissection and museum-based classes, are eminently compatible with active, student-centered learning strategies and the adoption of deep learning approaches by veterinary students. Thus the key to encouraging, for example, dissection for deep learning ("deep dissection") lies more in student motivation, personal engagement, curriculum structure, and "learning context" than in the nature of the learning activity itself.

A lower-extremities kinematic comparison of deep-water running styles and treadmill running.

PubMed

Killgore, Garry L; Wilcox, Anthony R; Caster, Brian L; Wood, Terry M

2006-11-01

The purpose of this investigation was to identify a deep-water running (DWR) style that most closely approximates terrestrial running, particularly relative to the lower extremities. Twenty intercollegiate distance runners (women, N = 12; men, N = 8) were videotaped from the right sagittal view while running on a treadmill (TR) and in deep water at 55-60% of their TR VO(2)max using 2 DWR styles: cross-country (CC) and high-knee (HK). Variables of interest were horizontal (X) and vertical (Y) displacement of the knee and ankle, stride rate (SR), VO(2), heart rate (HR), and rating of perceived exertion (RPE). Multivariate omnibus tests revealed statistically significant differences for RPE (p < 0.001). The post hoc pairwise comparisons revealed significant differences between TR and both DWR styles (p < 0.001). The kinematic variables multivariate omnibus tests were found to be statistically significant (p < 0.001 to p < 0.019). The post hoc pairwise comparisons revealed significant differences in SR (p < 0.001) between TR (1.25 +/- 0.08 Hz) and both DWR styles and also between the CC (0.81 +/- 0.08 Hz) and HK (1.14 +/- 0.10 Hz) styles of DWR. The CC style of DWR was found to be similar to TR with respect to linear ankle displacement, whereas the HK style was significantly different from TR in all comparisons made for ankle and knee displacement. The CC style of DWR is recommended as an adjunct to distance running training if the goal is to mimic the specificity of the ankle linear horizontal displacement of land-based running, but the SR will be slower at a comparable percentage of VO(2)max.

Learning with Multiple Representations: Extending Multimedia Learning beyond the Lab

ERIC Educational Resources Information Center

Eilam, Billie; Poyas, Yael

2008-01-01

The present study extended multimedia learning principles beyond the lab to an ecologically valid setting (homework). Eighteen information cards were used to perform three homework tasks. The control group students learned from single representation (SR) cards that presented all information as printed text. The multiple representation (MR) group…

Cross-Cultural Service Learning: American and Russian Students Learn Applied Organizational Communication.

ERIC Educational Resources Information Center

Stevens, Betsy

2001-01-01

Describes how American and Russian students engaged in service learning in their own communities as part of an organizational communication class in which they learned communication principles and applied their skills to assist non-profit organizations. Describes both projects, stumbling blocks, and course outcomes. (SR)

Deep learning based classification for head and neck cancer detection with hyperspectral imaging in an animal model

NASA Astrophysics Data System (ADS)

Ma, Ling; Lu, Guolan; Wang, Dongsheng; Wang, Xu; Chen, Zhuo Georgia; Muller, Susan; Chen, Amy; Fei, Baowei

2017-03-01

Hyperspectral imaging (HSI) is an emerging imaging modality that can provide a noninvasive tool for cancer detection and image-guided surgery. HSI acquires high-resolution images at hundreds of spectral bands, providing big data to differentiating different types of tissue. We proposed a deep learning based method for the detection of head and neck cancer with hyperspectral images. Since the deep learning algorithm can learn the feature hierarchically, the learned features are more discriminative and concise than the handcrafted features. In this study, we adopt convolutional neural networks (CNN) to learn the deep feature of pixels for classifying each pixel into tumor or normal tissue. We evaluated our proposed classification method on the dataset containing hyperspectral images from 12 tumor-bearing mice. Experimental results show that our method achieved an average accuracy of 91.36%. The preliminary study demonstrated that our deep learning method can be applied to hyperspectral images for detecting head and neck tumors in animal models.

Impacts of Dust on Tropical Volcanic Soil Formation: Insights from Strontium and Uranium-Series Isotopes in Soils from Basse-Terre Island, French Guadeloupe

NASA Astrophysics Data System (ADS)

Pereyra, Y.; Ma, L.; Sak, P. B.; Gaillardet, J.; Buss, H. L.; Brantley, S. L.

2015-12-01

Dust inputs play an important role in soil formation, especially for thick soils developed on tropical volcanic islands. In these regions, soils are highly depleted due to intensive chemical weathering, and mineral nutrients from dusts have been known to be important in sustaining soil fertility and productivity. Tropical volcanic soils are an ideal system to study the impacts of dust inputs on the ecosystem. Sr and U-series isotopes are excellent tracers to identify sources of materials in an open system if the end-members have distinctive isotope signatures. These two isotope systems are particularly useful to trace the origin of atmospheric inputs into soils and to determine rates and timescales of soil formation. This study analyzes major elemental concentrations, Sr and U-series isotope ratios in highly depleted soils in the tropical volcanic island of Basse-Terre in French Guadeloupe to determine atmospheric input sources and identify key soil formation processes. We focus on three soil profiles (8 to 12 m thick) from the Bras-David, Moustique Petit-Bourg, and Deshaies watersheds; and on the adjacent rivers to these sites. Results have shown a significant depletion of U, Sr, and major elements in the deep profile (12 to 4 m) attributed to rapid chemical weathering. The top soil profiles (4 m to the surface) all show addition of elements such as Ca, Mg, U, and Sr due to atmospheric dust. More importantly, the topsoil profiles have distinct Sr and U-series isotope compositions from the deep soils. Sr and U-series isotope ratios of the top soils and sequential extraction fractions confirm that the sources of the dust are from the Saharan dessert, through long distance transport from Africa to the Caribbean region across the Atlantic Ocean. During the transport, some dust isotope signatures may also have been modified by local volcanic ashes and marine aerosols. Our study highlights that dusts and marine aerosols play important roles in element cycles and nutrient sources in the highly depleted surface soils of tropical oceanic islands.

Revealing the magmas degassing below closed-conduit active volcanoes: noble gases in volcanic rocks versus fumarolic fluids at Vulcano (Aeolian Islands, Italy)

NASA Astrophysics Data System (ADS)

Mandarano, Michela; Paonita, Antonio; Martelli, Mauro; Viccaro, Marco; Nicotra, Eugenio; Millar, Ian L.

2016-04-01

With the aim to constrain the nature of magma currently feeding the fumarolic field of Vulcano, we measured the elemental and isotopic compositions of noble gases (He, Ne, and Ar) in olivine- and clinopyroxene-hosted fluid inclusions in high-K calcalcaline-shoshonitic and shoshonitic-potassic series so as to cover the entire volcanological history of Vulcano Island (Italy). The major and trace-element concentrations and the Sr- and Pb-isotope compositions for whole rocks were integrated with data obtained from the fluid inclusions. 3He/4He in fluid inclusions is within the range of 3.30 and 5.94 R/Ra, being lower than the value for the deep magmatic source expected for Vulcano Island (6.0-6.2 R/Ra). 3He/4He of the magmatic source is almost constant throughout the volcanic record of Vulcano. Integration of the He- and Sr-isotope systematics leads to the conclusion that a decrease in the He-isotope ratio of the rocks is mainly due to the assimilation of 10-25% of a crustal component similar to the Calabrian basement. 3He/4He shows a negative correlation with Sr isotopes except for the last-emitted Vulcanello latites (Punta del Roveto), which have high He- and Sr-isotope ratios. This anomaly has been attributed to a flushing process by fluids coming from the deepest reservoirs. Indeed, an input of deep magmatic volatiles with high 3He/4He values increases the He-isotope ratio without changing 87Sr/86Sr. A comparison of the He isotope ratios between fluid inclusions and fumarolic gases showed that only the basalts of La Sommata and the latites of Vulcanello have comparable values. Taking into account that the latites of Vulcanello relate to one of the most-recent eruptions at Vulcano (in the 17th century), we infer that that the most probable magma which actually feeds the fumarolic emissions is a latitic body ponding at about 3-3.5 km of depth and flushed by fluids coming from a deeper and basic magma.

Deglacial dust provenance changes in the Eastern Equatorial Pacific and implications for ITCZ movement

NASA Astrophysics Data System (ADS)

Xie, Ruifang C.; Marcantonio, Franco

2012-02-01

The provenance of eolian dust supplied to deep-sea sediments has the potential to offer insights into changes in past atmospheric circulation. Specifically, measuring temporal changes in dust provenance can shed light on changes in the mean position of the Intertropical Convergence Zone (ITCZ), a region acting as a barrier separating wind-blown material derived from northern versus southern hemisphere sources. Here we have analyzed Nd, Sr, and Pb isotope ratios in the operationally-defined detrital component extracted from deep-sea sediments in the eastern equatorial Pacific (EEP) along a meridional transect at 110°W from 3°S to 7°N (ODP Leg 138, sites 848-853). Sr isotope results show that barite Sr has a significant influence on 87Sr/86Sr isotope ratios of samples in the upwelling zone of the EEP. However, sites located >3° or more away from the equator (sites 852 and 853) are believed to not be affected by barite Sr and provide useful detrital Sr signals. 208Pb/206Pb and 207Pb/206Pb ratios in all cores fall into the Pb-isotope space of five potential dust sources (Asia, North and Central/South America, Sahara, and Australia), with no distinct isotopic fingerprinting of the dominant source(s). ɛNd values were most valuable for discerning detrital source provenance, and their values at all sites, ranging from -5.46 to -3.25, were more unradiogenic for sediments deposited during the last glacial than for those deposited during the Holocene. There are distinct latitudinal trends in the ɛNd values, with more radiogenic values further south and less radiogenic values further north, excluding site 848. This distinction holds true for both Holocene and last glacial periods. For the most southerly site, 848, we invoke, for the first time, a distinct southern hemisphere Australian source as being responsible for the unradiogenic Nd isotope ratios. Both average last glacial and Holocene ɛNd values show similar sharp gradients along the transect between 5.29°N and 2.77°N, suggesting little movement of the glacial ITCZ in the EEP. However, during the deglacial, this gradient is stronger and shifted further north between 5.29°N and 7.21°N, suggesting a more northerly, possibly stronger, deglacial ITCZ.

Hydrologic and environmental controls on uranium-series and strontium isotope ratios in a natural weathering environment

NASA Astrophysics Data System (ADS)

White, A. M.; Ma, L.; Moravec, B. G.; McIntosh, J. C.; Chorover, J.

2017-12-01

In a remote, volcanic headwater catchment of the Jemez River Basin Critical Zone Observatory (JRB-CZO) in NM, stable water isotopes and solute chemistry have shown that snowmelt infiltrates and is stored before later discharging into springs and streams via subsurface flowpaths that vary seasonally. Therefore, water-rock reactions are also expected to change with season as hydrologic flowpaths transport water, gases and solutes through different biogeochemical conditions, rock types and fracture networks. Uranium-series isotopes have been shown to be a novel tracer of water-rock reactions and source water contributions while strontium isotopes are frequently used as indicators of chemical weathering and bedrock geology. This study combines both isotopes to understand how U and Sr isotope signatures evolve through the Critical Zone (CZ). More specifically, this work examines the relationship between seasonality, water transit time (WTT), and U-series and Sr isotopes in stream and spring waters from three catchments within the JRB-CZO, as well as lithology, rock type and CZ structure in solid phase cores. Samples from ten springs with known WTTs were analyzed for U and Sr isotopes to determine the effect of WTT on the isotopic composition of natural waters. Results suggest that WTT alone cannot explain the variability of U and Sr isotopes in JRB-CZO springs. Stream samples were also collected across two water years to establish how seasonality controls surface water isotopic composition. U and Sr isotope values vary with season, consistent with a previous study from the La Jara catchment; however, this study revealed that these changes do not show a systematic pattern among the three catchments suggesting that differences in the mineralogy and structure of the deep CZ in individual catchments, and partitioning of water along deep vs surficial and fracture vs matrix flow paths, likely also control isotopic variability. The distribution of U-series and Sr isotopes in core samples with depth shows distinct weathering profiles with variable 234U/238U activity and Sr isotope ratios. Comparison of the isotopic composition of cores and groundwaters from similar depths, as well as surface waters in the JRB-CZO will be vital for the characterization of hydrogeologic controls on isotopic composition in this complex terrain.

The Mediating Effect of Intrinsic Motivation to Learn on the Relationship between Student´s Autonomy Support and Vitality and Deep Learning.

PubMed

Núñez, Juan L; León, Jaime

2016-07-18

Self-determination theory has shown that autonomy support in the classroom is associated with an increase of students' intrinsic motivation. Moreover, intrinsic motivation is related with positive outcomes. This study examines the relationships between autonomy support, intrinsic motivation to learn and two motivational consequences, deep learning and vitality. Specifically, the hypotheses were that autonomy support predicts the two types of consequences, and that autonomy support directly and indirectly predicts the vitality and the deep learning through intrinsic motivation to learn. Participants were 276 undergraduate students. The mean age was 21.80 years (SD = 2.94). Structural equation modeling was used to test the relationships between variables and delta method was used to analyze the mediating effect of intrinsic motivation to learn. Results indicated that student perception of autonomy support had a positive effect on deep learning and vitality (p < .001). In addition, these associations were mediated by intrinsic motivation to learn. These findings suggest that teachers are key elements in generating of autonomy support environment to promote intrinsic motivation, deep learning, and vitality in classroom. Educational implications are discussed.

Deep Learning in Medical Image Analysis.

PubMed

Shen, Dinggang; Wu, Guorong; Suk, Heung-Il

2017-06-21

This review covers computer-assisted analysis of images in the field of medical imaging. Recent advances in machine learning, especially with regard to deep learning, are helping to identify, classify, and quantify patterns in medical images. At the core of these advances is the ability to exploit hierarchical feature representations learned solely from data, instead of features designed by hand according to domain-specific knowledge. Deep learning is rapidly becoming the state of the art, leading to enhanced performance in various medical applications. We introduce the fundamentals of deep learning methods and review their successes in image registration, detection of anatomical and cellular structures, tissue segmentation, computer-aided disease diagnosis and prognosis, and so on. We conclude by discussing research issues and suggesting future directions for further improvement.

Deep learning with convolutional neural networks for EEG decoding and visualization

PubMed Central

Springenberg, Jost Tobias; Fiederer, Lukas Dominique Josef; Glasstetter, Martin; Eggensperger, Katharina; Tangermann, Michael; Hutter, Frank; Burgard, Wolfram; Ball, Tonio

2017-01-01

Abstract Deep learning with convolutional neural networks (deep ConvNets) has revolutionized computer vision through end‐to‐end learning, that is, learning from the raw data. There is increasing interest in using deep ConvNets for end‐to‐end EEG analysis, but a better understanding of how to design and train ConvNets for end‐to‐end EEG decoding and how to visualize the informative EEG features the ConvNets learn is still needed. Here, we studied deep ConvNets with a range of different architectures, designed for decoding imagined or executed tasks from raw EEG. Our results show that recent advances from the machine learning field, including batch normalization and exponential linear units, together with a cropped training strategy, boosted the deep ConvNets decoding performance, reaching at least as good performance as the widely used filter bank common spatial patterns (FBCSP) algorithm (mean decoding accuracies 82.1% FBCSP, 84.0% deep ConvNets). While FBCSP is designed to use spectral power modulations, the features used by ConvNets are not fixed a priori. Our novel methods for visualizing the learned features demonstrated that ConvNets indeed learned to use spectral power modulations in the alpha, beta, and high gamma frequencies, and proved useful for spatially mapping the learned features by revealing the topography of the causal contributions of features in different frequency bands to the decoding decision. Our study thus shows how to design and train ConvNets to decode task‐related information from the raw EEG without handcrafted features and highlights the potential of deep ConvNets combined with advanced visualization techniques for EEG‐based brain mapping. Hum Brain Mapp 38:5391–5420, 2017. © 2017 Wiley Periodicals, Inc. PMID:28782865

Deep learning with convolutional neural networks for EEG decoding and visualization.

PubMed

Schirrmeister, Robin Tibor; Springenberg, Jost Tobias; Fiederer, Lukas Dominique Josef; Glasstetter, Martin; Eggensperger, Katharina; Tangermann, Michael; Hutter, Frank; Burgard, Wolfram; Ball, Tonio

2017-11-01

Deep learning with convolutional neural networks (deep ConvNets) has revolutionized computer vision through end-to-end learning, that is, learning from the raw data. There is increasing interest in using deep ConvNets for end-to-end EEG analysis, but a better understanding of how to design and train ConvNets for end-to-end EEG decoding and how to visualize the informative EEG features the ConvNets learn is still needed. Here, we studied deep ConvNets with a range of different architectures, designed for decoding imagined or executed tasks from raw EEG. Our results show that recent advances from the machine learning field, including batch normalization and exponential linear units, together with a cropped training strategy, boosted the deep ConvNets decoding performance, reaching at least as good performance as the widely used filter bank common spatial patterns (FBCSP) algorithm (mean decoding accuracies 82.1% FBCSP, 84.0% deep ConvNets). While FBCSP is designed to use spectral power modulations, the features used by ConvNets are not fixed a priori. Our novel methods for visualizing the learned features demonstrated that ConvNets indeed learned to use spectral power modulations in the alpha, beta, and high gamma frequencies, and proved useful for spatially mapping the learned features by revealing the topography of the causal contributions of features in different frequency bands to the decoding decision. Our study thus shows how to design and train ConvNets to decode task-related information from the raw EEG without handcrafted features and highlights the potential of deep ConvNets combined with advanced visualization techniques for EEG-based brain mapping. Hum Brain Mapp 38:5391-5420, 2017. © 2017 Wiley Periodicals, Inc. © 2017 The Authors Human Brain Mapping Published by Wiley Periodicals, Inc.

Task- and self-related pathways to deep learning: the mediating role of achievement goals, classroom attentiveness, and group participation.

PubMed

Lau, Shun; Liem, Arief Darmanegara; Nie, Youyan

2008-12-01

The expectancy-value and achievement goal theories are arguably the two most dominant theories of achievement motivation in the contemporary literature. However, very few studies have examined how the constructs derived from both theories are related to deep learning. Moreover, although there is evidence demonstrating the links between achievement goals and deep learning, little research has examined the mediating processes involved. The aims of this research were to: (a) investigate the role of task- and self-related beliefs (task value and self-efficacy) as well as achievement goals in predicting deep learning in mathematics and (b) examine how classroom attentiveness and group participation mediated the relations between achievement goals and deep learning. The sample comprised 1,476 Grade-9 students from 39 schools in Singapore. Students' self-efficacy, task value, achievement goals, classroom attentiveness, group participation, and deep learning in mathematics were assessed by a self-reported questionnaire administered on-line. Structural equation modelling was performed to test the hypothesized model linking these variables. Task value was predictive of task-related achievement goals whereas self-efficacy was predictive of task-approach, performance-approach, and performance-avoidance goals. Achievement goals were found to fully mediate the relations between task value and self-efficacy on the one hand, and classroom attentiveness, group participation, and deep learning on the other. Classroom attentiveness and group participation partially mediated the relations between achievement goal adoption and deep learning. The findings suggest that (a) task- and self-related pathways are two possible routes through which students could be motivated to learn and (b) like task-approach goals, performance-approach goals could lead to adaptive processes and outcomes.

Using Cooperative Structures to Promote Deep Learning

ERIC Educational Resources Information Center

Millis, Barbara J.

2014-01-01

The author explores concrete ways to help students learn more and have fun doing it while they support each other's learning. The article specifically shows the relationships between cooperative learning and deep learning. Readers will become familiar with the tenets of cooperative learning and its power to enhance learning--even more so when…

Enhancing Deep Learning: Lessons from the Introduction of Learning Teams in Management Education in France

ERIC Educational Resources Information Center

Borredon, Liz; Deffayet, Sylvie; Baker, Ann C.; Kolb, David

2011-01-01

Drawing from the reflective teaching and learning practices recommended in influential publications on learning styles, experiential learning, deep learning, and dialogue, the authors tested the concept of "learning teams" in the framework of a leadership program implemented for the first time in a top French management school…

Factors Contributing to Changes in a Deep Approach to Learning in Different Learning Environments

ERIC Educational Resources Information Center

Postareff, Liisa; Parpala, Anna; Lindblom-Ylänne, Sari

2015-01-01

The study explored factors explaining changes in a deep approach to learning. The data consisted of interviews with 12 students from four Bachelor-level courses representing different disciplines. We analysed and compared descriptions of students whose deep approach either increased, decreased or remained relatively unchanged during their courses.…

Deep learning aided decision support for pulmonary nodules diagnosing: a review.

PubMed

Yang, Yixin; Feng, Xiaoyi; Chi, Wenhao; Li, Zhengyang; Duan, Wenzhe; Liu, Haiping; Liang, Wenhua; Wang, Wei; Chen, Ping; He, Jianxing; Liu, Bo

2018-04-01

Deep learning techniques have recently emerged as promising decision supporting approaches to automatically analyze medical images for different clinical diagnosing purposes. Diagnosing of pulmonary nodules by using computer-assisted diagnosing has received considerable theoretical, computational, and empirical research work, and considerable methods have been developed for detection and classification of pulmonary nodules on different formats of images including chest radiographs, computed tomography (CT), and positron emission tomography in the past five decades. The recent remarkable and significant progress in deep learning for pulmonary nodules achieved in both academia and the industry has demonstrated that deep learning techniques seem to be promising alternative decision support schemes to effectively tackle the central issues in pulmonary nodules diagnosing, including feature extraction, nodule detection, false-positive reduction, and benign-malignant classification for the huge volume of chest scan data. The main goal of this investigation is to provide a comprehensive state-of-the-art review of the deep learning aided decision support for pulmonary nodules diagnosing. As far as the authors know, this is the first time that a review is devoted exclusively to deep learning techniques for pulmonary nodules diagnosing.

Determination of 90Sr / 238U ratio by double isotope dilution inductively coupled plasma mass spectrometer with multiple collection in spent nuclear fuel samples with in situ 90Sr / 90Zr separation in a collision-reaction cell

NASA Astrophysics Data System (ADS)

Isnard, H.; Aubert, M.; Blanchet, P.; Brennetot, R.; Chartier, F.; Geertsen, V.; Manuguerra, F.

2006-02-01

Strontium-90 is one of the most important fission products generated in nuclear industry. In the research field concerning nuclear waste disposal in deep geological environment, it is necessary to quantify accurately and precisely its concentration (or the 90Sr / 238U atomic ratio) in irradiated fuels. To obtain accurate analysis of radioactive 90Sr, mass spectrometry associated with isotope dilution is the most appropriated method. But, in nuclear fuel samples the interference with 90Zr must be previously eliminated. An inductively coupled plasma mass spectrometer with multiple collection, equipped with an hexapole collision cell, has been used to eliminate the 90Sr / 90Zr interference by addition of oxygen in the collision cell as a reactant gas. Zr + ions are converted into ZrO +, whereas Sr + ions are not reactive. A mixed solution, prepared from a solution of enriched 84Sr and a solution of enriched 235U was then used to quantify the 90Sr / 238U ratio in spent fuel sample solutions using the double isotope dilution method. This paper shows the results, the reproducibility and the uncertainties that can be obtained with this method to quantify the 90Sr / 238U atomic ratio in an UOX (uranium oxide) and a MOX (mixed oxide) spent fuel samples using the collision cell of an inductively coupled plasma mass spectrometer with multiple collection to perform the 90Sr / 90Zr separation. A comparison with the results obtained by inductively coupled plasma mass spectrometer with multiple collection after a chemical separation of strontium from zirconium using a Sr spec resin (Eichrom) has been performed. Finally, to validate the analytical procedure developed, measurements of the same samples have been performed by thermal ionization mass spectrometry, used as an independent technique, after chemical separation of Sr.

Multiple isotopes (O, C, Li, Sr) as tracers of CO2 and brine leakage from CO2-enhanced oil recovery activities in Permian Basin, Texas, USA

NASA Astrophysics Data System (ADS)

Phan, T. T.; Sharma, S.; Gardiner, J. B.; Thomas, R. B.; Stuckman, M.; Spaulding, R.; Lopano, C. L.; Hakala, A.

2017-12-01

Potential CO2 and brine migration or leakage into shallow groundwater is a critical issue associated with CO2 injection at both enhanced oil recovery (EOR) and carbon sequestration sites. The effectiveness of multiple isotope systems (δ18OH2O, δ13C, δ7Li, 87Sr/86Sr) in monitoring CO2 and brine leakage at a CO2-EOR site located within the Permian basin (Seminole, Texas, USA) was studied. Water samples collected from an oil producing formation (San Andres), a deep groundwater formation (Santa Rosa), and a shallow groundwater aquifer (Ogallala) over a four-year period were analyzed for elemental and isotopic compositions. The absence of any change in δ18OH2O or δ13CDIC values of water in the overlying Ogallala aquifer after CO2 injection indicates that injected CO2 did not leak into this aquifer. The range of Ogallala water δ7Li (13-17‰) overlaps the San Andres water δ7Li (13-15‰) whereas 87Sr/86Sr of Ogallala (0.70792±0.00005) significantly differs from San Andres water (0.70865±0.00003). This observation demonstrates that Sr isotopes are much more sensitive than Li isotopes in tracking brine leakage into shallow groundwater at the studied site. In contrast, deep groundwater δ7Li (21-25‰) is isotopically distinct from San Andres produced water; thus, monitoring this intermitted formation water can provide an early indication of CO2 injection-induced brine migration from the underlying oil producing formation. During water alternating with gas (WAG) operations, a significant shift towards more positive δ13CDIC values was observed in the produced water from several of the San Andres formation wells. The carbon isotope trend suggests that the 13C enriched injected CO2 and formation carbonates became the primary sources of dissolved inorganic carbon in the area surrounding the injection wells. Moreover, one-way ANOVA statistical analysis shows that the differences in δ7Li (F(1,16) = 2.09, p = 0.17) and 87Sr/86Sr (F(1,18) = 4.47, p = 0.05) values of shallow groundwater collected before and during the WAG period are not statistically significant. The results to date suggest that the water chemistry of shallow groundwater has not been influenced by the CO2 injection activities. The efficacy of each isotope system as a monitoring tool will be evaluated and discussed using a Bayesian mixing model.

Learning Business Etiquette at Dinner.

ERIC Educational Resources Information Center

McPherson, William

1998-01-01

Describes helping business students prepare for job interviews and other business events by learning and practicing business etiquette at a dinner set up by business-communication instructors and the university food service. (SR)

Application of Deep Learning in Automated Analysis of Molecular Images in Cancer: A Survey

PubMed Central

Xue, Yong; Chen, Shihui; Liu, Yong

2017-01-01

Molecular imaging enables the visualization and quantitative analysis of the alterations of biological procedures at molecular and/or cellular level, which is of great significance for early detection of cancer. In recent years, deep leaning has been widely used in medical imaging analysis, as it overcomes the limitations of visual assessment and traditional machine learning techniques by extracting hierarchical features with powerful representation capability. Research on cancer molecular images using deep learning techniques is also increasing dynamically. Hence, in this paper, we review the applications of deep learning in molecular imaging in terms of tumor lesion segmentation, tumor classification, and survival prediction. We also outline some future directions in which researchers may develop more powerful deep learning models for better performance in the applications in cancer molecular imaging. PMID:29114182

Retrieval practice as an effective memory strategy in children and adolescents with traumatic brain injury.

PubMed

Coyne, Julia H; Borg, Jacquelyn M; DeLuca, John; Glass, Leslie; Sumowski, James F

2015-04-01

To investigate whether retrieval practice (RP) is a more effective memory strategy than restudy in children and adolescents with traumatic brain injury (TBI). Three × two within-subjects experiment: 3 (learning condition: massed restudy [MR], spaced restudy [SR], retrieval practice [RP]) × 2 (stimulus type: verbal paired associates [VPAs] and face-name pairs [FNPs]). The dependent measure was delayed recall of VPAs and FNPs. Subacute pediatric neurorehabilitation center. Pediatric survivors of TBI (N=15) aged 8 to 16 years with below-average memory. During RP, participants were quizzed on to-be-learned information (VPAs and FNPs) shortly after it was presented, such that they practiced retrieval during the learning phase. MR consisted of repeated restudy (tantamount to cramming). SR consisted of restudy trials separated in time (ie, distributed learning). Delayed recall of 24 VPAs and 24 FNPs after a 25-minute delay. VPAs and FNPs were equally divided across 3 learning conditions (16 per condition). There was a large main effect of learning condition on delayed recall (P<.001; ηp(2)=.84), with better mean recall of VPAs and FNPs studied through RP (6.23±1.39) relative to MR (3.60±1.53; P<.001) and SR (4.77±1.39; P<.001). Moreover, RP was the single best learning strategy for every participant. Memory problems and related academic learning difficulties are common after pediatric TBI. Herein, we identify RP as a promising and simple strategy to support learning and improve memory in children and adolescents with TBI. Our experimental findings were quite robust and set the stage for subsequent randomized controlled trials of RP in pediatric TBI. Copyright © 2015 American Congress of Rehabilitation Medicine. Published by Elsevier Inc. All rights reserved.

Developing Deep Learning Applications for Life Science and Pharma Industry.

PubMed

Siegismund, Daniel; Tolkachev, Vasily; Heyse, Stephan; Sick, Beate; Duerr, Oliver; Steigele, Stephan

2018-06-01

Deep Learning has boosted artificial intelligence over the past 5 years and is seen now as one of the major technological innovation areas, predicted to replace lots of repetitive, but complex tasks of human labor within the next decade. It is also expected to be 'game changing' for research activities in pharma and life sciences, where large sets of similar yet complex data samples are systematically analyzed. Deep learning is currently conquering formerly expert domains especially in areas requiring perception, previously not amenable to standard machine learning. A typical example is the automated analysis of images which are typically produced en-masse in many domains, e. g., in high-content screening or digital pathology. Deep learning enables to create competitive applications in so-far defined core domains of 'human intelligence'. Applications of artificial intelligence have been enabled in recent years by (i) the massive availability of data samples, collected in pharma driven drug programs (='big data') as well as (ii) deep learning algorithmic advancements and (iii) increase in compute power. Such applications are based on software frameworks with specific strengths and weaknesses. Here, we introduce typical applications and underlying frameworks for deep learning with a set of practical criteria for developing production ready solutions in life science and pharma research. Based on our own experience in successfully developing deep learning applications we provide suggestions and a baseline for selecting the most suited frameworks for a future-proof and cost-effective development. © Georg Thieme Verlag KG Stuttgart · New York.

A Deep Learning Approach to on-Node Sensor Data Analytics for Mobile or Wearable Devices.

PubMed

Ravi, Daniele; Wong, Charence; Lo, Benny; Yang, Guang-Zhong

2017-01-01

The increasing popularity of wearable devices in recent years means that a diverse range of physiological and functional data can now be captured continuously for applications in sports, wellbeing, and healthcare. This wealth of information requires efficient methods of classification and analysis where deep learning is a promising technique for large-scale data analytics. While deep learning has been successful in implementations that utilize high-performance computing platforms, its use on low-power wearable devices is limited by resource constraints. In this paper, we propose a deep learning methodology, which combines features learned from inertial sensor data together with complementary information from a set of shallow features to enable accurate and real-time activity classification. The design of this combined method aims to overcome some of the limitations present in a typical deep learning framework where on-node computation is required. To optimize the proposed method for real-time on-node computation, spectral domain preprocessing is used before the data are passed onto the deep learning framework. The classification accuracy of our proposed deep learning approach is evaluated against state-of-the-art methods using both laboratory and real world activity datasets. Our results show the validity of the approach on different human activity datasets, outperforming other methods, including the two methods used within our combined pipeline. We also demonstrate that the computation times for the proposed method are consistent with the constraints of real-time on-node processing on smartphones and a wearable sensor platform.

Lattice-Induced Frequency Shifts in Sr Optical Lattice Clocks at the 10{sup -17} Level

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

Westergaard, P. G.; Lodewyck, J.; Lecallier, A.

2011-05-27

We present a comprehensive study of the frequency shifts associated with the lattice potential in a Sr lattice clock by comparing two such clocks with a frequency stability reaching 5x10{sup -17} after a 1 h integration time. We put the first experimental upper bound on the multipolar M1 and E2 interactions, significantly smaller than the recently predicted theoretical upper limit, and give a 30-fold improved upper limit on the effect of hyperpolarizability. Finally, we report on the first observation of the vector and tensor shifts in a Sr lattice clock. Combining these measurements, we show that all known lattice relatedmore » perturbations will not affect the clock accuracy down to the 10{sup -17} level, even for lattices as deep as 150 recoil energies.« less

Parental social anxiety disorder prospectively predicts toddlers' fear/avoidance in a social referencing paradigm.

PubMed

Aktar, Evin; Majdandžić, Mirjana; de Vente, Wieke; Bögels, Susan M

2014-01-01

Anxiety runs in families. Observational learning of anxious behavior from parents with anxiety disorders plays an important role in the intergenerational transmission of anxiety. We investigated the link between parental anxiety (parental lifetime anxiety disorders and expressed parental anxiety) and toddler fear/avoidance during social referencing (SR) situations. Toddlers (N = 117) participated with both parents (with lifetime social anxiety disorder, other nonsocial anxiety disorders, lifetime comorbid social and other anxiety disorders, or without anxiety disorders) in a longitudinal study. Behavioral inhibition (BI) was measured at 12 months via observational tasks. At 30 months, children were confronted with a stranger and a remote-control robot in SR situations, separately with each parent. Children's fear and avoidance, and parents' expressions of anxiety, encouragement, and overcontrol were observed. Toddlers of parents with lifetime social anxiety disorder (alone and comorbid with other anxiety disorders) showed more fear/avoidance in SR situations than toddlers of parents without anxiety disorders, while the effect of other anxiety disorders alone was not significant. Although expressed parental anxiety at 30 months in SR situations did not significantly predict toddlers' fear/avoidance, higher levels of expressed anxiety at 12 months in SR situations by parents with comorbid social and other anxiety disorders predicted higher levels of fear/avoidance. BI at 12 months predicted toddlers' fear/avoidance only with mothers, but not with fathers. Parental lifetime social anxiety disorders may be a stronger predictor of children's fear/avoidance than parents' expressions of anxiety in SR situations in toddlerhood. End of infancy may be a sensitive time for learning of anxiety from parents with comorbid lifetime social and nonsocial anxiety disorders in SR situations. Fathers are as important as mothers in the transmission of anxiety via SR. Furthermore, children may act relatively free of their early temperament in SR situations with fathers. © 2013 The Authors. Journal of Child Psychology and Psychiatry © 2013 Association for Child and Adolescent Mental Health.

Deep Learning to Predict Falls in Older Adults Based on Daily-Life Trunk Accelerometry.

PubMed

Nait Aicha, Ahmed; Englebienne, Gwenn; van Schooten, Kimberley S; Pijnappels, Mirjam; Kröse, Ben

2018-05-22

Early detection of high fall risk is an essential component of fall prevention in older adults. Wearable sensors can provide valuable insight into daily-life activities; biomechanical features extracted from such inertial data have been shown to be of added value for the assessment of fall risk. Body-worn sensors such as accelerometers can provide valuable insight into fall risk. Currently, biomechanical features derived from accelerometer data are used for the assessment of fall risk. Here, we studied whether deep learning methods from machine learning are suited to automatically derive features from raw accelerometer data that assess fall risk. We used an existing dataset of 296 older adults. We compared the performance of three deep learning model architectures (convolutional neural network (CNN), long short-term memory (LSTM) and a combination of these two (ConvLSTM)) to each other and to a baseline model with biomechanical features on the same dataset. The results show that the deep learning models in a single-task learning mode are strong in recognition of identity of the subject, but that these models only slightly outperform the baseline method on fall risk assessment. When using multi-task learning, with gender and age as auxiliary tasks, deep learning models perform better. We also found that preprocessing of the data resulted in the best performance (AUC = 0.75). We conclude that deep learning models, and in particular multi-task learning, effectively assess fall risk on the basis of wearable sensor data.

Deep Learning to Predict Falls in Older Adults Based on Daily-Life Trunk Accelerometry

PubMed Central

Englebienne, Gwenn; Pijnappels, Mirjam

2018-01-01

Early detection of high fall risk is an essential component of fall prevention in older adults. Wearable sensors can provide valuable insight into daily-life activities; biomechanical features extracted from such inertial data have been shown to be of added value for the assessment of fall risk. Body-worn sensors such as accelerometers can provide valuable insight into fall risk. Currently, biomechanical features derived from accelerometer data are used for the assessment of fall risk. Here, we studied whether deep learning methods from machine learning are suited to automatically derive features from raw accelerometer data that assess fall risk. We used an existing dataset of 296 older adults. We compared the performance of three deep learning model architectures (convolutional neural network (CNN), long short-term memory (LSTM) and a combination of these two (ConvLSTM)) to each other and to a baseline model with biomechanical features on the same dataset. The results show that the deep learning models in a single-task learning mode are strong in recognition of identity of the subject, but that these models only slightly outperform the baseline method on fall risk assessment. When using multi-task learning, with gender and age as auxiliary tasks, deep learning models perform better. We also found that preprocessing of the data resulted in the best performance (AUC = 0.75). We conclude that deep learning models, and in particular multi-task learning, effectively assess fall risk on the basis of wearable sensor data. PMID:29786659

Radiocarbon-based ages and growth rates of bamboo corals from the Gulf of Alaska

NASA Astrophysics Data System (ADS)

Roark, E. Brendan; Guilderson, Thomas P.; Flood-Page, Sarah; Dunbar, Robert B.; Ingram, B. Lynn; Fallon, Stewart J.; McCulloch, Malcolm

2005-02-01

Deep-sea coral communities have long been recognized by fisherman as areas that support large populations of commercial fish. As a consequence, many deep-sea coral communities are threatened by bottom trawling. Successful management and conservation of this widespread deep-sea habitat requires knowledge of the age and growth rates of deep-sea corals. These organisms also contain important archives of intermediate and deep-water variability, and are thus of interest in the context of decadal to century-scale climate dynamics. Here, we present Δ14C data that suggest that bamboo corals from the Gulf of Alaska are long-lived (75-126 years) and that they acquire skeletal carbon from two distinct sources. Independent verification of our growth rate estimates and coral ages is obtained by counting seasonal Sr/Ca cycles and probable lunar cycle growth bands.

Radiocarbon-Based Ages and Growth Rates of Bamboo Corals from the Gulf of Alaska

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

Roark, E B; Guilderson, T P; Flood-Page, S

2004-12-12

Deep-sea coral communities have long been recognized by fisherman as areas that support large populations of commercial fish. As a consequence, many deep-sea coral communities are threatened by bottom trawling. Successful management and conservation of this widespread deep-sea habitat requires knowledge of the age and growth rates of deep-sea corals. These organisms also contain important archives of intermediate and deep-water variability, and are thus of interest in the context of decadal to century-scale climate dynamics. Here, we present {Delta}{sup 14}C data that suggest that bamboo corals from the Gulf of Alaska are long-lived (75-126 years) and that they acquire skeletalmore » carbon from two distinct sources. Independent verification of our growth rate estimates and coral ages is obtained by counting seasonal Sr/Ca cycles and probable lunar cycle growth bands.« less

Preferences for Deep-Surface Learning: A Vocational Education Case Study Using a Multimedia Assessment Activity

ERIC Educational Resources Information Center

Hamm, Simon; Robertson, Ian

2010-01-01

This research tests the proposition that the integration of a multimedia assessment activity into a Diploma of Events Management program promotes a deep learning approach. Firstly, learners' preferences for deep or surface learning were evaluated using the revised two-factor Study Process Questionnaire. Secondly, after completion of an assessment…

Digitally Inspired Thinking: Can Social Media Lead to Deep Learning in Higher Education?

ERIC Educational Resources Information Center

Samuels-Peretz, Debbie; Dvorkin Camiel, Lana; Teeley, Karen; Banerjee, Gouri

2017-01-01

In this study, students from a variety of disciplines, who were enrolled in six courses that incorporate the use of social media, were surveyed to evaluate their perception of how the integration of social-media tools supports deep approaches to learning. Students reported that social media supports deep learning both directly and indirectly,…

Moving beyond the Deep and Surface Dichotomy; Using Q Methodology to Explore Students' Approaches to Studying

ERIC Educational Resources Information Center

Godor, Brian P.

2016-01-01

Student learning approaches research has been built upon the notions of deep and surface learning. Despite its status as part of the educational research canon, the dichotomy of deep/surface has been critiqued as constraining the debate surrounding student learning. Additionally, issues of content validity have been expressed concerning…

Reciprocal Effects of Self-Regulation, Semantic Knowledge, and Reading Comprehension in Early Elementary School

ERIC Educational Resources Information Center

Connor, Carol McDonald; Day, Stephanie L.; Phillips, Beth; Sparapani, Nicole; Ingebrand, Sarah W.; McLean, Leigh; Barrus, Angela; Kaschak, Michael P.

2016-01-01

Many assume that cognitive and linguistic processes, such as semantic knowledge (SK) and self-regulation (SR), subserve learned skills like reading. However, complex models of interacting and bootstrapping effects of SK, SR, instruction, and reading hypothesize reciprocal effects. Testing this "lattice" model with children (n = 852)…

Theory of Self- vs. Externally-Regulated LearningTM: Fundamentals, Evidence, and Applicability

PubMed Central

de la Fuente-Arias, Jesús

2017-01-01

The Theory of Self- vs. Externally-Regulated LearningTM has integrated the variables of SRL theory, the DEDEPRO model, and the 3P model. This new Theory has proposed: (a) in general, the importance of the cyclical model of individual self-regulation (SR) and of external regulation stemming from the context (ER), as two different and complementary variables, both in combination and in interaction; (b) specifically, in the teaching-learning context, the relevance of different types of combinations between levels of self-regulation (SR) and of external regulation (ER) in the prediction of self-regulated learning (SRL), and of cognitive-emotional achievement. This review analyzes the assumptions, conceptual elements, empirical evidence, benefits and limitations of SRL vs. ERL Theory. Finally, professional fields of application and future lines of research are suggested. PMID:29033872

White blood cells identification system based on convolutional deep neural learning networks.

PubMed

Shahin, A I; Guo, Yanhui; Amin, K M; Sharawi, Amr A

2017-11-16

White blood cells (WBCs) differential counting yields valued information about human health and disease. The current developed automated cell morphology equipments perform differential count which is based on blood smear image analysis. Previous identification systems for WBCs consist of successive dependent stages; pre-processing, segmentation, feature extraction, feature selection, and classification. There is a real need to employ deep learning methodologies so that the performance of previous WBCs identification systems can be increased. Classifying small limited datasets through deep learning systems is a major challenge and should be investigated. In this paper, we propose a novel identification system for WBCs based on deep convolutional neural networks. Two methodologies based on transfer learning are followed: transfer learning based on deep activation features and fine-tuning of existed deep networks. Deep acrivation featues are extracted from several pre-trained networks and employed in a traditional identification system. Moreover, a novel end-to-end convolutional deep architecture called "WBCsNet" is proposed and built from scratch. Finally, a limited balanced WBCs dataset classification is performed through the WBCsNet as a pre-trained network. During our experiments, three different public WBCs datasets (2551 images) have been used which contain 5 healthy WBCs types. The overall system accuracy achieved by the proposed WBCsNet is (96.1%) which is more than different transfer learning approaches or even the previous traditional identification system. We also present features visualization for the WBCsNet activation which reflects higher response than the pre-trained activated one. a novel WBCs identification system based on deep learning theory is proposed and a high performance WBCsNet can be employed as a pre-trained network. Copyright © 2017. Published by Elsevier B.V.

Survey on deep learning for radiotherapy.

PubMed

Meyer, Philippe; Noblet, Vincent; Mazzara, Christophe; Lallement, Alex

2018-07-01

More than 50% of cancer patients are treated with radiotherapy, either exclusively or in combination with other methods. The planning and delivery of radiotherapy treatment is a complex process, but can now be greatly facilitated by artificial intelligence technology. Deep learning is the fastest-growing field in artificial intelligence and has been successfully used in recent years in many domains, including medicine. In this article, we first explain the concept of deep learning, addressing it in the broader context of machine learning. The most common network architectures are presented, with a more specific focus on convolutional neural networks. We then present a review of the published works on deep learning methods that can be applied to radiotherapy, which are classified into seven categories related to the patient workflow, and can provide some insights of potential future applications. We have attempted to make this paper accessible to both radiotherapy and deep learning communities, and hope that it will inspire new collaborations between these two communities to develop dedicated radiotherapy applications. Copyright © 2018 Elsevier Ltd. All rights reserved.

[Advantages and Application Prospects of Deep Learning in Image Recognition and Bone Age Assessment].

PubMed

Hu, T H; Wan, L; Liu, T A; Wang, M W; Chen, T; Wang, Y H

2017-12-01

Deep learning and neural network models have been new research directions and hot issues in the fields of machine learning and artificial intelligence in recent years. Deep learning has made a breakthrough in the applications of image and speech recognitions, and also has been extensively used in the fields of face recognition and information retrieval because of its special superiority. Bone X-ray images express different variations in black-white-gray gradations, which have image features of black and white contrasts and level differences. Based on these advantages of deep learning in image recognition, we combine it with the research of bone age assessment to provide basic datum for constructing a forensic automatic system of bone age assessment. This paper reviews the basic concept and network architectures of deep learning, and describes its recent research progress on image recognition in different research fields at home and abroad, and explores its advantages and application prospects in bone age assessment. Copyright© by the Editorial Department of Journal of Forensic Medicine.

Atomically precise metal nanoclusters: stable sizes and optical properties

NASA Astrophysics Data System (ADS)

Jin, Rongchao

2015-01-01

Controlling nanoparticles with atomic precision has long been a major dream of nanochemists. Breakthroughs have been made in the case of gold nanoparticles, at least for nanoparticles smaller than ~3 nm in diameter. Such ultrasmall gold nanoparticles indeed exhibit fundamentally different properties from those of the plasmonic counterparts owing to the quantum size effects as well as the extremely high surface-to-volume ratio. These unique nanoparticles are often called nanoclusters to distinguish them from conventional plasmonic nanoparticles. Intense work carried out in the last few years has generated a library of stable sizes (or stable stoichiometries) of atomically precise gold nanoclusters, which are opening up new exciting opportunities for both fundamental research and technological applications. In this review, we have summarized the recent progress in the research of thiolate (SR)-protected gold nanoclusters with a focus on the reported stable sizes and their optical absorption spectra. The crystallization of nanoclusters still remains challenging; nevertheless, a few more structures have been achieved since the earlier successes in Au102(SR)44, Au25(SR)18 and Au38(SR)24 nanoclusters, and the newly reported structures include Au20(SR)16, Au24(SR)20, Au28(SR)20, Au30S(SR)18, and Au36(SR)24. Phosphine-protected gold and thiolate-protected silver nanoclusters are also briefly discussed in this review. The reported gold nanocluster sizes serve as the basis for investigating their size dependent properties as well as the development of applications in catalysis, sensing, biological labelling, optics, etc. Future efforts will continue to address what stable sizes are existent, and more importantly, what factors determine their stability. Structural determination and theoretical simulations will help to gain deep insight into the structure-property relationships.

Super-Resolution Person Re-Identification With Semi-Coupled Low-Rank Discriminant Dictionary Learning.

PubMed

Jing, Xiao-Yuan; Zhu, Xiaoke; Wu, Fei; Hu, Ruimin; You, Xinge; Wang, Yunhong; Feng, Hui; Yang, Jing-Yu

2017-03-01

Person re-identification has been widely studied due to its importance in surveillance and forensics applications. In practice, gallery images are high resolution (HR), while probe images are usually low resolution (LR) in the identification scenarios with large variation of illumination, weather, or quality of cameras. Person re-identification in this kind of scenarios, which we call super-resolution (SR) person re-identification, has not been well studied. In this paper, we propose a semi-coupled low-rank discriminant dictionary learning (SLD 2 L) approach for SR person re-identification task. With the HR and LR dictionary pair and mapping matrices learned from the features of HR and LR training images, SLD 2 L can convert the features of the LR probe images into HR features. To ensure that the converted features have favorable discriminative capability and the learned dictionaries can well characterize intrinsic feature spaces of the HR and LR images, we design a discriminant term and a low-rank regularization term for SLD 2 L. Moreover, considering that low resolution results in different degrees of loss for different types of visual appearance features, we propose a multi-view SLD 2 L (MVSLD 2 L) approach, which can learn the type-specific dictionary pair and mappings for each type of feature. Experimental results on multiple publicly available data sets demonstrate the effectiveness of our proposed approaches for the SR person re-identification task.

The evolution of sustainable remediation in Australia and New Zealand: A storyline.

PubMed

Smith, Garry; Nadebaum, Peter

2016-12-15

This article describes the 'storyline' of the early and recent growth of sustainable remediation (SR) practice in Australia and New Zealand (ANZ), in order to inform and support other SR stakeholders, and to identify some lessons learned. Achievement of full acceptance and consistency across relevant ANZ regulatory jurisdictions and industry sectors will take time and will require publication of successful examples of SR application. The article describes the respective policy and regulatory contexts for sustainable remediation practice in Australia and in New Zealand; several milestone activities and events in the growth of SR in ANZ; and example SR methodologies and policies produced by stakeholders and remediation practitioners including the Sustainable Remediation Forum of Australia and New Zealand (SuRF ANZ). Copyright © 2016 Elsevier Ltd. All rights reserved.

Machine Learning Based Single-Frame Super-Resolution Processing for Lensless Blood Cell Counting

PubMed Central

Huang, Xiwei; Jiang, Yu; Liu, Xu; Xu, Hang; Han, Zhi; Rong, Hailong; Yang, Haiping; Yan, Mei; Yu, Hao

2016-01-01

A lensless blood cell counting system integrating microfluidic channel and a complementary metal oxide semiconductor (CMOS) image sensor is a promising technique to miniaturize the conventional optical lens based imaging system for point-of-care testing (POCT). However, such a system has limited resolution, making it imperative to improve resolution from the system-level using super-resolution (SR) processing. Yet, how to improve resolution towards better cell detection and recognition with low cost of processing resources and without degrading system throughput is still a challenge. In this article, two machine learning based single-frame SR processing types are proposed and compared for lensless blood cell counting, namely the Extreme Learning Machine based SR (ELMSR) and Convolutional Neural Network based SR (CNNSR). Moreover, lensless blood cell counting prototypes using commercial CMOS image sensors and custom designed backside-illuminated CMOS image sensors are demonstrated with ELMSR and CNNSR. When one captured low-resolution lensless cell image is input, an improved high-resolution cell image will be output. The experimental results show that the cell resolution is improved by 4×, and CNNSR has 9.5% improvement over the ELMSR on resolution enhancing performance. The cell counting results also match well with a commercial flow cytometer. Such ELMSR and CNNSR therefore have the potential for efficient resolution improvement in lensless blood cell counting systems towards POCT applications. PMID:27827837

Deep learning for healthcare applications based on physiological signals: A review.

PubMed

Faust, Oliver; Hagiwara, Yuki; Hong, Tan Jen; Lih, Oh Shu; Acharya, U Rajendra

2018-07-01

We have cast the net into the ocean of knowledge to retrieve the latest scientific research on deep learning methods for physiological signals. We found 53 research papers on this topic, published from 01.01.2008 to 31.12.2017. An initial bibliometric analysis shows that the reviewed papers focused on Electromyogram(EMG), Electroencephalogram(EEG), Electrocardiogram(ECG), and Electrooculogram(EOG). These four categories were used to structure the subsequent content review. During the content review, we understood that deep learning performs better for big and varied datasets than classic analysis and machine classification methods. Deep learning algorithms try to develop the model by using all the available input. This review paper depicts the application of various deep learning algorithms used till recently, but in future it will be used for more healthcare areas to improve the quality of diagnosis. Copyright © 2018 Elsevier B.V. All rights reserved.

Applications of Deep Learning in Biomedicine.

PubMed

Mamoshina, Polina; Vieira, Armando; Putin, Evgeny; Zhavoronkov, Alex

2016-05-02

Increases in throughput and installed base of biomedical research equipment led to a massive accumulation of -omics data known to be highly variable, high-dimensional, and sourced from multiple often incompatible data platforms. While this data may be useful for biomarker identification and drug discovery, the bulk of it remains underutilized. Deep neural networks (DNNs) are efficient algorithms based on the use of compositional layers of neurons, with advantages well matched to the challenges -omics data presents. While achieving state-of-the-art results and even surpassing human accuracy in many challenging tasks, the adoption of deep learning in biomedicine has been comparatively slow. Here, we discuss key features of deep learning that may give this approach an edge over other machine learning methods. We then consider limitations and review a number of applications of deep learning in biomedical studies demonstrating proof of concept and practical utility.

Assessing the Linguistic Productivity of Unsupervised Deep Neural Networks

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

Phillips, Lawrence A.; Hodas, Nathan O.

Increasingly, cognitive scientists have demonstrated interest in applying tools from deep learning. One use for deep learning is in language acquisition where it is useful to know if a linguistic phenomenon can be learned through domain-general means. To assess whether unsupervised deep learning is appropriate, we first pose a smaller question: Can unsupervised neural networks apply linguistic rules productively, using them in novel situations. We draw from the literature on determiner/noun productivity by training an unsupervised, autoencoder network measuring its ability to combine nouns with determiners. Our simple autoencoder creates combinations it has not previously encountered, displaying a degree ofmore » overlap similar to actual children. While this preliminary work does not provide conclusive evidence for productivity, it warrants further investigation with more complex models. Further, this work helps lay the foundations for future collaboration between the deep learning and cognitive science communities.« less

Teaching for Deep Learning

ERIC Educational Resources Information Center

Smith, Tracy Wilson; Colby, Susan A.

2007-01-01

The authors have been engaged in research focused on students' depth of learning as well as teachers' efforts to foster deep learning. Findings from a study examining the teaching practices and student learning outcomes of sixty-four teachers in seventeen different states (Smith et al. 2005) indicated that most of the learning in these classrooms…

Stimulating Deep Learning Using Active Learning Techniques

ERIC Educational Resources Information Center

Yew, Tee Meng; Dawood, Fauziah K. P.; a/p S. Narayansany, Kannaki; a/p Palaniappa Manickam, M. Kamala; Jen, Leong Siok; Hoay, Kuan Chin

2016-01-01

When students and teachers behave in ways that reinforce learning as a spectator sport, the result can often be a classroom and overall learning environment that is mostly limited to transmission of information and rote learning rather than deep approaches towards meaningful construction and application of knowledge. A group of college instructors…

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.

Rapid Mantle Source Variations During the Latest Episode of Kilauea's Prolonged Pu'u O'o Eruption, Hawaii

NASA Astrophysics Data System (ADS)

Marske, J. P.; Garcia, M. O.; Pietruszka, A. J.; Norman, M. D.; Rhodes, J. M.

2006-12-01

Nearly 24 years of continuous geochemical monitoring of lavas from the current Pu'u O'o eruption allow us to probe the mantle processes beneath Kilauea Volcano in unparalleled detail. Here we present new measurements Pb, Sr, and Nd isotope ratios and major- and trace-element abundances for lavas from episode 55 (1997-2006), which marks the longest and most voluminous interval of this eruption. Pu'u O'o lavas erupted since 1985 display systematic decreases in their TiO2, K2O, P2O5 and CaO abundances (normalized to 10 wt. % MgO to correct for olivine control) due to changes in the parental magma composition. Incompatible element ratios (e.g., Ba/Nb and La/Y) also show overall temporal decreases. Earlier erupted Pu'u O'o lavas displayed the most significant decrease in incompatible element ratios with near constant SiO2 contents, and a gradual increase in 87Sr/86Sr ratios. However, episode 55 lavas record significant increases in MgO- normalized SiO2 contents and 87Sr/86Sr with nearly constant (e.g. Ba/Nb) or a slightly reversed (e.g., TiO2 and K2O) trends in incompatible element ratios and abundances. There is little variation of 206Pb/204Pb ratios in lavas (18.38-18.43) erupted since 1985. Neither a single mantle source composition nor a change in partial melting conditions alone can explain these observations. Based on the isotopic and chemical variability, we conclude that early Pu'u O'o lavas originated from two distinct mantle source components: (1) a long-term depleted component (with relatively low 87Sr/86Sr ratios) that originated within the deep source of the Hawaiian plume that characterizes the earlier part of the eruption (1985-1992), and (2) a recently depleted component (i.e. a component that was recently depleted by prior melting) with low abundances of incompatible elements became increasingly important from 1992-1997. More recently, Pu'u O'o has tapped greater proportions of a new (3) long-term less depleted component (with higher 87Sr/86Sr ratios than observed from 1985-1992) that originated within the deep source region of the plume. This third component lies within typical Pb, Sr and Nd isotopic space for Kilauea, but represents a new source composition for the Pu'u O'o eruption. The systematic geochemical evolution of Pu'u O'o lavas reflects changes in the proportions of the mantle source components tapped throughout the eruption. The rapid isotope variations (on a time scale of years) in the most recent lavas suggest the mantle source components are heterogeneous on an extremely small scale, relative to the size of Kilauea's melting region.

A deep learning framework for financial time series using stacked autoencoders and long-short term memory.

PubMed

Bao, Wei; Yue, Jun; Rao, Yulei

2017-01-01

The application of deep learning approaches to finance has received a great deal of attention from both investors and researchers. This study presents a novel deep learning framework where wavelet transforms (WT), stacked autoencoders (SAEs) and long-short term memory (LSTM) are combined for stock price forecasting. The SAEs for hierarchically extracted deep features is introduced into stock price forecasting for the first time. The deep learning framework comprises three stages. First, the stock price time series is decomposed by WT to eliminate noise. Second, SAEs is applied to generate deep high-level features for predicting the stock price. Third, high-level denoising features are fed into LSTM to forecast the next day's closing price. Six market indices and their corresponding index futures are chosen to examine the performance of the proposed model. Results show that the proposed model outperforms other similar models in both predictive accuracy and profitability performance.

Comparison of Deep Learning With Multiple Machine Learning Methods and Metrics Using Diverse Drug Discovery Data Sets.

PubMed

Korotcov, Alexandru; Tkachenko, Valery; Russo, Daniel P; Ekins, Sean

2017-12-04

Machine learning methods have been applied to many data sets in pharmaceutical research for several decades. The relative ease and availability of fingerprint type molecular descriptors paired with Bayesian methods resulted in the widespread use of this approach for a diverse array of end points relevant to drug discovery. Deep learning is the latest machine learning algorithm attracting attention for many of pharmaceutical applications from docking to virtual screening. Deep learning is based on an artificial neural network with multiple hidden layers and has found considerable traction for many artificial intelligence applications. We have previously suggested the need for a comparison of different machine learning methods with deep learning across an array of varying data sets that is applicable to pharmaceutical research. End points relevant to pharmaceutical research include absorption, distribution, metabolism, excretion, and toxicity (ADME/Tox) properties, as well as activity against pathogens and drug discovery data sets. In this study, we have used data sets for solubility, probe-likeness, hERG, KCNQ1, bubonic plague, Chagas, tuberculosis, and malaria to compare different machine learning methods using FCFP6 fingerprints. These data sets represent whole cell screens, individual proteins, physicochemical properties as well as a data set with a complex end point. Our aim was to assess whether deep learning offered any improvement in testing when assessed using an array of metrics including AUC, F1 score, Cohen's kappa, Matthews correlation coefficient and others. Based on ranked normalized scores for the metrics or data sets Deep Neural Networks (DNN) ranked higher than SVM, which in turn was ranked higher than all the other machine learning methods. Visualizing these properties for training and test sets using radar type plots indicates when models are inferior or perhaps over trained. These results also suggest the need for assessing deep learning further using multiple metrics with much larger scale comparisons, prospective testing as well as assessment of different fingerprints and DNN architectures beyond those used.

From contraction to gene expression: nanojunctions of the sarco/endoplasmic reticulum deliver site- and function-specific calcium signals.

PubMed

Evans, A Mark; Fameli, Nicola; Ogunbayo, Oluseye A; Duan, Jingxian; Navarro-Dorado, Jorge

2016-08-01

Calcium signals determine, for example, smooth muscle contraction and changes in gene expression. How calcium signals select for these processes is enigmatic. We build on the "panjunctional sarcoplasmic reticulum" hypothesis, describing our view that different calcium pumps and release channels, with different kinetics and affinities for calcium, are strategically positioned within nanojunctions of the SR and help demarcate their respective cytoplasmic nanodomains. SERCA2b and RyR1 are preferentially targeted to the sarcoplasmic reticulum (SR) proximal to the plasma membrane (PM), i.e., to the superficial buffer barrier formed by PM-SR nanojunctions, and support vasodilation. In marked contrast, SERCA2a may be entirely restricted to the deep, perinuclear SR and may supply calcium to this sub-compartment in support of vasoconstriction. RyR3 is also preferentially targeted to the perinuclear SR, where its clusters associate with lysosome-SR nanojunctions. The distribution of RyR2 is more widespread and extends from this region to the wider cell. Therefore, perinuclear RyR3s most likely support the initiation of global calcium waves at L-SR junctions, which subsequently propagate by calcium-induced calcium release via RyR2 in order to elicit contraction. Data also suggest that unique SERCA and RyR are preferentially targeted to invaginations of the nuclear membrane. Site- and function-specific calcium signals may thus arise to modulate stimulus-response coupling and transcriptional cascades.

A Deep Learning Network Approach to ab initio Protein Secondary Structure Prediction

PubMed Central

Spencer, Matt; Eickholt, Jesse; Cheng, Jianlin

2014-01-01

Ab initio protein secondary structure (SS) predictions are utilized to generate tertiary structure predictions, which are increasingly demanded due to the rapid discovery of proteins. Although recent developments have slightly exceeded previous methods of SS prediction, accuracy has stagnated around 80% and many wonder if prediction cannot be advanced beyond this ceiling. Disciplines that have traditionally employed neural networks are experimenting with novel deep learning techniques in attempts to stimulate progress. Since neural networks have historically played an important role in SS prediction, we wanted to determine whether deep learning could contribute to the advancement of this field as well. We developed an SS predictor that makes use of the position-specific scoring matrix generated by PSI-BLAST and deep learning network architectures, which we call DNSS. Graphical processing units and CUDA software optimize the deep network architecture and efficiently train the deep networks. Optimal parameters for the training process were determined, and a workflow comprising three separately trained deep networks was constructed in order to make refined predictions. This deep learning network approach was used to predict SS for a fully independent test data set of 198 proteins, achieving a Q3 accuracy of 80.7% and a Sov accuracy of 74.2%. PMID:25750595

A Deep Learning Network Approach to ab initio Protein Secondary Structure Prediction.

PubMed

Spencer, Matt; Eickholt, Jesse; Jianlin Cheng

2015-01-01

Ab initio protein secondary structure (SS) predictions are utilized to generate tertiary structure predictions, which are increasingly demanded due to the rapid discovery of proteins. Although recent developments have slightly exceeded previous methods of SS prediction, accuracy has stagnated around 80 percent and many wonder if prediction cannot be advanced beyond this ceiling. Disciplines that have traditionally employed neural networks are experimenting with novel deep learning techniques in attempts to stimulate progress. Since neural networks have historically played an important role in SS prediction, we wanted to determine whether deep learning could contribute to the advancement of this field as well. We developed an SS predictor that makes use of the position-specific scoring matrix generated by PSI-BLAST and deep learning network architectures, which we call DNSS. Graphical processing units and CUDA software optimize the deep network architecture and efficiently train the deep networks. Optimal parameters for the training process were determined, and a workflow comprising three separately trained deep networks was constructed in order to make refined predictions. This deep learning network approach was used to predict SS for a fully independent test dataset of 198 proteins, achieving a Q3 accuracy of 80.7 percent and a Sov accuracy of 74.2 percent.

DeepNeuron: an open deep learning toolbox for neuron tracing.

PubMed

Zhou, Zhi; Kuo, Hsien-Chi; Peng, Hanchuan; Long, Fuhui

2018-06-06

Reconstructing three-dimensional (3D) morphology of neurons is essential for understanding brain structures and functions. Over the past decades, a number of neuron tracing tools including manual, semiautomatic, and fully automatic approaches have been developed to extract and analyze 3D neuronal structures. Nevertheless, most of them were developed based on coding certain rules to extract and connect structural components of a neuron, showing limited performance on complicated neuron morphology. Recently, deep learning outperforms many other machine learning methods in a wide range of image analysis and computer vision tasks. Here we developed a new Open Source toolbox, DeepNeuron, which uses deep learning networks to learn features and rules from data and trace neuron morphology in light microscopy images. DeepNeuron provides a family of modules to solve basic yet challenging problems in neuron tracing. These problems include but not limited to: (1) detecting neuron signal under different image conditions, (2) connecting neuronal signals into tree(s), (3) pruning and refining tree morphology, (4) quantifying the quality of morphology, and (5) classifying dendrites and axons in real time. We have tested DeepNeuron using light microscopy images including bright-field and confocal images of human and mouse brain, on which DeepNeuron demonstrates robustness and accuracy in neuron tracing.

Clinical Named Entity Recognition Using Deep Learning Models.

PubMed

Wu, Yonghui; Jiang, Min; Xu, Jun; Zhi, Degui; Xu, Hua

2017-01-01

Clinical Named Entity Recognition (NER) is a critical natural language processing (NLP) task to extract important concepts (named entities) from clinical narratives. Researchers have extensively investigated machine learning models for clinical NER. Recently, there have been increasing efforts to apply deep learning models to improve the performance of current clinical NER systems. This study examined two popular deep learning architectures, the Convolutional Neural Network (CNN) and the Recurrent Neural Network (RNN), to extract concepts from clinical texts. We compared the two deep neural network architectures with three baseline Conditional Random Fields (CRFs) models and two state-of-the-art clinical NER systems using the i2b2 2010 clinical concept extraction corpus. The evaluation results showed that the RNN model trained with the word embeddings achieved a new state-of-the- art performance (a strict F1 score of 85.94%) for the defined clinical NER task, outperforming the best-reported system that used both manually defined and unsupervised learning features. This study demonstrates the advantage of using deep neural network architectures for clinical concept extraction, including distributed feature representation, automatic feature learning, and long-term dependencies capture. This is one of the first studies to compare the two widely used deep learning models and demonstrate the superior performance of the RNN model for clinical NER.

Clinical Named Entity Recognition Using Deep Learning Models

PubMed Central

Wu, Yonghui; Jiang, Min; Xu, Jun; Zhi, Degui; Xu, Hua

2017-01-01

Clinical Named Entity Recognition (NER) is a critical natural language processing (NLP) task to extract important concepts (named entities) from clinical narratives. Researchers have extensively investigated machine learning models for clinical NER. Recently, there have been increasing efforts to apply deep learning models to improve the performance of current clinical NER systems. This study examined two popular deep learning architectures, the Convolutional Neural Network (CNN) and the Recurrent Neural Network (RNN), to extract concepts from clinical texts. We compared the two deep neural network architectures with three baseline Conditional Random Fields (CRFs) models and two state-of-the-art clinical NER systems using the i2b2 2010 clinical concept extraction corpus. The evaluation results showed that the RNN model trained with the word embeddings achieved a new state-of-the- art performance (a strict F1 score of 85.94%) for the defined clinical NER task, outperforming the best-reported system that used both manually defined and unsupervised learning features. This study demonstrates the advantage of using deep neural network architectures for clinical concept extraction, including distributed feature representation, automatic feature learning, and long-term dependencies capture. This is one of the first studies to compare the two widely used deep learning models and demonstrate the superior performance of the RNN model for clinical NER. PMID:29854252

Deep learning in pharmacogenomics: from gene regulation to patient stratification.

PubMed

Kalinin, Alexandr A; Higgins, Gerald A; Reamaroon, Narathip; Soroushmehr, Sayedmohammadreza; Allyn-Feuer, Ari; Dinov, Ivo D; Najarian, Kayvan; Athey, Brian D

2018-05-01

This Perspective provides examples of current and future applications of deep learning in pharmacogenomics, including: identification of novel regulatory variants located in noncoding domains of the genome and their function as applied to pharmacoepigenomics; patient stratification from medical records; and the mechanistic prediction of drug response, targets and their interactions. Deep learning encapsulates a family of machine learning algorithms that has transformed many important subfields of artificial intelligence over the last decade, and has demonstrated breakthrough performance improvements on a wide range of tasks in biomedicine. We anticipate that in the future, deep learning will be widely used to predict personalized drug response and optimize medication selection and dosing, using knowledge extracted from large and complex molecular, epidemiological, clinical and demographic datasets.

Deep Learning Neural Networks and Bayesian Neural Networks in Data Analysis

NASA Astrophysics Data System (ADS)

Chernoded, Andrey; Dudko, Lev; Myagkov, Igor; Volkov, Petr

2017-10-01

Most of the modern analyses in high energy physics use signal-versus-background classification techniques of machine learning methods and neural networks in particular. Deep learning neural network is the most promising modern technique to separate signal and background and now days can be widely and successfully implemented as a part of physical analysis. In this article we compare Deep learning and Bayesian neural networks application as a classifiers in an instance of top quark analysis.

BCDForest: a boosting cascade deep forest model towards the classification of cancer subtypes based on gene expression data.

PubMed

Guo, Yang; Liu, Shuhui; Li, Zhanhuai; Shang, Xuequn

2018-04-11

The classification of cancer subtypes is of great importance to cancer disease diagnosis and therapy. Many supervised learning approaches have been applied to cancer subtype classification in the past few years, especially of deep learning based approaches. Recently, the deep forest model has been proposed as an alternative of deep neural networks to learn hyper-representations by using cascade ensemble decision trees. It has been proved that the deep forest model has competitive or even better performance than deep neural networks in some extent. However, the standard deep forest model may face overfitting and ensemble diversity challenges when dealing with small sample size and high-dimensional biology data. In this paper, we propose a deep learning model, so-called BCDForest, to address cancer subtype classification on small-scale biology datasets, which can be viewed as a modification of the standard deep forest model. The BCDForest distinguishes from the standard deep forest model with the following two main contributions: First, a named multi-class-grained scanning method is proposed to train multiple binary classifiers to encourage diversity of ensemble. Meanwhile, the fitting quality of each classifier is considered in representation learning. Second, we propose a boosting strategy to emphasize more important features in cascade forests, thus to propagate the benefits of discriminative features among cascade layers to improve the classification performance. Systematic comparison experiments on both microarray and RNA-Seq gene expression datasets demonstrate that our method consistently outperforms the state-of-the-art methods in application of cancer subtype classification. The multi-class-grained scanning and boosting strategy in our model provide an effective solution to ease the overfitting challenge and improve the robustness of deep forest model working on small-scale data. Our model provides a useful approach to the classification of cancer subtypes by using deep learning on high-dimensional and small-scale biology data.

The Influence of Parents and Teachers on the Deep Learning Approach of Pupils in Norwegian Upper-Secondary Schools

ERIC Educational Resources Information Center

Elstad, Eyvind; Christophersen, Knut-Andreas; Turmo, Are

2012-01-01

Introduction: The purpose of this article was to explore the influence of parents and teachers on the deep learning approach of pupils by estimating the strength of the relationships between these factors and the motivation, volition and deep learning approach of Norwegian 16-year-olds. Method: Structural equation modeling for cross-sectional…

The Use of Deep Learning Strategies in Online Business Courses to Impact Student Retention

ERIC Educational Resources Information Center

DeLotell, Pam Jones; Millam, Loretta A.; Reinhardt, Michelle M.

2010-01-01

Interest, application and understanding--these are key elements in successful online classroom experiences and all part of what is commonly referred to as deep learning. Deep learning occurs when students are able to connect with course topics, find value in them and see how to apply them to real-world situations. Asynchronous discussion forums in…

Deep Unfolding for Topic Models.

PubMed

Chien, Jen-Tzung; Lee, Chao-Hsi

2018-02-01

Deep unfolding provides an approach to integrate the probabilistic generative models and the deterministic neural networks. Such an approach is benefited by deep representation, easy interpretation, flexible learning and stochastic modeling. This study develops the unsupervised and supervised learning of deep unfolded topic models for document representation and classification. Conventionally, the unsupervised and supervised topic models are inferred via the variational inference algorithm where the model parameters are estimated by maximizing the lower bound of logarithm of marginal likelihood using input documents without and with class labels, respectively. The representation capability or classification accuracy is constrained by the variational lower bound and the tied model parameters across inference procedure. This paper aims to relax these constraints by directly maximizing the end performance criterion and continuously untying the parameters in learning process via deep unfolding inference (DUI). The inference procedure is treated as the layer-wise learning in a deep neural network. The end performance is iteratively improved by using the estimated topic parameters according to the exponentiated updates. Deep learning of topic models is therefore implemented through a back-propagation procedure. Experimental results show the merits of DUI with increasing number of layers compared with variational inference in unsupervised as well as supervised topic models.

Prediction of residue-residue contact matrix for protein-protein interaction with Fisher score features and deep learning.

PubMed

Du, Tianchuan; Liao, Li; Wu, Cathy H; Sun, Bilin

2016-11-01

Protein-protein interactions play essential roles in many biological processes. Acquiring knowledge of the residue-residue contact information of two interacting proteins is not only helpful in annotating functions for proteins, but also critical for structure-based drug design. The prediction of the protein residue-residue contact matrix of the interfacial regions is challenging. In this work, we introduced deep learning techniques (specifically, stacked autoencoders) to build deep neural network models to tackled the residue-residue contact prediction problem. In tandem with interaction profile Hidden Markov Models, which was used first to extract Fisher score features from protein sequences, stacked autoencoders were deployed to extract and learn hidden abstract features. The deep learning model showed significant improvement over the traditional machine learning model, Support Vector Machines (SVM), with the overall accuracy increased by 15% from 65.40% to 80.82%. We showed that the stacked autoencoders could extract novel features, which can be utilized by deep neural networks and other classifiers to enhance learning, out of the Fisher score features. It is further shown that deep neural networks have significant advantages over SVM in making use of the newly extracted features. Copyright © 2016. Published by Elsevier Inc.

Using deep learning in image hyper spectral segmentation, classification, and detection

NASA Astrophysics Data System (ADS)

Zhao, Xiuying; Su, Zhenyu

2018-02-01

Recent years have shown that deep learning neural networks are a valuable tool in the field of computer vision. Deep learning method can be used in applications like remote sensing such as Land cover Classification, Detection of Vehicle in Satellite Images, Hyper spectral Image classification. This paper addresses the use of the deep learning artificial neural network in Satellite image segmentation. Image segmentation plays an important role in image processing. The hue of the remote sensing image often has a large hue difference, which will result in the poor display of the images in the VR environment. Image segmentation is a pre processing technique applied to the original images and splits the image into many parts which have different hue to unify the color. Several computational models based on supervised, unsupervised, parametric, probabilistic region based image segmentation techniques have been proposed. Recently, one of the machine learning technique known as, deep learning with convolution neural network has been widely used for development of efficient and automatic image segmentation models. In this paper, we focus on study of deep neural convolution network and its variants for automatic image segmentation rather than traditional image segmentation strategies.

Deep learning aided decision support for pulmonary nodules diagnosing: a review

PubMed Central

Yang, Yixin; Feng, Xiaoyi; Chi, Wenhao; Li, Zhengyang; Duan, Wenzhe; Liu, Haiping; Liang, Wenhua; Wang, Wei; Chen, Ping

2018-01-01

Deep learning techniques have recently emerged as promising decision supporting approaches to automatically analyze medical images for different clinical diagnosing purposes. Diagnosing of pulmonary nodules by using computer-assisted diagnosing has received considerable theoretical, computational, and empirical research work, and considerable methods have been developed for detection and classification of pulmonary nodules on different formats of images including chest radiographs, computed tomography (CT), and positron emission tomography in the past five decades. The recent remarkable and significant progress in deep learning for pulmonary nodules achieved in both academia and the industry has demonstrated that deep learning techniques seem to be promising alternative decision support schemes to effectively tackle the central issues in pulmonary nodules diagnosing, including feature extraction, nodule detection, false-positive reduction, and benign-malignant classification for the huge volume of chest scan data. The main goal of this investigation is to provide a comprehensive state-of-the-art review of the deep learning aided decision support for pulmonary nodules diagnosing. As far as the authors know, this is the first time that a review is devoted exclusively to deep learning techniques for pulmonary nodules diagnosing. PMID:29780633

Deep Learning for Prediction of Obstructive Disease From Fast Myocardial Perfusion SPECT: A Multicenter Study.

PubMed

Betancur, Julian; Commandeur, Frederic; Motlagh, Mahsaw; Sharir, Tali; Einstein, Andrew J; Bokhari, Sabahat; Fish, Mathews B; Ruddy, Terrence D; Kaufmann, Philipp; Sinusas, Albert J; Miller, Edward J; Bateman, Timothy M; Dorbala, Sharmila; Di Carli, Marcelo; Germano, Guido; Otaki, Yuka; Tamarappoo, Balaji K; Dey, Damini; Berman, Daniel S; Slomka, Piotr J

2018-03-12

The study evaluated the automatic prediction of obstructive disease from myocardial perfusion imaging (MPI) by deep learning as compared with total perfusion deficit (TPD). Deep convolutional neural networks trained with a large multicenter population may provide improved prediction of per-patient and per-vessel coronary artery disease from single-photon emission computed tomography MPI. A total of 1,638 patients (67% men) without known coronary artery disease, undergoing stress 99m Tc-sestamibi or tetrofosmin MPI with new generation solid-state scanners in 9 different sites, with invasive coronary angiography performed within 6 months of MPI, were studied. Obstructive disease was defined as ≥70% narrowing of coronary arteries (≥50% for left main artery). Left ventricular myocardium was segmented using clinical nuclear cardiology software and verified by an expert reader. Stress TPD was computed using sex- and camera-specific normal limits. Deep learning was trained using raw and quantitative polar maps and evaluated for prediction of obstructive stenosis in a stratified 10-fold cross-validation procedure. A total of 1,018 (62%) patients and 1,797 of 4,914 (37%) arteries had obstructive disease. Area under the receiver-operating characteristic curve for disease prediction by deep learning was higher than for TPD (per patient: 0.80 vs. 0.78; per vessel: 0.76 vs. 0.73: p < 0.01). With deep learning threshold set to the same specificity as TPD, per-patient sensitivity improved from 79.8% (TPD) to 82.3% (deep learning) (p < 0.05), and per-vessel sensitivity improved from 64.4% (TPD) to 69.8% (deep learning) (p < 0.01). Deep learning has the potential to improve automatic interpretation of MPI as compared with current clinical methods. Copyright © 2018 American College of Cardiology Foundation. Published by Elsevier Inc. All rights reserved.

Intelligent Detection of Structure from Remote Sensing Images Based on Deep Learning Method

NASA Astrophysics Data System (ADS)

Xin, L.

2018-04-01

Utilizing high-resolution remote sensing images for earth observation has become the common method of land use monitoring. It requires great human participation when dealing with traditional image interpretation, which is inefficient and difficult to guarantee the accuracy. At present, the artificial intelligent method such as deep learning has a large number of advantages in the aspect of image recognition. By means of a large amount of remote sensing image samples and deep neural network models, we can rapidly decipher the objects of interest such as buildings, etc. Whether in terms of efficiency or accuracy, deep learning method is more preponderant. This paper explains the research of deep learning method by a great mount of remote sensing image samples and verifies the feasibility of building extraction via experiments.

Dental students' perception of their approaches to learning in a PBL programme.

PubMed

Haghparast, H; Ghorbani, A; Rohlin, M

2017-08-01

To compare dental students' perceptions of their learning approaches between different years of a problem-based learning (PBL) programme. The hypothesis was that in a comparison between senior and junior students, the senior students would perceive themselves as having a higher level of deep learning approach and a lower level of surface learning approach than junior students would. This hypothesis was based on the fact that senior students have longer experience of a student-centred educational context, which is supposed to underpin student learning. Students of three cohorts (first year, third year and fifth year) of a PBL-based dental programme were asked to respond to a questionnaire (R-SPQ-2F) developed to analyse students' learning approaches, that is deep approach and surface approach, using four subscales including deep strategy, surface strategy, deep motive and surface motive. The results of the three cohorts were compared using a one-way analysis of variance (ANOVA). A P-value was set at <0.05 for statistical significance. The fifth-year students demonstrated a lower surface approach than the first-year students (P = 0.020). There was a significant decrease in surface strategy from the first to the fifth year (P = 0.003). No differences were found concerning deep approach or its subscales (deep strategy and deep motive) between the mean scores of the three cohorts. The results did not show the expected increased depth in learning approaches over the programme years. © 2016 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd.

Deep Learning in Gastrointestinal Endoscopy.

PubMed

Patel, Vivek; Armstrong, David; Ganguli, Malika; Roopra, Sandeep; Kantipudi, Neha; Albashir, Siwar; Kamath, Markad V

2016-01-01

Gastrointestinal (GI) endoscopy is used to inspect the lumen or interior of the GI tract for several purposes, including, (1) making a clinical diagnosis, in real time, based on the visual appearances; (2) taking targeted tissue samples for subsequent histopathological examination; and (3) in some cases, performing therapeutic interventions targeted at specific lesions. GI endoscopy is therefore predicated on the assumption that the operator-the endoscopist-is able to identify and characterize abnormalities or lesions accurately and reproducibly. However, as in other areas of clinical medicine, such as histopathology and radiology, many studies have documented marked interobserver and intraobserver variability in lesion recognition. Thus, there is a clear need and opportunity for techniques or methodologies that will enhance the quality of lesion recognition and diagnosis and improve the outcomes of GI endoscopy. Deep learning models provide a basis to make better clinical decisions in medical image analysis. Biomedical image segmentation, classification, and registration can be improved with deep learning. Recent evidence suggests that the application of deep learning methods to medical image analysis can contribute significantly to computer-aided diagnosis. Deep learning models are usually considered to be more flexible and provide reliable solutions for image analysis problems compared to conventional computer vision models. The use of fast computers offers the possibility of real-time support that is important for endoscopic diagnosis, which has to be made in real time. Advanced graphics processing units and cloud computing have also favored the use of machine learning, and more particularly, deep learning for patient care. This paper reviews the rapidly evolving literature on the feasibility of applying deep learning algorithms to endoscopic imaging.

New Techniques for Deep Learning with Geospatial Data using TensorFlow, Earth Engine, and Google Cloud Platform

NASA Astrophysics Data System (ADS)

Hancher, M.

2017-12-01

Recent years have seen promising results from many research teams applying deep learning techniques to geospatial data processing. In that same timeframe, TensorFlow has emerged as the most popular framework for deep learning in general, and Google has assembled petabytes of Earth observation data from a wide variety of sources and made them available in analysis-ready form in the cloud through Google Earth Engine. Nevertheless, developing and applying deep learning to geospatial data at scale has been somewhat cumbersome to date. We present a new set of tools and techniques that simplify this process. Our approach combines the strengths of several underlying tools: TensorFlow for its expressive deep learning framework; Earth Engine for data management, preprocessing, postprocessing, and visualization; and other tools in Google Cloud Platform to train TensorFlow models at scale, perform additional custom parallel data processing, and drive the entire process from a single familiar Python development environment. These tools can be used to easily apply standard deep neural networks, convolutional neural networks, and other custom model architectures to a variety of geospatial data structures. We discuss our experiences applying these and related tools to a range of machine learning problems, including classic problems like cloud detection, building detection, land cover classification, as well as more novel problems like illegal fishing detection. Our improved tools will make it easier for geospatial data scientists to apply modern deep learning techniques to their own problems, and will also make it easier for machine learning researchers to advance the state of the art of those techniques.

A hybrid deep learning approach to predict malignancy of breast lesions using mammograms

NASA Astrophysics Data System (ADS)

Wang, Yunzhi; Heidari, Morteza; Mirniaharikandehei, Seyedehnafiseh; Gong, Jing; Qian, Wei; Qiu, Yuchen; Zheng, Bin

2018-03-01

Applying deep learning technology to medical imaging informatics field has been recently attracting extensive research interest. However, the limited medical image dataset size often reduces performance and robustness of the deep learning based computer-aided detection and/or diagnosis (CAD) schemes. In attempt to address this technical challenge, this study aims to develop and evaluate a new hybrid deep learning based CAD approach to predict likelihood of a breast lesion detected on mammogram being malignant. In this approach, a deep Convolutional Neural Network (CNN) was firstly pre-trained using the ImageNet dataset and serve as a feature extractor. A pseudo-color Region of Interest (ROI) method was used to generate ROIs with RGB channels from the mammographic images as the input to the pre-trained deep network. The transferred CNN features from different layers of the CNN were then obtained and a linear support vector machine (SVM) was trained for the prediction task. By applying to a dataset involving 301 suspicious breast lesions and using a leave-one-case-out validation method, the areas under the ROC curves (AUC) = 0.762 and 0.792 using the traditional CAD scheme and the proposed deep learning based CAD scheme, respectively. An ensemble classifier that combines the classification scores generated by the two schemes yielded an improved AUC value of 0.813. The study results demonstrated feasibility and potentially improved performance of applying a new hybrid deep learning approach to develop CAD scheme using a relatively small dataset of medical images.

The effects of deep network topology on mortality prediction.

PubMed

Hao Du; Ghassemi, Mohammad M; Mengling Feng

2016-08-01

Deep learning has achieved remarkable results in the areas of computer vision, speech recognition, natural language processing and most recently, even playing Go. The application of deep-learning to problems in healthcare, however, has gained attention only in recent years, and it's ultimate place at the bedside remains a topic of skeptical discussion. While there is a growing academic interest in the application of Machine Learning (ML) techniques to clinical problems, many in the clinical community see little incentive to upgrade from simpler methods, such as logistic regression, to deep learning. Logistic regression, after all, provides odds ratios, p-values and confidence intervals that allow for ease of interpretation, while deep nets are often seen as `black-boxes' which are difficult to understand and, as of yet, have not demonstrated performance levels far exceeding their simpler counterparts. If deep learning is to ever take a place at the bedside, it will require studies which (1) showcase the performance of deep-learning methods relative to other approaches and (2) interpret the relationships between network structure, model performance, features and outcomes. We have chosen these two requirements as the goal of this study. In our investigation, we utilized a publicly available EMR dataset of over 32,000 intensive care unit patients and trained a Deep Belief Network (DBN) to predict patient mortality at discharge. Utilizing an evolutionary algorithm, we demonstrate automated topology selection for DBNs. We demonstrate that with the correct topology selection, DBNs can achieve better prediction performance compared to several bench-marking methods.

Hydrothermal sediments as a potential record of seawater Nd isotope compositions: The Rainbow vent site (36°14'N, Mid-Atlantic Ridge)

NASA Astrophysics Data System (ADS)

Chavagnac, ValéRie; Palmer, Martin R.; Milton, J. Andrew; Green, Darryl R. H.; German, Christopher R.

2006-09-01

Geochemical compositions and Sr and Nd isotopes were measured in two cores collected ˜2 and 5 km from the Rainbow hydrothermal vent site on the Mid-Atlantic Ridge. Overall, the cores record enrichments in Fe and other metals from hydrothermal fallout, but sequential dissolution of the sediments allows discrimination between a leach phase (easily leachable) and a residue phase (refractory). The oxy-anion and transition metal distribution combined with rare earth element (REE) patterns suggest that (1) the leach fraction is a mixture of biogenic carbonate and hydrothermal Fe-Mn oxy-hydroxide with no significant contribution from detrital material and (2) >99.5% of the REE content of the leach fraction is of seawater origin. In addition, the leach fraction has an average 87Sr/86Sr ratio indistinguishable from modern seawater at 0.70916. Although we lack the ɛNd value of present-day deep water at the Rainbow vent site, we believe that the REE budget of the leach fraction is predominantly of seawater origin. We suggest therefore that the leach fraction provides a record of local seawater ɛNd values. Nd isotope data from these cores span the period of 4-14 ka (14C ages) and yield ɛNd values for North East Atlantic Deep Water (NEADW) that are higher (-9.3 to -11.1) than those observed in the nearby Madeira Abyssal Plain from the same depth (-12.4 ± 0.9). This observation suggests that either the Iceland-Scotland Overflow Water (ISOW) and Lower Deep Water contributions to the formation of NEADW are higher along the Mid-Atlantic Ridge than in the surrounding basins or that the relative proportion of ISOW was higher during this period than is observed today. This study indicates that hydrothermal sediments have the potential to provide a higher-resolution record of deep water ɛNd values, and hence deepwater circulation patterns in the oceans, than is possible from other types of sediments.

Geodynamic Implications of Himu Mantle In The Source of Tertiary Volcanics From The Veneto Region (south Eastern Alps)

NASA Astrophysics Data System (ADS)

Macera, P.; Gasperini, D.; Blichert-Toft; Bosch, D.; del Moro, A.; Dini, G.; Martin, S.; Piromallo, C.

DuringTertiary times extensive mafic volcanism took place in the South-Eastern Alps, along a half-graben structure bounded by the Schio-Vicenza main fault. This mag- matism gave rise to four main volcanic centers: Lessini, Berici, Euganei, and Maros- tica. The dominating rock types are alkali basalts, basanites and transitional basalts, with hawaiites, trachybasalts, tephrites, basaltic andesites, and differentiated rocks be- ing less common. Major and trace element and Sr-Nd-Hf-Pb isotopic data for the most primitive lavas from each volcanic center show the typical features of HIMU hotspot volcanism, variably diluted by a depleted asthenospheric mantle component (87Sr/86Sr48Ma = 0.70314-0.70321; eNd48Ma = +6.4 to +6.5; eHf48Ma = +6.4 to +8.1, 206Pb/204Pb48Ma = 18.786-19.574). Since the HIMU component is consid- ered to be of deep mantle origin, its presence in a tectonic environment dominated by subduction (the Alpine subduction of the European plate below the Adria plate) has significant geodynamic implications. Slab detachment and ensuing rise of deep man- tle material into the lithospheric gap is proposed to be a viable mechanism of hotspot magmatism in a subduction zone setting. Interaction between deep-seated plume ma- terial and shallow depleted asthenospheric mantle may account for the geochemical features of the Veneto volcanics, as well as those of the so-called enriched astheno- spheric reservoir (EAR) component. Ascending counterflow of deep mantle material through the lithospheric gap to the top of the subducting slab further may induce heat- ing of the overriding plate and trigger it to partially melt. Upwelling of the resulting mafic magmas and their subsequent underplating at the mantle-lower crust bound- ary would favor partial melting of the lower crust, thereby giving rise to the bimodal mafic-felsic magmatism that characterizes the whole Periadriatic province. According to this model, the HIMU-like magmatism of the Alpine foreland is therefore closely related to the calc-alkaline magmatism of the Periadriatic Lineament, and caused by the same mechanism of Tertiary Alpine convergence tectonics.

Deep and Surface Learning in Problem-Based Learning: A Review of the Literature

ERIC Educational Resources Information Center

Dolmans, Diana H. J. M.; Loyens, Sofie M. M.; Marcq, Hélène; Gijbels, David

2016-01-01

In problem-based learning (PBL), implemented worldwide, students learn by discussing professionally relevant problems enhancing application and integration of knowledge, which is assumed to encourage students towards a deep learning approach in which students are intrinsically interested and try to understand what is being studied. This review…

Deciphering fluid sources of hydrothermal systems: A combined Sr- and S-isotope study on barite (Schwarzwald, SW Germany)

USGS Publications Warehouse

Staude, S.; Gob, S.; Pfaff, K.; Strobele, F.; Premo, W.R.; Markl, G.

2011-01-01

Primary and secondary barites from hydrothermal mineralizations in SW Germany were investigated, for the first time, by a combination of strontium (Sr) isotope systematics (87Sr/86Sr), Sr contents and δ34S values to distinguish fluid sources and precipitation mechanisms responsible for their formation. Barite of Permian age derived its Sr solely from crystalline basement rocks, whereas all younger barite also incorporate Sr from formation waters of the overlying sediments. In fact, most of the Sr in younger barite is leached from Lower and Middle Triassic sediments. In contrast, most of the sulfur (S) of Permian, Jurassic and northern Schwarzwald Miocene barite originated from basement rocks. The S source of Upper Rhinegraben (URG)-related Paleogene barite differs depending on geographic position: for veins of the southern URG, it is the Oligocene evaporitic sequence, while central URG mineralizations derived its S from Middle Triassic evaporites. Using Sr isotopes of barite of known age combined with estimates on the Sr contents and Sr isotopic ratios of the fluids' source rocks, we were able to quantify mixing ratios of basement-derived fluids and sedimentary formation waters for the first time. These calculations show that Jurassic barite formed by mixing of 75–95% ascending basement-derived fluids with 5–25% sedimentary formation water, but that only 20–55% of the Sr was brought by the basement-derived fluid to the depositional site. Miocene barite formed by mixing of an ascending basement-derived brine (60–70%) with 30–40% sedimentary formation waters. In this case, only 8–15% of the Sr was derived from the deep brine. This fluid-mixing calculation is an example for deposits in which the fluid source is known. This method applied to a greater number of deposits formed at different times and in various geological settings may shed light on more general causes of fluid movement in the Earth's crust and on the formation of hydrothermal ore deposits.

Deep Unsupervised Learning on a Desktop PC: A Primer for Cognitive Scientists.

PubMed

Testolin, Alberto; Stoianov, Ivilin; De Filippo De Grazia, Michele; Zorzi, Marco

2013-01-01

Deep belief networks hold great promise for the simulation of human cognition because they show how structured and abstract representations may emerge from probabilistic unsupervised learning. These networks build a hierarchy of progressively more complex distributed representations of the sensory data by fitting a hierarchical generative model. However, learning in deep networks typically requires big datasets and it can involve millions of connection weights, which implies that simulations on standard computers are unfeasible. Developing realistic, medium-to-large-scale learning models of cognition would therefore seem to require expertise in programing parallel-computing hardware, and this might explain why the use of this promising approach is still largely confined to the machine learning community. Here we show how simulations of deep unsupervised learning can be easily performed on a desktop PC by exploiting the processors of low cost graphic cards (graphic processor units) without any specific programing effort, thanks to the use of high-level programming routines (available in MATLAB or Python). We also show that even an entry-level graphic card can outperform a small high-performance computing cluster in terms of learning time and with no loss of learning quality. We therefore conclude that graphic card implementations pave the way for a widespread use of deep learning among cognitive scientists for modeling cognition and behavior.

Deep Unsupervised Learning on a Desktop PC: A Primer for Cognitive Scientists

PubMed Central

Testolin, Alberto; Stoianov, Ivilin; De Filippo De Grazia, Michele; Zorzi, Marco

2013-01-01

Deep belief networks hold great promise for the simulation of human cognition because they show how structured and abstract representations may emerge from probabilistic unsupervised learning. These networks build a hierarchy of progressively more complex distributed representations of the sensory data by fitting a hierarchical generative model. However, learning in deep networks typically requires big datasets and it can involve millions of connection weights, which implies that simulations on standard computers are unfeasible. Developing realistic, medium-to-large-scale learning models of cognition would therefore seem to require expertise in programing parallel-computing hardware, and this might explain why the use of this promising approach is still largely confined to the machine learning community. Here we show how simulations of deep unsupervised learning can be easily performed on a desktop PC by exploiting the processors of low cost graphic cards (graphic processor units) without any specific programing effort, thanks to the use of high-level programming routines (available in MATLAB or Python). We also show that even an entry-level graphic card can outperform a small high-performance computing cluster in terms of learning time and with no loss of learning quality. We therefore conclude that graphic card implementations pave the way for a widespread use of deep learning among cognitive scientists for modeling cognition and behavior. PMID:23653617

The optical properties of Sr3SiAl10O20 and Sr3SiAl10O20:Mn4+

NASA Astrophysics Data System (ADS)

Jansen, Thomas; Jüstel, Thomas

2017-11-01

Mn4+-activated luminescent materials have attracted significant attention recently. In particular, alkaline earth aluminates, such as Sr4Al14O25:Mn4+ and CaAl12O19:Mn4+, emit light in the red region, which can be exploited in phosphor-converted LEDs. We applied a sol-gel precursor followed by a ceramic method to synthesize highly crystalline Mn4+-doped Sr3SiAl10O20. The compound Sr3SiAl10O20:Mn4+ exhibits deep red photoluminescence that peaks at 663 nm, which can be assigned to the 2Eg → 4A2g intraconfigurational transition of Mn4+ ([Ar]3d.3 configuration) within the [MnO6]8- octahedra on the aluminum site in the Sr3SiAl10O20 (Space group C12/m1) host structure. The photoluminescence properties, such as the temperature dependence of the luminescence intensity and luminescence lifetime, are presented. Furthermore, the luminescence intensity as function of the activator concentration was investigated. Additionally, the band structure of the undoped host material was treated with Density Functional Theory (DFT). The theoretical results were evaluated experimentally with diffuse UV reflectance spectroscopy. Finally, the crystal field and Racah parameters were extracted to compare with those reported in the literature.

Factors controlling Li concentration and isotopic composition in formation waters and host rocks of Marcellus Shale, Appalachian Basin

USGS Publications Warehouse

Phan, Thai T.; Capo, Rosemary C; Stewart, Brian W.; Macpherson, Gwen; Rowan, Elisabeth L.; Hammack, Richard W.

2015-01-01

In Greene Co., southwest Pennsylvania, the Upper Devonian sandstone formation waters have δ7Li values of + 14.6 ± 1.2 (2SD, n = 25), and are distinct from Marcellus Shale formation waters which have δ7Li of + 10.0 ± 0.8 (2SD, n = 12). These two formation waters also maintain distinctive 87Sr/86Sr ratios suggesting hydrologic separation between these units. Applying temperature-dependent illitilization model to Marcellus Shale, we found that Li concentration in clay minerals increased with Li concentration in pore fluid during diagenetic illite-smectite transition. Samples from north central PA show a much smaller range in both δ7Li and 87Sr/86Sr than in southwest Pennsylvania. Spatial variations in Li and δ7Li values show that Marcellus formation waters are not homogeneous across the Appalachian Basin. Marcellus formation waters in the northeastern Pennsylvania portion of the basin show a much smaller range in both δ7Li and 87Sr/86Sr, suggesting long term, cross-formational fluid migration in this region. Assessing the impact of potential mixing of fresh water with deep formation water requires establishment of a geochemical and isotopic baseline in the shallow, fresh water aquifers, and site specific characterization of formation water, followed by long-term monitoring, particularly in regions of future shale gas development.

Determining Solute Sources and Water Flowpaths in a Forested Headwater Catchment: Advances With the Ca-Sr-Ba Multi-tracer

NASA Astrophysics Data System (ADS)

Bullen, T. D.; Bailey, S. W.; McGuire, K. J.; Zimmer, M. A.; Ross, D. S.

2011-12-01

Determining solute sources and water flowpaths in catchments is of critical importance to development of models that effectively describe catchment function. For solutes in soil water and stream water, simple mass balance models that compare precipitation input to catchment outlet compositions can predict average mineral weathering contributions for the catchment as a whole, but fail to provide information about either variability of contributions from different portions of the catchment and different soil depths or processes such as ion exchange and biological cycling. In order to better understand how forested headwater catchments function, we are interpreting concentration and isotope ratios of the alkaline earth elements Ca, Sr and Ba in streamwater, groundwater, the soil ion exchange pool and plants in a hydropedologic context at the 41 hectare hydrologic reference catchment (Watershed 3) at the Hubbard Brook Experimental Forest, New Hampshire, USA. This forested headwater catchment consists of a beech-birch-maple-spruce forest growing on vertically- and laterally-developed Spodosols and Inceptisols formed on granitoid glacial till that mantles Paleozoic metamorphic bedrock. Across the watershed in terms of the soil ion exchange pool, the forest floor has high Sr/Ba and Ca/Sr ratios, mineral soils have intermediate Sr/Ba and low Ca/Sr, and relatively unweathered till in the C horizon has low Sr/Ba and high Ca/Sr. Waters moving through these various compartments will obtain Sr/Ba and Ca/Sr ratios reflecting these characteristics, and thus variations of Sr/Ba and Ca/Sr of streamwater provide evidence of the depth of water flowpaths feeding the streams. 87Sr/86Sr of exchangeable Sr spans a broad range from 0.715 to 0.725, with highest values along the mid-to upper flanks of the catchment and lowest values in a broad zone along the central axis of the catchment associated with numerous groundwater seeps. Thus, variations of 87Sr/86Sr in streamwater provide evidence of the spatial distribution of water flowpaths feeding the streams. In addition, we are exploring the use of Sr and Ba stable isotope ratios (88Sr/86Sr, 138Ba/134Ba) as novel tracers of Sr and Ba sources in catchments. Initial results indicate that both Sr and Ba stable isotopes are fractionated by plants similarly to patterns observed globally for Ca stable isotopes. We hypothesize that while biologically-cycled Ca is efficiently retained in the organic soil-plant system, biologically-cycled Sr and especially Ba will be more easily leached by soil waters and delivered to the streams and thus their stable isotope ratios may provide an additional means to distinguish between shallow and deep water flowpaths in forested catchments.

Is Multitask Deep Learning Practical for Pharma?

PubMed

Ramsundar, Bharath; Liu, Bowen; Wu, Zhenqin; Verras, Andreas; Tudor, Matthew; Sheridan, Robert P; Pande, Vijay

2017-08-28

Multitask deep learning has emerged as a powerful tool for computational drug discovery. However, despite a number of preliminary studies, multitask deep networks have yet to be widely deployed in the pharmaceutical and biotech industries. This lack of acceptance stems from both software difficulties and lack of understanding of the robustness of multitask deep networks. Our work aims to resolve both of these barriers to adoption. We introduce a high-quality open-source implementation of multitask deep networks as part of the DeepChem open-source platform. Our implementation enables simple python scripts to construct, fit, and evaluate sophisticated deep models. We use our implementation to analyze the performance of multitask deep networks and related deep models on four collections of pharmaceutical data (three of which have not previously been analyzed in the literature). We split these data sets into train/valid/test using time and neighbor splits to test multitask deep learning performance under challenging conditions. Our results demonstrate that multitask deep networks are surprisingly robust and can offer strong improvement over random forests. Our analysis and open-source implementation in DeepChem provide an argument that multitask deep networks are ready for widespread use in commercial drug discovery.

Geological and Geochemical Characteristics of Skarn Type Lead-Zinc Deposit in Baoshan Block, Yunnan Province

NASA Astrophysics Data System (ADS)

Yao, Xue; Wang, Peng

2017-11-01

Baoshan block is an important Pb-Zn-Fe-Cu polymetallic ore-concentration area which is located in southern of the Sanjiang metallogenic belt in western Yunnan. The article is studying about the geological and geochemical characteristics of the skarn type lead-zinc deposit in Baoshan block. The skarn-type lead-zinc deposit Baoshan block is characterized by skarn and skarn marble, and the orebodies are layered, or bedded along the interlayer fault, which are significantly controlled by structure. The research about Stable isotope S, H and O indicates that the ore-forming fluids are mainly derived from magmatic water, partly mixed with parts of metamorphic water and atmospheric precipitation. The initial Sr isotopic Sr87/Sr86 ratio suggests that the ore-forming materials derived from deep concealed magmatic rock, age of Rb-Sr mineralization is similar to that of Yanshanian granite. In conclusion, the Yanshanian tectonic-magmatic-fluid coupling mineralization of Yanshan formation is the main reason for the skarn-type lead-zinc deposit in the Baoshan block.

Some Consequences of a Time Dependent Speed of Light

NASA Astrophysics Data System (ADS)

Smith, Felix T.

2007-06-01

For reasons connected with both cosmology (the flatness and horizon problems) and atomic physics (n-body Dirac equation, etc.), various proposals have been made to modify general or special relativity(SR) to accommodate a cosmologically decreasing light speed [J. Magueijo, Rep. Prog. Phys. 66, 2025 (2003)]. Two such theories, projective SR [S.N. Manida, gr-qc/9905046; S. S. Stepanov, physics/9909009 and Phys. Rev. D, 62, 023507 (2000)] and symmetric SR [F.T. Smith, Ann. Fond. L. de Broglie, 30, 179 (2005)] adapt special relativity to in different ways to an expanding, hyperbolically curved position space and predict time-dependences of c within reach of measurement but differing by a factor of two. Both theories bring in a new constant λ-1=σ=c^2H0-1. As Magueijo points, out the role of c in physics and cosmology is so profound that many deep changes must follow if is not absolutely invariant in space and time. In particular, symmetric SR brings a new light to the Dirac large-number relationship between the constants of gravitation and atomic physics.

Rb-Sr and Sm-Nd Isotopic Studies of Lunar Green and Orange Glasses

NASA Technical Reports Server (NTRS)

Shih, C.-Y.; Nyquist, L. E.; Reese, Y.

2012-01-01

Lunar volcanic glassy beads have been considered as quenched basaltic magmas derived directly from deep lunar mantle during fire-fountaining eruptions [1]. Since these sub-mm size glassy melt droplets were cooled in a hot gaseous medium during free flight [2], they have not been subject to mineral fractionations. Thus, they represent primary magmas and are the best samples for the investigation of the lunar mantle. Previously, we presented preliminary Rb- Sr and Sm-Nd isotopic results for green and orange glassy samples from green glass clod 15426,63 and orange soil 74220,44, respectively [3]. Using these isotopic data, initial Sr-87/Sr-86 and Nd ratios for these pristine mare glass sources can be calculated from their respective crystallization ages previously determined by other age-dating techniques. These isotopic data were used to evaluate the mineralogy of the mantle sources. In this report, we analyzed additional glassy samples in order to further characterize isotopic signatures of their source regions. Also, we'll postulate a relationship between these two major mare basalt source mineralogies in the context of lunar magma ocean dynamics.

Soft resonator of omnidirectional resonance for acoustic metamaterials with a negative bulk modulus

PubMed Central

Jing, Xiaodong; Meng, Yang; Sun, Xiaofeng

2015-01-01

Monopolar resonance is of fundamental importance in the acoustic field. Here, we present the realization of a monopolar resonance that goes beyond the concept of Helmholtz resonators. The balloon-like soft resonator (SR) oscillates omnidirectionally and radiates from all parts of its spherical surface, eliminating the need for a hard wall for the cavity and baffle effects. For airborne sound, such a low-modulus resonator can be made extremely lightweight. Deep subwavelength resonance is achieved when the SR is tuned by adjusting the shell thickness, benefiting from the large density contrast between the shell material and the encapsulated gas. The SR resonates with near-perfect monopole symmetry, as demonstrated by the theoretical and experimental results, which are in excellent agreement. For a lattice of SRs, a band gap occurs and blocks near-total transmission, and the effective bulk modulus exhibits a prominent negative band, while the effective mass density remains unchanged. Our study shows that the SR is suitable for building 3D acoustic metamaterials and provides a basis for constructing left-handed materials as a new means of creating a negative bulk modulus. PMID:26538085

Distribution of 90Sr, 137Cs and 239,240Pu in Caspian Sea water and biota

NASA Astrophysics Data System (ADS)

Povinec, Pavel P.; Froehlich, Klaus; Gastaud, Janine; Oregioni, Beniamino; Pagava, Samson V.; Pham, Mai K.; Rusetski, Vladimir

2003-09-01

Two sampling expeditions were carried out in the Caspian Sea in 1995 and 1996. The aim was to investigate oceanographic conditions, water dynamics of the Sea and to measure radionuclide concentrations using 90Sr, 137Cs and 239,240Pu as tracers in the water column. Of the three basins comprising the Caspian Sea, the two deep basins (the central and southern basins) appear to be rapidly ventilated on a time scale of about 30 years, as shown by the penetration of radionuclides to bottom waters. The main source of radionuclides in the Sea has been global fallout and subsequent river run-off from catchment areas. At the stations visited, there were no signs of radioactive waste dumping, although the 90Sr levels found were higher than expected from global fallout, which may be due to remobilization of 90Sr from soil and its transport by rivers to the Sea. Radionuclide concentrations in fish and caviar are within the expected ranges and are not of radiological importance for consumption of fish and caviar from the Caspian Sea.

Using Literature to Build Self-Esteem in Adolescents with Learning and Behavior Problems.

ERIC Educational Resources Information Center

Miller, Darcy

1994-01-01

Describes the Literature Project, a literature-based thematic approach to reading instruction designed to improve self-esteem in adolescents with learning and behavior problems. Discusses the effectiveness of the program. (SR)

Active appearance model and deep learning for more accurate prostate segmentation on MRI

NASA Astrophysics Data System (ADS)

Cheng, Ruida; Roth, Holger R.; Lu, Le; Wang, Shijun; Turkbey, Baris; Gandler, William; McCreedy, Evan S.; Agarwal, Harsh K.; Choyke, Peter; Summers, Ronald M.; McAuliffe, Matthew J.

2016-03-01

Prostate segmentation on 3D MR images is a challenging task due to image artifacts, large inter-patient prostate shape and texture variability, and lack of a clear prostate boundary specifically at apex and base levels. We propose a supervised machine learning model that combines atlas based Active Appearance Model (AAM) with a Deep Learning model to segment the prostate on MR images. The performance of the segmentation method is evaluated on 20 unseen MR image datasets. The proposed method combining AAM and Deep Learning achieves a mean Dice Similarity Coefficient (DSC) of 0.925 for whole 3D MR images of the prostate using axial cross-sections. The proposed model utilizes the adaptive atlas-based AAM model and Deep Learning to achieve significant segmentation accuracy.

Simulation of noisy dynamical system by Deep Learning

NASA Astrophysics Data System (ADS)

Yeo, Kyongmin

2017-11-01

Deep learning has attracted huge attention due to its powerful representation capability. However, most of the studies on deep learning have been focused on visual analytics or language modeling and the capability of the deep learning in modeling dynamical systems is not well understood. In this study, we use a recurrent neural network to model noisy nonlinear dynamical systems. In particular, we use a long short-term memory (LSTM) network, which constructs internal nonlinear dynamics systems. We propose a cross-entropy loss with spatial ridge regularization to learn a non-stationary conditional probability distribution from a noisy nonlinear dynamical system. A Monte Carlo procedure to perform time-marching simulations by using the LSTM is presented. The behavior of the LSTM is studied by using noisy, forced Van der Pol oscillator and Ikeda equation.

Deep Learning in Medical Image Analysis

PubMed Central

Shen, Dinggang; Wu, Guorong; Suk, Heung-Il

2016-01-01

The computer-assisted analysis for better interpreting images have been longstanding issues in the medical imaging field. On the image-understanding front, recent advances in machine learning, especially, in the way of deep learning, have made a big leap to help identify, classify, and quantify patterns in medical images. Specifically, exploiting hierarchical feature representations learned solely from data, instead of handcrafted features mostly designed based on domain-specific knowledge, lies at the core of the advances. In that way, deep learning is rapidly proving to be the state-of-the-art foundation, achieving enhanced performances in various medical applications. In this article, we introduce the fundamentals of deep learning methods; review their successes to image registration, anatomical/cell structures detection, tissue segmentation, computer-aided disease diagnosis or prognosis, and so on. We conclude by raising research issues and suggesting future directions for further improvements. PMID:28301734

Deep learning of support vector machines with class probability output networks.

PubMed

Kim, Sangwook; Yu, Zhibin; Kil, Rhee Man; Lee, Minho

2015-04-01

Deep learning methods endeavor to learn features automatically at multiple levels and allow systems to learn complex functions mapping from the input space to the output space for the given data. The ability to learn powerful features automatically is increasingly important as the volume of data and range of applications of machine learning methods continues to grow. This paper proposes a new deep architecture that uses support vector machines (SVMs) with class probability output networks (CPONs) to provide better generalization power for pattern classification problems. As a result, deep features are extracted without additional feature engineering steps, using multiple layers of the SVM classifiers with CPONs. The proposed structure closely approaches the ideal Bayes classifier as the number of layers increases. Using a simulation of classification problems, the effectiveness of the proposed method is demonstrated. Copyright © 2014 Elsevier Ltd. All rights reserved.

Diverse assessment and active student engagement sustain deep learning: A comparative study of outcomes in two parallel introductory biochemistry courses.

PubMed

Bevan, Samantha J; Chan, Cecilia W L; Tanner, Julian A

2014-01-01

Although there is increasing evidence for a relationship between courses that emphasize student engagement and achievement of student deep learning, there is a paucity of quantitative comparative studies in a biochemistry and molecular biology context. Here, we present a pedagogical study in two contrasting parallel biochemistry introductory courses to compare student surface and deep learning. Surface and deep learning were measured quantitatively by a study process questionnaire at the start and end of the semester, and qualitatively by questionnaires and interviews with students. In the traditional lecture/examination based course, there was a dramatic shift to surface learning approaches through the semester. In the course that emphasized student engagement and adopted multiple forms of assessment, a preference for deep learning was sustained with only a small reduction through the semester. Such evidence for the benefits of implementing student engagement and more diverse non-examination based assessment has important implications for the design, delivery, and renewal of introductory courses in biochemistry and molecular biology. © 2014 The International Union of Biochemistry and Molecular Biology.

Airline Passenger Profiling Based on Fuzzy Deep Machine Learning.

PubMed

Zheng, Yu-Jun; Sheng, Wei-Guo; Sun, Xing-Ming; Chen, Sheng-Yong

2017-12-01

Passenger profiling plays a vital part of commercial aviation security, but classical methods become very inefficient in handling the rapidly increasing amounts of electronic records. This paper proposes a deep learning approach to passenger profiling. The center of our approach is a Pythagorean fuzzy deep Boltzmann machine (PFDBM), whose parameters are expressed by Pythagorean fuzzy numbers such that each neuron can learn how a feature affects the production of the correct output from both the positive and negative sides. We propose a hybrid algorithm combining a gradient-based method and an evolutionary algorithm for training the PFDBM. Based on the novel learning model, we develop a deep neural network (DNN) for classifying normal passengers and potential attackers, and further develop an integrated DNN for identifying group attackers whose individual features are insufficient to reveal the abnormality. Experiments on data sets from Air China show that our approach provides much higher learning ability and classification accuracy than existing profilers. It is expected that the fuzzy deep learning approach can be adapted for a variety of complex pattern analysis tasks.

Using deep learning for content-based medical image retrieval

NASA Astrophysics Data System (ADS)

Sun, Qinpei; Yang, Yuanyuan; Sun, Jianyong; Yang, Zhiming; Zhang, Jianguo

2017-03-01

Content-Based medical image retrieval (CBMIR) is been highly active research area from past few years. The retrieval performance of a CBMIR system crucially depends on the feature representation, which have been extensively studied by researchers for decades. Although a variety of techniques have been proposed, it remains one of the most challenging problems in current CBMIR research, which is mainly due to the well-known "semantic gap" issue that exists between low-level image pixels captured by machines and high-level semantic concepts perceived by human[1]. Recent years have witnessed some important advances of new techniques in machine learning. One important breakthrough technique is known as "deep learning". Unlike conventional machine learning methods that are often using "shallow" architectures, deep learning mimics the human brain that is organized in a deep architecture and processes information through multiple stages of transformation and representation. This means that we do not need to spend enormous energy to extract features manually. In this presentation, we propose a novel framework which uses deep learning to retrieval the medical image to improve the accuracy and speed of a CBIR in integrated RIS/PACS.

Deep Convolutional Extreme Learning Machine and Its Application in Handwritten Digit Classification

PubMed Central

Yang, Xinyi

2016-01-01

In recent years, some deep learning methods have been developed and applied to image classification applications, such as convolutional neuron network (CNN) and deep belief network (DBN). However they are suffering from some problems like local minima, slow convergence rate, and intensive human intervention. In this paper, we propose a rapid learning method, namely, deep convolutional extreme learning machine (DC-ELM), which combines the power of CNN and fast training of ELM. It uses multiple alternate convolution layers and pooling layers to effectively abstract high level features from input images. Then the abstracted features are fed to an ELM classifier, which leads to better generalization performance with faster learning speed. DC-ELM also introduces stochastic pooling in the last hidden layer to reduce dimensionality of features greatly, thus saving much training time and computation resources. We systematically evaluated the performance of DC-ELM on two handwritten digit data sets: MNIST and USPS. Experimental results show that our method achieved better testing accuracy with significantly shorter training time in comparison with deep learning methods and other ELM methods. PMID:27610128

Deep Convolutional Extreme Learning Machine and Its Application in Handwritten Digit Classification.

PubMed

Pang, Shan; Yang, Xinyi

2016-01-01

In recent years, some deep learning methods have been developed and applied to image classification applications, such as convolutional neuron network (CNN) and deep belief network (DBN). However they are suffering from some problems like local minima, slow convergence rate, and intensive human intervention. In this paper, we propose a rapid learning method, namely, deep convolutional extreme learning machine (DC-ELM), which combines the power of CNN and fast training of ELM. It uses multiple alternate convolution layers and pooling layers to effectively abstract high level features from input images. Then the abstracted features are fed to an ELM classifier, which leads to better generalization performance with faster learning speed. DC-ELM also introduces stochastic pooling in the last hidden layer to reduce dimensionality of features greatly, thus saving much training time and computation resources. We systematically evaluated the performance of DC-ELM on two handwritten digit data sets: MNIST and USPS. Experimental results show that our method achieved better testing accuracy with significantly shorter training time in comparison with deep learning methods and other ELM methods.

Deep learning

NASA Astrophysics Data System (ADS)

Lecun, Yann; Bengio, Yoshua; Hinton, Geoffrey

2015-05-01

Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.

A survey on deep learning in medical image analysis.

PubMed

Litjens, Geert; Kooi, Thijs; Bejnordi, Babak Ehteshami; Setio, Arnaud Arindra Adiyoso; Ciompi, Francesco; Ghafoorian, Mohsen; van der Laak, Jeroen A W M; van Ginneken, Bram; Sánchez, Clara I

2017-12-01

Deep learning algorithms, in particular convolutional networks, have rapidly become a methodology of choice for analyzing medical images. This paper reviews the major deep learning concepts pertinent to medical image analysis and summarizes over 300 contributions to the field, most of which appeared in the last year. We survey the use of deep learning for image classification, object detection, segmentation, registration, and other tasks. Concise overviews are provided of studies per application area: neuro, retinal, pulmonary, digital pathology, breast, cardiac, abdominal, musculoskeletal. We end with a summary of the current state-of-the-art, a critical discussion of open challenges and directions for future research. Copyright © 2017 Elsevier B.V. All rights reserved.

Deep learning.

PubMed

LeCun, Yann; Bengio, Yoshua; Hinton, Geoffrey

2015-05-28

Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.

Long-Term Changes in Chemical Weathering in the Himalayan Region from Indus Fan Sediments

NASA Astrophysics Data System (ADS)

Carter, S.; Griffith, E. M.; Scher, H.; Dellapenna, T. M.; Clift, P. D.

2017-12-01

The Asian Monsoon reflects large-scale interactions between the atmosphere, land, and ocean systems. Increasing our understanding of this system, how and why it has evolved through time, is critically important in order to understand how it may evolve in the future. The radiogenic strontium isotopic signature (87Sr/86Sr) of the clay fraction in deep sea sediment cores within submarine fans has been used as a record of riverine 87Sr/86Sr composition to gain information about Himalayan weathering intensity. Strontium exists in clay minerals primarily in interlayer sites or adsorbed onto mineral surfaces. Interlayer cation exchange is thought to be completed within rivers during recrystallization or neoformation of clays. A record of chemical weathering intensity in the Himalayas is presented by analyzing the 87Sr/86Sr signature of the clay fraction in sediments from International Ocean Discovery Program (IODP) Expedition 355 Sites U1456 and U1457, located on the Indus Fan, eastern Arabian Sea. This record will be coupled with additional records of bulk grain size and K/Al ratios of clay as potentially additional indicators of weathering intensity. The Sr isotopes in the interstitial waters at each site have also been measured (Carter et al., in press) to verify that the Sr in the treated clay fraction is not being reset by diagenesis in the sedimentary column. Initial results verify that the 87Sr/86Sr values of the clay are less than those in the bulk sediment, as expected, but are not similar to pore fluid Sr. 87Sr/86Sr values of the clays show trends suggesting fluctuations in chemical weathering intensity through time. However, bulk grain size and K/Al ratios results conflict with the 87Sr/86Sr values. If the additional proxy records continue to show conflicting results for "weathering intensity", together they may reveal more information regarding the sedimentary system. Ultimately, the various records will either agree, providing strong evidence for changes in chemical weathering and the evolution of the monsoon, or disagree, allowing for further investigation into the relationships between chemical weathering, evolution of the flood plain, and sediment deposition in the fan. These new records will aid in the correlation of Himalayan exhumation and monsoon intensity and help to constrain this dynamic system.

Using Student-Centred Learning Environments to Stimulate Deep Approaches to Learning: Factors Encouraging or Discouraging Their Effectiveness

ERIC Educational Resources Information Center

Baeten, Marlies; Kyndt, Eva; Struyven, Katrien; Dochy, Filip

2010-01-01

This review outlines encouraging and discouraging factors in stimulating the adoption of deep approaches to learning in student-centred learning environments. Both encouraging and discouraging factors can be situated in the context of the learning environment, in students' perceptions of that context and in characteristics of the students…

Understanding Cognitive Presence in an Online and Blended Community of Inquiry: Assessing Outcomes and Processes for Deep Approaches to Learning

ERIC Educational Resources Information Center

Akyol, Zehra; Garrison, D. Randy

2011-01-01

This paper focuses on deep and meaningful learning approaches and outcomes associated with online and blended communities of inquiry. Applying mixed methodology for the research design, the study used transcript analysis, learning outcomes, perceived learning, satisfaction, and interviews to assess learning processes and outcomes. The findings for…

MusiteDeep: a deep-learning framework for general and kinase-specific phosphorylation site prediction.

PubMed

Wang, Duolin; Zeng, Shuai; Xu, Chunhui; Qiu, Wangren; Liang, Yanchun; Joshi, Trupti; Xu, Dong

2017-12-15

Computational methods for phosphorylation site prediction play important roles in protein function studies and experimental design. Most existing methods are based on feature extraction, which may result in incomplete or biased features. Deep learning as the cutting-edge machine learning method has the ability to automatically discover complex representations of phosphorylation patterns from the raw sequences, and hence it provides a powerful tool for improvement of phosphorylation site prediction. We present MusiteDeep, the first deep-learning framework for predicting general and kinase-specific phosphorylation sites. MusiteDeep takes raw sequence data as input and uses convolutional neural networks with a novel two-dimensional attention mechanism. It achieves over a 50% relative improvement in the area under the precision-recall curve in general phosphorylation site prediction and obtains competitive results in kinase-specific prediction compared to other well-known tools on the benchmark data. MusiteDeep is provided as an open-source tool available at https://github.com/duolinwang/MusiteDeep. xudong@missouri.edu. Supplementary data are available at Bioinformatics online. © The Author 2017. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com

Teaching Real-World Applications of Business Statistics Using Communication to Scaffold Learning

ERIC Educational Resources Information Center

Green, Gareth P.; Jones, Stacey; Bean, John C.

2015-01-01

Our assessment research suggests that quantitative business courses that rely primarily on algorithmic problem solving may not produce the deep learning required for addressing real-world business problems. This article illustrates a strategy, supported by recent learning theory, for promoting deep learning by moving students gradually from…

Reconstruction of the Eocene Arctic Ocean Using Ichthyolith Isotope Analyses

NASA Astrophysics Data System (ADS)

Gleason, J. D.; Thomas, D. J.; Moore, T. C.; Waddell, L. M.; Blum, J. D.; Haley, B. A.

2007-12-01

Nd, Sr, O and C isotopic compositions of Eocene fish debris (teeth, bones, scales), and their reduced organic coatings, have been used to reconstruct water mass composition, water column structure, surface productivity and salinities of the Arctic Ocean Basin at Lomonosov Ridge between 55 and 44 Ma. Cleaned ichthyolith samples from IODP Expedition 302 (ACEX) record epsilon Nd values that range from -5.7 to -7.8, distinct from modern Arctic Intermediate Water (-10.5) and North Atlantic Deep Water. These Nd values may record some exchange with Pacific/Tethyan water masses, but inputs from local continental sources are more likely. Sr isotopic values are consistent with a brackish-to-fresh water surface layer (87Sr/86Sr = 0.7079-0.7087) that was poorly mixed with Eocene global seawater (0.7077-0.7078). Leaching experiments show reduced organic coatings to be more radiogenic (>0.7090) than cleaned ichthyolith phosphate. Ichthyolith Sr isotopic variations likely reflect changes in localized river input as a function of shifts in the Arctic hydrologic cycle, and 87Sr/86Sr values might be used as a proxy for surface water salinity. Model mixing calculations indicate salinities of 5 to 20 per mil, lower than estimates based on O isotopes from fish bone carbonate (16 to 26 per mil). Significant salinity drops (i.e., 55 Ma PETM and 48.5 Ma Azolla event) registered in oxygen isotopes do not show large excursions in the 87Sr/86Sr data. Carbon isotopes in fish debris record a spike in organic activity at 48.5 Ma (Azolla event), and otherwise high-productivity waters between 55 and 44 Ma. The combined Sr-Nd-O-C isotopic record is consistent with highly restricted basin-wide circulation in the Eocene, indicative of a highly stratified water column with anoxic bottom waters, a "fresh" water upper layer, and enhanced continental runoff during warm intervals until the first appearance of ice rafted debris at 45 Ma.

Modeling language and cognition with deep unsupervised learning: a tutorial overview

PubMed Central

Zorzi, Marco; Testolin, Alberto; Stoianov, Ivilin P.

2013-01-01

Deep unsupervised learning in stochastic recurrent neural networks with many layers of hidden units is a recent breakthrough in neural computation research. These networks build a hierarchy of progressively more complex distributed representations of the sensory data by fitting a hierarchical generative model. In this article we discuss the theoretical foundations of this approach and we review key issues related to training, testing and analysis of deep networks for modeling language and cognitive processing. The classic letter and word perception problem of McClelland and Rumelhart (1981) is used as a tutorial example to illustrate how structured and abstract representations may emerge from deep generative learning. We argue that the focus on deep architectures and generative (rather than discriminative) learning represents a crucial step forward for the connectionist modeling enterprise, because it offers a more plausible model of cortical learning as well as a way to bridge the gap between emergentist connectionist models and structured Bayesian models of cognition. PMID:23970869

Modeling language and cognition with deep unsupervised learning: a tutorial overview.

PubMed

Zorzi, Marco; Testolin, Alberto; Stoianov, Ivilin P

2013-01-01

Deep unsupervised learning in stochastic recurrent neural networks with many layers of hidden units is a recent breakthrough in neural computation research. These networks build a hierarchy of progressively more complex distributed representations of the sensory data by fitting a hierarchical generative model. In this article we discuss the theoretical foundations of this approach and we review key issues related to training, testing and analysis of deep networks for modeling language and cognitive processing. The classic letter and word perception problem of McClelland and Rumelhart (1981) is used as a tutorial example to illustrate how structured and abstract representations may emerge from deep generative learning. We argue that the focus on deep architectures and generative (rather than discriminative) learning represents a crucial step forward for the connectionist modeling enterprise, because it offers a more plausible model of cortical learning as well as a way to bridge the gap between emergentist connectionist models and structured Bayesian models of cognition.

Action-Driven Visual Object Tracking With Deep Reinforcement Learning.

PubMed

Yun, Sangdoo; Choi, Jongwon; Yoo, Youngjoon; Yun, Kimin; Choi, Jin Young

2018-06-01

In this paper, we propose an efficient visual tracker, which directly captures a bounding box containing the target object in a video by means of sequential actions learned using deep neural networks. The proposed deep neural network to control tracking actions is pretrained using various training video sequences and fine-tuned during actual tracking for online adaptation to a change of target and background. The pretraining is done by utilizing deep reinforcement learning (RL) as well as supervised learning. The use of RL enables even partially labeled data to be successfully utilized for semisupervised learning. Through the evaluation of the object tracking benchmark data set, the proposed tracker is validated to achieve a competitive performance at three times the speed of existing deep network-based trackers. The fast version of the proposed method, which operates in real time on graphics processing unit, outperforms the state-of-the-art real-time trackers with an accuracy improvement of more than 8%.

Sublayer-Specific Coding Dynamics during Spatial Navigation and Learning in Hippocampal Area CA1.

PubMed

Danielson, Nathan B; Zaremba, Jeffrey D; Kaifosh, Patrick; Bowler, John; Ladow, Max; Losonczy, Attila

2016-08-03

The mammalian hippocampus is critical for spatial information processing and episodic memory. Its primary output cells, CA1 pyramidal cells (CA1 PCs), vary in genetics, morphology, connectivity, and electrophysiological properties. It is therefore possible that distinct CA1 PC subpopulations encode different features of the environment and differentially contribute to learning. To test this hypothesis, we optically monitored activity in deep and superficial CA1 PCs segregated along the radial axis of the mouse hippocampus and assessed the relationship between sublayer dynamics and learning. Superficial place maps were more stable than deep during head-fixed exploration. Deep maps, however, were preferentially stabilized during goal-oriented learning, and representation of the reward zone by deep cells predicted task performance. These findings demonstrate that superficial CA1 PCs provide a more stable map of an environment, while their counterparts in the deep sublayer provide a more flexible representation that is shaped by learning about salient features in the environment. VIDEO ABSTRACT. Copyright © 2016 Elsevier Inc. All rights reserved.

Sr/Ca ratios in cold-water corals - a 'low-resolution' temperature archive?

NASA Astrophysics Data System (ADS)

Rüggeberg, Andres; Riethdorf, Jan-Rainer; Raddatz, Jacek; López Correa, Matthias; Montagna, Paolo; Dullo, Wolf-Christian; Freiwald, André

2010-05-01

One of the basic data to understand global change and past global changes is the measurement and the reconstruction of temperature of marine water masses. E.g. seawater temperature controls the density of seawater and in combination with salinity is the major driving force for the oceans circulation system. Geochemical investigations on cold-water corals Lophelia pertusa and Desmophyllum cristagalli indicated the potential of these organisms as high-resolution archives of environmental parameters from intermediate and deeper water masses (Adkins and Boyle 1997). Some studies tried to use cold-water corals as a high-resolution archive of temperature and salinity (Smith et al. 2000, 2002; Blamart et al. 2005; Lutringer et al. 2005). However, the fractionation of stable isotopes (delta18O and delta13C) and element ratios (Sr/Ca, Mg/Ca, U/Ca) are strongly influenced by vital effects (Shirai et al. 2005; Cohen et al. 2006), and difficult to interpret. Nevertheless, ongoing studies indicate the potential of a predominant temperature dependent fractionation of distinct isotopes and elements (e.g. Li/Ca, Montagna et al. 2008; U/Ca, Mg/Ca, delta18O, Lòpez Correa et al. 2008; delta88/86Sr, Rüggeberg et al. 2008). Within the frame of DFG-Project TRISTAN and Paläo-TRISTAN (Du 129/37-2 and 37-3) we investigated live-collected specimens of cold-water coral L. pertusa from all along the European continental margin (Northern and mid Norwegian shelves, Skagerrak, Rockall and Porcupine Bank, Galicia Bank, Gulf of Cadiz, Mediterranean Sea). These coral samples grew in waters characterized by temperatures between 6°C and 14°C. Electron Microprobe investigations along the growth direction of individual coral polyps were applied to determine the relationship between the incorporation of distinct elements (Sr, Ca, Mg, S). Cohen et al. (2006) showed for L. pertusa from the Kosterfjord, Skagerrak, that ~25% of the coral's Sr/Ca ratio is related to temperature, while 75% are influenced by the calcification rate of the organism. However, the Sr/Ca-temperature relation of our L. pertusa specimens suggest, that mean values are more reliable for temperature reconstruction along a larger temperature range than local high-resolution investigations. Additionally, our results plot on same line of Sr/Ca-temperature relationship like tropical corals indicating a similar behaviour of element incorporation during calcification. References: Adkins JF, Boyle EA (1997) Changing atmospheric ∆14C and the record of deep water paleoventilation ages. Paleoceanography 12:337-344 Blamart D, Rollion-Bard C, Cuif J-P, Juillet-Leclerc A, Lutringer A, Weering Tv, Henriet J-P (2005) C and O isotopes in a deep-sea coral (Lophelia pertusa) related to skeletal microstructure. In: Freiwald A, Roberts JM (eds) Cold-water Corals and Ecosystems. Springer-Verlag, Berlin Heidelberg, p 1005-1020 Cohen AL, Gaetani GA, Lundälv T, Corliss BH, George RY (2006) Compositional variability in a cold-water scleractinian, Lophelia pertusa: New insights into vital effects. Geochemistry, Geophysics, Geosystems 7:Q12004, doi:12010.11029/12006GC001354 López Correa M, Montagna P, Rüggeberg A, McCulloch M, Taviani M, Freiwald A (2008) Trace elements and stable isotopes in recent North Atlantic Lophelia pertusa along a latitudal gradient and from fossil Mediterranean sites. ASLO 2008 Summer Meeting, St. John's, Newfoundland & Labrador, Canada, 08.06.-13.06.2008, p. 47 Lutringer A, Blamart D, Frank N, Labeyrie L (2005) Paleotemperatures from deep-sea corals: scale effects. In: Freiwald A, Roberts JM (eds) Cold-water Corals and Ecosystems. Springer-Verlag, Berlin, Heidelberg, p 1081-1096 Montagna P, López-Correa M, Rüggeberg A, McCulloch M, Rodolfo-Metalpa R, Dullo W-C, Ferrier-Pagès C, Freiwald A, Henderson G, Mazzoli C, Russo S, Silenzi S, Taviani M (2008) Coral Li/Ca in micro-structural domains as a temperature proxy. Goldschmidt Conference, Vancouver, British Columbia, Canada Rüggeberg A, Fietzke J, Liebetrau V, Eisenhauer A, Dullo W-C, Freiwald A (2008) Stable strontium isotopes (delta88/86Sr) in cold-water corals — A new proxy for reconstruction of intermediate ocean water temperatures. Earth and Planetary Science Letters 269:569-574 Shirai K, Kusakabe M, Nakai S, Ishii T, Watanabe T, Hiyagon H, Sano Y (2005) Deep-sea coral geochemistry: Implication for the vital effect. Chemical Geology 224:212-222 Smith JE, Schwarcz HP, Risk MJ (2002) Patterns of isotopic disequilibria in azooxanhtellate coral skeletons. Hydrobiologia 471:111-115 Smith JE, Schwarcz HP, Risk MJ, McConnaughey TA, Keller N (2000) Paleotemperatures from deep-sea corals: Overcoming 'vital effects'. Palaios 15:25-32

The Use of Deep and Surface Learning Strategies among Students Learning English as a Foreign Language in an Internet Environment

ERIC Educational Resources Information Center

Aharony, Noa

2006-01-01

Background: The learning context is learning English in an Internet environment. The examination of this learning process was based on the Biggs and Moore's teaching-learning model (Biggs & Moore, 1993). Aim: The research aims to explore the use of the deep and surface strategies in an Internet environment among EFL students who come from…

Deep learning decision fusion for the classification of urban remote sensing data

NASA Astrophysics Data System (ADS)

Abdi, Ghasem; Samadzadegan, Farhad; Reinartz, Peter

2018-01-01

Multisensor data fusion is one of the most common and popular remote sensing data classification topics by considering a robust and complete description about the objects of interest. Furthermore, deep feature extraction has recently attracted significant interest and has become a hot research topic in the geoscience and remote sensing research community. A deep learning decision fusion approach is presented to perform multisensor urban remote sensing data classification. After deep features are extracted by utilizing joint spectral-spatial information, a soft-decision made classifier is applied to train high-level feature representations and to fine-tune the deep learning framework. Next, a decision-level fusion classifies objects of interest by the joint use of sensors. Finally, a context-aware object-based postprocessing is used to enhance the classification results. A series of comparative experiments are conducted on the widely used dataset of 2014 IEEE GRSS data fusion contest. The obtained results illustrate the considerable advantages of the proposed deep learning decision fusion over the traditional classifiers.

A deep learning framework for financial time series using stacked autoencoders and long-short term memory

PubMed Central

Bao, Wei; Rao, Yulei

2017-01-01

The application of deep learning approaches to finance has received a great deal of attention from both investors and researchers. This study presents a novel deep learning framework where wavelet transforms (WT), stacked autoencoders (SAEs) and long-short term memory (LSTM) are combined for stock price forecasting. The SAEs for hierarchically extracted deep features is introduced into stock price forecasting for the first time. The deep learning framework comprises three stages. First, the stock price time series is decomposed by WT to eliminate noise. Second, SAEs is applied to generate deep high-level features for predicting the stock price. Third, high-level denoising features are fed into LSTM to forecast the next day’s closing price. Six market indices and their corresponding index futures are chosen to examine the performance of the proposed model. Results show that the proposed model outperforms other similar models in both predictive accuracy and profitability performance. PMID:28708865

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.

Learning approach among health sciences students in a medical college in Nepal: a cross-sectional study.

PubMed

Shah, Dev Kumar; Yadav, Ram Lochan; Sharma, Deepak; Yadav, Prakash Kumar; Sapkota, Niraj Khatri; Jha, Rajesh Kumar; Islam, Md Nazrul

2016-01-01

Many factors shape the quality of learning. The intrinsically motivated students adopt a deep approach to learning, while students who fear failure in assessments adopt a surface approach to learning. In the area of health science education in Nepal, there is still a lack of studies on learning approach that can be used to transform the students to become better learners and improve the effectiveness of teaching. Therefore, we aimed to explore the learning approaches among medical, dental, and nursing students of Chitwan Medical College, Nepal using Biggs's Revised Two-Factor Study Process Questionnaire (R-SPQ-2F) after testing its reliability. R-SPQ-2F containing 20 items represented two main scales of learning approaches, deep and surface, with four subscales: deep motive, deep strategy, surface motive, and surface strategy. Each subscale had five items and each item was rated on a 5-point Likert scale. The data were analyzed using Student's t-test and analysis of variance. Reliability of the administered questionnaire was checked using Cronbach's alpha. The Cronbach's alpha value (0.6) for 20 items of R-SPQ-2F was found to be acceptable for its use. The participants predominantly had a deep approach to learning regardless of their age and sex (deep: 32.62±6.33 versus surface: 25.14±6.81, P<0.001). The level of deep approach among medical students (33.26±6.40) was significantly higher than among dental (31.71±6.51) and nursing (31.36±4.72) students. In comparison to first-year students, deep approach among second-year medical (34.63±6.51 to 31.73±5.93; P<0.001) and dental (33.47±6.73 to 29.09±5.62; P=0.002) students was found to be significantly decreased. On the other hand, surface approach significantly increased (25.55±8.19 to 29.34±6.25; P=0.023) among second-year dental students compared to first-year dental students. Medical students were found to adopt a deeper approach to learning than dental and nursing students. However, irrespective of disciplines and personal characteristics of participants, the primarily deep learning approach was found to be shifting progressively toward a surface approach after completion of an academic year, which should be avoided.

Transport of Sr 2+ and SrEDTA 2- in partially-saturated and heterogeneous sediments

NASA Astrophysics Data System (ADS)

Pace, M. N.; Mayes, M. A.; Jardine, P. M.; McKay, L. D.; Yin, X. L.; Mehlhorn, T. L.; Liu, Q.; Gürleyük, H.

2007-05-01

Strontium-90 has migrated deep into the unsaturated subsurface beneath leaking storage tanks in the Waste Management Areas (WMA) at the U.S. Department of Energy's (DOE) Hanford Reservation. Faster than expected transport of contaminants in the vadose zone is typically attributed to either physical hydrologic processes such as development of preferential flow pathways, or to geochemical processes such as the formation of stable, anionic complexes with organic chelates, e.g., ethylenediaminetetraacetic acid (EDTA). The goal of this paper is to determine whether hydrological processes in the Hanford sediments can influence the geochemistry of the system and hence control transport of Sr 2+ and SrEDTA 2-. The study used batch isotherms, saturated packed column experiments, and an unsaturated transport experiment in an undisturbed core. Isotherms and repacked column experiments suggested that the SrEDTA 2- complex was unstable in the presence of Hanford sediments, resulting in dissociation and transport of Sr 2+ as a divalent cation. A decrease in sorption with increasing solid:solution ratio for Sr 2+ and SrEDTA 2- suggested mineral dissolution resulted in competition for sorption sites and the formation of stable aqueous complexes. This was confirmed by detection of MgEDTA 2-, MnEDTA 2-, PbEDTA 2-, and unidentified Sr and Ca complexes. Displacement of Sr 2+ through a partially-saturated undisturbed core resulted in less retardation and more irreversible sorption than was observed in the saturated repacked columns, and model results suggested a significant reservoir (49%) of immobile water was present during transport through the heterogeneous layered sediments. The undisturbed core was subsequently disassembled along distinct bedding planes and subjected to sequential extractions. Strontium was unequally distributed between carbonates (49%), ion exchange sites (37%), and the oxide (14%) fraction. An inverse relationship between mass wetness and Sr suggested that sandy sediments of low water content constituted the immobile flow regime. Our results suggested that the sequestration of Sr 2+ in partially-saturated, heterogeneous sediments was most likely due to the formation of immobile water in drier regions having low hydraulic conductivities.

Measuring New-Voter Learning via Three Channels of Political Information.

ERIC Educational Resources Information Center

Martinelli, Kathleen A.; Chaffee, Steven H.

1995-01-01

Reports on a cross-sectional survey of newly naturalized United States citizens before the 1988 United States presidential election. Finds that newspaper, television news, and television ads each made a separate, significant contribution to issue learning. (SR)

Deep learning guided stroke management: a review of clinical applications.

PubMed

Feng, Rui; Badgeley, Marcus; Mocco, J; Oermann, Eric K

2018-04-01

Stroke is a leading cause of long-term disability, and outcome is directly related to timely intervention. Not all patients benefit from rapid intervention, however. Thus a significant amount of attention has been paid to using neuroimaging to assess potential benefit by identifying areas of ischemia that have not yet experienced cellular death. The perfusion-diffusion mismatch, is used as a simple metric for potential benefit with timely intervention, yet penumbral patterns provide an inaccurate predictor of clinical outcome. Machine learning research in the form of deep learning (artificial intelligence) techniques using deep neural networks (DNNs) excel at working with complex inputs. The key areas where deep learning may be imminently applied to stroke management are image segmentation, automated featurization (radiomics), and multimodal prognostication. The application of convolutional neural networks, the family of DNN architectures designed to work with images, to stroke imaging data is a perfect match between a mature deep learning technique and a data type that is naturally suited to benefit from deep learning's strengths. These powerful tools have opened up exciting opportunities for data-driven stroke management for acute intervention and for guiding prognosis. Deep learning techniques are useful for the speed and power of results they can deliver and will become an increasingly standard tool in the modern stroke specialist's arsenal for delivering personalized medicine to patients with ischemic stroke. © Article author(s) (or their employer(s) unless otherwise stated in the text of the article) 2018. All rights reserved. No commercial use is permitted unless otherwise expressly granted.

Geochemical constraints on depth of origin of oceanic carbonatites: The Cape Verde case

NASA Astrophysics Data System (ADS)

Doucelance, Régis; Hammouda, Tahar; Moreira, Manuel; Martins, João C.

2010-12-01

We present new Sr-Nd isotope compositions together with major- and trace element concentrations measured for whole rocks and mineral separate phases (apatite, biotite and calcite) from fifteen Cape Verde oceanic carbonatites (Atlantic Ocean). Trace element patterns of calcio- and magnesio-carbonatites present a strong depletion in K, Hf, Zr and Ti and an overall enrichment in Sr and REE relative to Cape Verde basalts, arguing for distinct source components between carbonatites and basalts. Sr and Nd isotopic ratios show small, but significant variations defining a binary mixing between a depleted end-member with unradiogenic Sr and radiogenic Nd values and a ''enriched'' end-member compatible with old marine carbonates. We interpret the depleted end-member as the Cape Verde oceanic lithosphere by comparison with previous studies on Cape Verde basalts. We thus propose that oceanic carbonatites are resulting from the interaction of a deep rooted mantle plume carrying a lower 4He/ 3He signature from the lower mantle and a carbonated metasomatized lithosphere, which by low degree melting produced carbonatite magmas. Sr-Nd compositions and trace element patterns of carbonatites argue in favor of a metasomatic agent originating from partial melting of recycled, carbonated oceanic crust. We have successfully reproduced the main geochemical features of this model using a Monte-Carlo-type simulation.

DeepARG: a deep learning approach for predicting antibiotic resistance genes from metagenomic data.

PubMed

Arango-Argoty, Gustavo; Garner, Emily; Pruden, Amy; Heath, Lenwood S; Vikesland, Peter; Zhang, Liqing

2018-02-01

Growing concerns about increasing rates of antibiotic resistance call for expanded and comprehensive global monitoring. Advancing methods for monitoring of environmental media (e.g., wastewater, agricultural waste, food, and water) is especially needed for identifying potential resources of novel antibiotic resistance genes (ARGs), hot spots for gene exchange, and as pathways for the spread of ARGs and human exposure. Next-generation sequencing now enables direct access and profiling of the total metagenomic DNA pool, where ARGs are typically identified or predicted based on the "best hits" of sequence searches against existing databases. Unfortunately, this approach produces a high rate of false negatives. To address such limitations, we propose here a deep learning approach, taking into account a dissimilarity matrix created using all known categories of ARGs. Two deep learning models, DeepARG-SS and DeepARG-LS, were constructed for short read sequences and full gene length sequences, respectively. Evaluation of the deep learning models over 30 antibiotic resistance categories demonstrates that the DeepARG models can predict ARGs with both high precision (> 0.97) and recall (> 0.90). The models displayed an advantage over the typical best hit approach, yielding consistently lower false negative rates and thus higher overall recall (> 0.9). As more data become available for under-represented ARG categories, the DeepARG models' performance can be expected to be further enhanced due to the nature of the underlying neural networks. Our newly developed ARG database, DeepARG-DB, encompasses ARGs predicted with a high degree of confidence and extensive manual inspection, greatly expanding current ARG repositories. The deep learning models developed here offer more accurate antimicrobial resistance annotation relative to current bioinformatics practice. DeepARG does not require strict cutoffs, which enables identification of a much broader diversity of ARGs. The DeepARG models and database are available as a command line version and as a Web service at http://bench.cs.vt.edu/deeparg .

Sparse Representations-Based Super-Resolution of Key-Frames Extracted from Frames-Sequences Generated by a Visual Sensor Network

PubMed Central

Sajjad, Muhammad; Mehmood, Irfan; Baik, Sung Wook

2014-01-01

Visual sensor networks (VSNs) usually generate a low-resolution (LR) frame-sequence due to energy and processing constraints. These LR-frames are not very appropriate for use in certain surveillance applications. It is very important to enhance the resolution of the captured LR-frames using resolution enhancement schemes. In this paper, an effective framework for a super-resolution (SR) scheme is proposed that enhances the resolution of LR key-frames extracted from frame-sequences captured by visual-sensors. In a VSN, a visual processing hub (VPH) collects a huge amount of visual data from camera sensors. In the proposed framework, at the VPH, key-frames are extracted using our recent key-frame extraction technique and are streamed to the base station (BS) after compression. A novel effective SR scheme is applied at BS to produce a high-resolution (HR) output from the received key-frames. The proposed SR scheme uses optimized orthogonal matching pursuit (OOMP) for sparse-representation recovery in SR. OOMP does better in terms of detecting true sparsity than orthogonal matching pursuit (OMP). This property of the OOMP helps produce a HR image which is closer to the original image. The K-SVD dictionary learning procedure is incorporated for dictionary learning. Batch-OMP improves the dictionary learning process by removing the limitation in handling a large set of observed signals. Experimental results validate the effectiveness of the proposed scheme and show its superiority over other state-of-the-art schemes. PMID:24566632

Sparse representations-based super-resolution of key-frames extracted from frames-sequences generated by a visual sensor network.

PubMed

Sajjad, Muhammad; Mehmood, Irfan; Baik, Sung Wook

2014-02-21

Visual sensor networks (VSNs) usually generate a low-resolution (LR) frame-sequence due to energy and processing constraints. These LR-frames are not very appropriate for use in certain surveillance applications. It is very important to enhance the resolution of the captured LR-frames using resolution enhancement schemes. In this paper, an effective framework for a super-resolution (SR) scheme is proposed that enhances the resolution of LR key-frames extracted from frame-sequences captured by visual-sensors. In a VSN, a visual processing hub (VPH) collects a huge amount of visual data from camera sensors. In the proposed framework, at the VPH, key-frames are extracted using our recent key-frame extraction technique and are streamed to the base station (BS) after compression. A novel effective SR scheme is applied at BS to produce a high-resolution (HR) output from the received key-frames. The proposed SR scheme uses optimized orthogonal matching pursuit (OOMP) for sparse-representation recovery in SR. OOMP does better in terms of detecting true sparsity than orthogonal matching pursuit (OMP). This property of the OOMP helps produce a HR image which is closer to the original image. The K-SVD dictionary learning procedure is incorporated for dictionary learning. Batch-OMP improves the dictionary learning process by removing the limitation in handling a large set of observed signals. Experimental results validate the effectiveness of the proposed scheme and show its superiority over other state-of-the-art schemes.

Dahl (S × R) rat congenic strain analysis confirms and defines a chromosome 17 spatial navigation quantitative trait locus to <10 Mbp.

PubMed

Herrera, Victoria L; Pasion, Khristine A; Tan, Glaiza A; Ruiz-Opazo, Nelson

2013-01-01

A quantitative trait locus (QTL) linked with ability to find a platform in the Morris Water Maze (MWM) was located on chromosome 17 (Nav-5 QTL) using intercross between Dahl S and Dahl R rats. We developed two congenic strains, S.R17A and S.R17B introgressing Dahl R-chromosome 17 segments into Dahl S chromosome 17 region spanning putative Nav-5 QTL. Performance analysis of S.R17A, S.R17B and Dahl S rats in the Morris water maze (MWM) task showed a significantly decreased spatial navigation performance in S.R17B congenic rats when compared with Dahl S controls (P = 0.02). The S.R17A congenic segment did not affect MWM performance delimiting Nav-5 to the chromosome 17 65.02-74.66 Mbp region. Additional fine mapping is necessary to identify the specific gene variant accounting for Nav-5 effect on spatial learning and memory in Dahl rats.

Fifty years of computer analysis in chest imaging: rule-based, machine learning, deep learning.

PubMed

van Ginneken, Bram

2017-03-01

Half a century ago, the term "computer-aided diagnosis" (CAD) was introduced in the scientific literature. Pulmonary imaging, with chest radiography and computed tomography, has always been one of the focus areas in this field. In this study, I describe how machine learning became the dominant technology for tackling CAD in the lungs, generally producing better results than do classical rule-based approaches, and how the field is now rapidly changing: in the last few years, we have seen how even better results can be obtained with deep learning. The key differences among rule-based processing, machine learning, and deep learning are summarized and illustrated for various applications of CAD in the chest.

Epilogue: A Speech by Ben Cameron.

ERIC Educational Resources Information Center

Cameron, Ben

1999-01-01

Reflects on the state of the arts in general and theatre arts in particular. Examines the role of controversy in the arts; discusses "elitism," issues of diversity and being committed to community. Argues that deep, substantive discussions are needed on the value to society of the arts and the values that the arts uphold and exemplify. (SR)

Quantum neuromorphic hardware for quantum artificial intelligence

NASA Astrophysics Data System (ADS)

Prati, Enrico

2017-08-01

The development of machine learning methods based on deep learning boosted the field of artificial intelligence towards unprecedented achievements and application in several fields. Such prominent results were made in parallel with the first successful demonstrations of fault tolerant hardware for quantum information processing. To which extent deep learning can take advantage of the existence of a hardware based on qubits behaving as a universal quantum computer is an open question under investigation. Here I review the convergence between the two fields towards implementation of advanced quantum algorithms, including quantum deep learning.

Plume-related mantle source of super-large rare metal deposits from the Lovozero and Khibina massifs on the Kola Peninsula, Eastern part of Baltic Shield: Sr, Nd and Hf isotope systematics

NASA Astrophysics Data System (ADS)

Kogarko, L. N.; Lahaye, Y.; Brey, G. P.

2010-03-01

The two world’s largest complexes of highly alkaline nepheline syenites and related rare metal loparite and eudialyte deposits, the Khibina and Lovozero massifs, occur in the central part of the Kola Peninsula. We measured for the first time in situ the trace element concentrations and the Sr, Nd and Hf isotope ratios by LA-ICP-MS (laser ablation inductively coupled plasma mass spectrometer) in loparite, eudialyte an in some other pegmatitic minerals. The results are in aggreement with the whole rock Sr and Nd isotope which suggests the formation of these superlarge rare metal deposits in a magmatic closed system. The initial Hf, Sr, Nd isotope ratios are similar to the isotopic signatures of OIB indicating depleted mantle as a source. This leads to the suggestion that the origin of these gigantic alkaline intrusions is connected to a deep seated mantle source—possibly to a lower mantle plume. The required combination of a depleted mantle and high rare metal enrichment in the source can be explained by the input of incompatible elements by metasomatising melts/fluids into the zones of alkaline magma generation shortly before the partial melting event (to avoid ingrowth of radiogenic isotopes). The minerals belovite and pyrochlore from the pegmatites are abnormally high in 87Sr /86Sr ratios. This may be explained by closed system isotope evolution as a result of a significant increase in Rb/Sr during the evolution of the peralkaline magma.

A Comparative Study of Learning Strategies Used by Romanian and Hungarian Preuniversity Students in Science Learning

ERIC Educational Resources Information Center

Lingvay, Mónika; Timofte, Roxana S.; Ciascai, Liliana; Predescu, Constantin

2015-01-01

Development of pupils' deep learning approach is an important goal of education nowadays, considering that a deep learning approach is mediating conceptual understanding and transfer. Different performance at PISA tests of Romanian and Hungarian pupils cause us to commence a study for the analysis of learning approaches employed by these pupils.…

Making Information Systems Less Scrugged: Reflecting on the Processes of Change in Teaching and Learning

ERIC Educational Resources Information Center

Houghton, Luke; Ruth, Alison

2010-01-01

Deep and shallow learner approaches are useful for different purposes. Shallow learning can be good where fact memorization is appropriate, learning how to swim or play the guitar for example. Deep learning is much more appropriate when the learning material present involves going beyond simple facts and into what lies below the surface. When…

Europium and strontium anomalies in the MORB source mantle

NASA Astrophysics Data System (ADS)

Tang, Ming; McDonough, William F.; Ash, Richard D.

2017-01-01

Lower crustal recycling depletes the continental crust of Eu and Sr and returns Eu and Sr enriched materials into the mantle (e.g., Tang et al., 2015, Geology). To test the hypothesis that the MORB source mantle balances the Eu and Sr deficits in the continental crust, we carried out high precision Eu/Eu∗ and Sr/Sr∗ measurement for 72 MORB glasses with MgO >8.5% from the Pacific, Indian, and Atlantic mid-ocean ridges. MORB glasses with MgO ⩾ 9 wt.% have a mean Eu/Eu∗ of 1.025 ± 0.025 (2 σm, n = 46) and Sr/Sr∗ of 1.242 ± 0.093 (2 σm, n = 41) and these ratios are positively correlated. These samples show both positive and negative Eu and Sr anomalies, with no correlations between Eu/Eu∗ vs. MgO or Sr/Sr∗ vs. MgO, suggesting that the anomalies are not produced by plagioclase fractionation at MgO >9 wt.% and, thus, other processes must be responsible for generating the anomalies. We term these MORB samples primitive MORBs, as they record the melt Eu/Eu∗ and Sr/Sr∗ before plagioclase fractionation. Consequently, the mean oceanic crust, including cumulates, has a bulk Eu/Eu∗ of ∼1 and 20% Sr excess. Considering that divalent Sr and Eu(II) diffuse faster than trivalent Pr, Nd, Sm, and Gd, we evaluated this kinetic effect on Sm-Eu-Gd and Pr-Sr-Nd fractionations during spinel peridotite partial melting in the MORB source mantle. Our modeling shows that the correlated Eu and Sr anomalies seen in primitive MORBs may result from disequilibrium mantle melting. Melt fractions produced during early- and late-stage melting may carry positive and negative Eu and Sr anomalies, respectively, that overlap with the ranges documented in primitive MORBs. Because the net effect of disequilibrium melting is to produce partial melts with bulk positive Eu and Sr anomalies, the MORB source mantle must have Eu/Eu∗ < 1.025 ± 0.025 (2 σm) and Sr/Sr∗ < 1.242 ± 0.093 (2 σm). Although we cannot rule out the possibility that recycled lower continental crustal materials, which have positive Eu and Sr anomalies, are partially mixed into the upper mantle (i.e., MORB source region), a significant amount of this crustal component must have been sequestered into the deep mantle, as supported by the negative 206Pb/204Pb-Eu/Eu∗ and 206Pb/204Pb-Sr/Sr∗ correlations in ocean island basalts.

Recent paleoseismicity record in Prince William Sound, Alaska, USA

NASA Astrophysics Data System (ADS)

Kuehl, Steven A.; Miller, Eric J.; Marshall, Nicole R.; Dellapenna, Timothy M.

2017-12-01

Sedimentological and geochemical investigation of sediment cores collected in the deep (>400 m) central basin of Prince William Sound, along with geochemical fingerprinting of sediment source areas, are used to identify earthquake-generated sediment gravity flows. Prince William Sound receives sediment from two distinct sources: from offshore (primarily Copper River) through Hinchinbrook Inlet, and from sources within the Sound (primarily Columbia Glacier). These sources are found to have diagnostic elemental ratios indicative of provenance; Copper River Basin sediments were significantly higher in Sr/Pb and Cu/Pb, whereas Prince William Sound sediments were significantly higher in K/Ca and Rb/Sr. Within the past century, sediment gravity flows deposited within the deep central channel of Prince William Sound have robust geochemical (provenance) signatures that can be correlated with known moderate to large earthquakes in the region. Given the thick Holocene sequence in the Sound ( 200 m) and correspondingly high sedimentation rates (>1 cm year-1), this relationship suggests that sediments within the central basin of Prince William Sound may contain an extraordinary high-resolution record of paleoseismicity in the region.

Retrieval practice improves memory in multiple sclerosis: clinical application of the testing effect.

PubMed

Sumowski, James F; Chiaravalloti, Nancy; Deluca, John

2010-03-01

The testing effect is a robust cognitive phenomenon by which memory retrieval on a test improves subsequent recall more than restudying. Also known as retrieval practice, the testing effect has been studied almost exclusively in healthy undergraduates. The current study investigated whether retrieval practice during testing leads to better delayed recall than restudy among persons with multiple sclerosis (MS), a neurologic disease associated with memory dysfunction. In a within-subjects design, 32 persons with MS and 16 demographically matched healthy controls (HC) studied 48 verbal paired associates (VPA) divided across 3 learning conditions: massed restudy (MR), spaced restudy (SR), and spaced testing (ST). Delayed VPA cued recall was measured after 45 min. There was a large main effect of learning condition (etap2 = .54, p < .001) such that both MS and HC participants produced better delayed recall for VPAs learned through ST relative to MR and SR; and SR relative to MR. This same pattern was observed for MS participants with objective memory impairment (n = 16), thereby providing the first evidence that retrieval practice improves memory more than restudy among persons with neurologically based memory impairment. Copyright 2010 APA, all rights reserved

Is the University System in Australia Producing Deep Thinkers?

ERIC Educational Resources Information Center

Lake, Warren W.; Boyd, William E.

2015-01-01

Teaching and learning research since the 1980s has established a trend in students' learning approach tendencies, characterised by decreasing surface learning and increasing deep learning with increasing age. This is an important trend in higher education, especially at a time of increasing numbers of older students: are we graduating more deep…

How Enterprise Education Can Promote Deep Learning to Improve Student Employability

ERIC Educational Resources Information Center

Moon, Rob; Curtis, Vic; Dupernex, Simon

2013-01-01

This paper focuses on identifying the approaches students take to their learning, with particular regard to issues of enterprise, entrepreneurship and innovation when comparing the traditional lecture format to a more applied, practice-based case study format. The notions of deep and surface learning are used to explain student learning. More…

Emotion and the Internet: A Model of Learning

ERIC Educational Resources Information Center

Tran, Thuhang T.; Ward, Cheryl B.

2005-01-01

This conceptual paper examines the link between emotion and surface-deep learning in the context of the international business curriculum. We propose that 1) emotion and learning have a curvilinear relationship, and 2) the reflective abilities and attitude transformations related to deep-level learning can only arise if the student is emotionally…

Changing Students' Approaches to Learning: A Two-Year Study within a University Teacher Training Course

ERIC Educational Resources Information Center

Gijbels, David; Coertjens, Liesje; Vanthournout, Gert; Struyf, Elke; Van Petegem, Peter

2009-01-01

Inciting a deep approach to learning in students is difficult. The present research poses two questions: can a constructivist learning-assessment environment change students' approaches towards a more deep approach? What effect does additional feedback have on the changes in learning approaches? Two cohorts of students completed questionnaires…

Zircon U-Pb ages and geochemistry of migmatites and granites in the Foping dome: Evidence for Late Triassic crustal evolution in South Qinling, China

NASA Astrophysics Data System (ADS)

Zhang, He; Li, Shuang-Qing; Fang, Bo-Wen; He, Jian-Feng; Xue, Ying-Yu; Siebel, Wolfgang; Chen, Fukun

2018-01-01

Migmatites provide a record of melt formation and crustal rheology. In this study we present zircon U-Pb ages and geochemical composition of migmatites from the Foping dome and granites from the Wulong pluton. U-Pb results from migmatite zircons indicate two episodes of partial melting. Rim domains from a leucosome in the Longcaoping area yield an age of ca. 209 Ma. Migmatites collected from the Foping dome yield U-Pb zircon ages of 2910 to 190 Ma, suggesting the involvement of meta-sedimentary source components. Rim domains of the zircons with low Th/U ratios (< 0.1) give ages of 225-190 Ma and the youngest age domains (ca. 195 Ma) are characterized by low contents of heavy rare earth elements, which is related to crystallization of garnet. Magmatic rocks from the Wulong pluton can be subdivided into high Sr/Y and low Sr/Y granites. U-Pb zircon ages vary from 219 to 214 Ma for the high Sr/Y granites and from 214 to 192 Ma for the low Sr/Y granites. High Sr/Y granites have higher Na2O and Sr contents than the low Sr/Y granites. They also lack negative Eu anomalies and are depleted in HREE compared to the low Sr/Y granites. Initial 87Sr/86Sr ratios and εNd values of all the samples roughly overlap with those of Neoproterozoic basement rocks exposed in South Qinling. Including previous studies, we propose that the high and low Sr/Y granites formed by melting of thickened and normal crust, respectively. Close temporal-spatial relationship of the high and low Sr/Y granites with the two-stage migmatization events implies variation of crustal thickness and thermal overprints of the orogenic crust in post-collisional collapse. Following the collision of South Qinling and the Yangtze block prior to 219 Ma, partial melting of the deep crust occurred. The melts migrated upwards to form the high Sr/Y granites. This process occurred rapidly and caused collapse of the thickened crust and carried heat upwards, leading to further partial melting within the shallower crust and formation of the low Sr/Y granites.

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.

Students' Approaches to Learning in Problem-Based Learning: Taking into Account Professional Behavior in the Tutorial Groups, Self-Study Time, and Different Assessment Aspects

ERIC Educational Resources Information Center

Loyens, Sofie M. M.; Gijbels, David; Coertjens, Liesje; Cote, Daniel J.

2013-01-01

Problem-based learning (PBL) represents a major development in higher educational practice and is believed to promote deep learning in students. However, empirical findings on the promotion of deep learning in PBL remain unclear. The aim of the present study is to investigate the relationships between students' approaches to learning (SAL) and…

Iterative deep convolutional encoder-decoder network for medical image segmentation.

PubMed

Jung Uk Kim; Hak Gu Kim; Yong Man Ro

2017-07-01

In this paper, we propose a novel medical image segmentation using iterative deep learning framework. We have combined an iterative learning approach and an encoder-decoder network to improve segmentation results, which enables to precisely localize the regions of interest (ROIs) including complex shapes or detailed textures of medical images in an iterative manner. The proposed iterative deep convolutional encoder-decoder network consists of two main paths: convolutional encoder path and convolutional decoder path with iterative learning. Experimental results show that the proposed iterative deep learning framework is able to yield excellent medical image segmentation performances for various medical images. The effectiveness of the proposed method has been proved by comparing with other state-of-the-art medical image segmentation methods.

Deep ECGNet: An Optimal Deep Learning Framework for Monitoring Mental Stress Using Ultra Short-Term ECG Signals.

PubMed

Hwang, Bosun; You, Jiwoo; Vaessen, Thomas; Myin-Germeys, Inez; Park, Cheolsoo; Zhang, Byoung-Tak

2018-02-08

Stress recognition using electrocardiogram (ECG) signals requires the intractable long-term heart rate variability (HRV) parameter extraction process. This study proposes a novel deep learning framework to recognize the stressful states, the Deep ECGNet, using ultra short-term raw ECG signals without any feature engineering methods. The Deep ECGNet was developed through various experiments and analysis of ECG waveforms. We proposed the optimal recurrent and convolutional neural networks architecture, and also the optimal convolution filter length (related to the P, Q, R, S, and T wave durations of ECG) and pooling length (related to the heart beat period) based on the optimization experiments and analysis on the waveform characteristics of ECG signals. The experiments were also conducted with conventional methods using HRV parameters and frequency features as a benchmark test. The data used in this study were obtained from Kwangwoon University in Korea (13 subjects, Case 1) and KU Leuven University in Belgium (9 subjects, Case 2). Experiments were designed according to various experimental protocols to elicit stressful conditions. The proposed framework to recognize stress conditions, the Deep ECGNet, outperformed the conventional approaches with the highest accuracy of 87.39% for Case 1 and 73.96% for Case 2, respectively, that is, 16.22% and 10.98% improvements compared with those of the conventional HRV method. We proposed an optimal deep learning architecture and its parameters for stress recognition, and the theoretical consideration on how to design the deep learning structure based on the periodic patterns of the raw ECG data. Experimental results in this study have proved that the proposed deep learning model, the Deep ECGNet, is an optimal structure to recognize the stress conditions using ultra short-term ECG data.

Learning better deep features for the prediction of occult invasive disease in ductal carcinoma in situ through transfer learning

NASA Astrophysics Data System (ADS)

Shi, Bibo; Hou, Rui; Mazurowski, Maciej A.; Grimm, Lars J.; Ren, Yinhao; Marks, Jeffrey R.; King, Lorraine M.; Maley, Carlo C.; Hwang, E. Shelley; Lo, Joseph Y.

2018-02-01

Purpose: To determine whether domain transfer learning can improve the performance of deep features extracted from digital mammograms using a pre-trained deep convolutional neural network (CNN) in the prediction of occult invasive disease for patients with ductal carcinoma in situ (DCIS) on core needle biopsy. Method: In this study, we collected digital mammography magnification views for 140 patients with DCIS at biopsy, 35 of which were subsequently upstaged to invasive cancer. We utilized a deep CNN model that was pre-trained on two natural image data sets (ImageNet and DTD) and one mammographic data set (INbreast) as the feature extractor, hypothesizing that these data sets are increasingly more similar to our target task and will lead to better representations of deep features to describe DCIS lesions. Through a statistical pooling strategy, three sets of deep features were extracted using the CNNs at different levels of convolutional layers from the lesion areas. A logistic regression classifier was then trained to predict which tumors contain occult invasive disease. The generalization performance was assessed and compared using repeated random sub-sampling validation and receiver operating characteristic (ROC) curve analysis. Result: The best performance of deep features was from CNN model pre-trained on INbreast, and the proposed classifier using this set of deep features was able to achieve a median classification performance of ROC-AUC equal to 0.75, which is significantly better (p<=0.05) than the performance of deep features extracted using ImageNet data set (ROCAUC = 0.68). Conclusion: Transfer learning is helpful for learning a better representation of deep features, and improves the prediction of occult invasive disease in DCIS.

Wishart Deep Stacking Network for Fast POLSAR Image Classification.

PubMed

Jiao, Licheng; Liu, Fang

2016-05-11

Inspired by the popular deep learning architecture - Deep Stacking Network (DSN), a specific deep model for polarimetric synthetic aperture radar (POLSAR) image classification is proposed in this paper, which is named as Wishart Deep Stacking Network (W-DSN). First of all, a fast implementation of Wishart distance is achieved by a special linear transformation, which speeds up the classification of POLSAR image and makes it possible to use this polarimetric information in the following Neural Network (NN). Then a single-hidden-layer neural network based on the fast Wishart distance is defined for POLSAR image classification, which is named as Wishart Network (WN) and improves the classification accuracy. Finally, a multi-layer neural network is formed by stacking WNs, which is in fact the proposed deep learning architecture W-DSN for POLSAR image classification and improves the classification accuracy further. In addition, the structure of WN can be expanded in a straightforward way by adding hidden units if necessary, as well as the structure of the W-DSN. As a preliminary exploration on formulating specific deep learning architecture for POLSAR image classification, the proposed methods may establish a simple but clever connection between POLSAR image interpretation and deep learning. The experiment results tested on real POLSAR image show that the fast implementation of Wishart distance is very efficient (a POLSAR image with 768000 pixels can be classified in 0.53s), and both the single-hidden-layer architecture WN and the deep learning architecture W-DSN for POLSAR image classification perform well and work efficiently.

Deep generative learning for automated EHR diagnosis of traditional Chinese medicine.

PubMed

Liang, Zhaohui; Liu, Jun; Ou, Aihua; Zhang, Honglai; Li, Ziping; Huang, Jimmy Xiangji

2018-05-04

Computer-aided medical decision-making (CAMDM) is the method to utilize massive EMR data as both empirical and evidence support for the decision procedure of healthcare activities. Well-developed information infrastructure, such as hospital information systems and disease surveillance systems, provides abundant data for CAMDM. However, the complexity of EMR data with abstract medical knowledge makes the conventional model incompetent for the analysis. Thus a deep belief networks (DBN) based model is proposed to simulate the information analysis and decision-making procedure in medical practice. The purpose of this paper is to evaluate a deep learning architecture as an effective solution for CAMDM. A two-step model is applied in our study. At the first step, an optimized seven-layer deep belief network (DBN) is applied as an unsupervised learning algorithm to perform model training to acquire feature representation. Then a support vector machine model is adopted to DBN at the second step of the supervised learning. There are two data sets used in the experiments. One is a plain text data set indexed by medical experts. The other is a structured dataset on primary hypertension. The data are randomly divided to generate the training set for the unsupervised learning and the testing set for the supervised learning. The model performance is evaluated by the statistics of mean and variance, the average precision and coverage on the data sets. Two conventional shallow models (support vector machine / SVM and decision tree / DT) are applied as the comparisons to show the superiority of our proposed approach. The deep learning (DBN + SVM) model outperforms simple SVM and DT on two data sets in terms of all the evaluation measures, which confirms our motivation that the deep model is good at capturing the key features with less dependence when the index is built up by manpower. Our study shows the two-step deep learning model achieves high performance for medical information retrieval over the conventional shallow models. It is able to capture the features of both plain text and the highly-structured database of EMR data. The performance of the deep model is superior to the conventional shallow learning models such as SVM and DT. It is an appropriate knowledge-learning model for information retrieval of EMR system. Therefore, deep learning provides a good solution to improve the performance of CAMDM systems. Copyright © 2018. Published by Elsevier B.V.

A bioavailable strontium isoscape for Western Europe: A machine learning approach

PubMed Central

von Holstein, Isabella C. C.; Laffoon, Jason E.; Willmes, Malte; Liu, Xiao-Ming; Davies, Gareth R.

2018-01-01

Strontium isotope ratios (87Sr/86Sr) are gaining considerable interest as a geolocation tool and are now widely applied in archaeology, ecology, and forensic research. However, their application for provenance requires the development of baseline models predicting surficial 87Sr/86Sr variations (“isoscapes”). A variety of empirically-based and process-based models have been proposed to build terrestrial 87Sr/86Sr isoscapes but, in their current forms, those models are not mature enough to be integrated with continuous-probability surface models used in geographic assignment. In this study, we aim to overcome those limitations and to predict 87Sr/86Sr variations across Western Europe by combining process-based models and a series of remote-sensing geospatial products into a regression framework. We find that random forest regression significantly outperforms other commonly used regression and interpolation methods, and efficiently predicts the multi-scale patterning of 87Sr/86Sr variations by accounting for geological, geomorphological and atmospheric controls. Random forest regression also provides an easily interpretable and flexible framework to integrate different types of environmental auxiliary variables required to model the multi-scale patterning of 87Sr/86Sr variability. The method is transferable to different scales and resolutions and can be applied to the large collection of geospatial data available at local and global levels. The isoscape generated in this study provides the most accurate 87Sr/86Sr predictions in bioavailable strontium for Western Europe (R2 = 0.58 and RMSE = 0.0023) to date, as well as a conservative estimate of spatial uncertainty by applying quantile regression forest. We anticipate that the method presented in this study combined with the growing numbers of bioavailable 87Sr/86Sr data and satellite geospatial products will extend the applicability of the 87Sr/86Sr geo-profiling tool in provenance applications. PMID:29847595

Long-Lived Mantle Plumes Sample Multiple Deep Mantle Geochemical Domains: The Example of the Hawaiian-Emperor Chain

NASA Astrophysics Data System (ADS)

Harrison, L.; Weis, D.

2017-12-01

Oceanic island basalts provide the opportunity for the geochemist to study the deep mantle source removed from continental sources of contamination and, for long-lived systems, the evolution of mantle sources with time. In the case of the Hawaiian-Emperor (HE) chain, formation by a long-lived (>81 Myr), deeply-sourced mantle plume allows for insight into plume dynamics and deep mantle geochemistry. The geochemical record of the entire chain is now complete with analysis of Pb-Hf-Nd-Sr isotopes and elemental compositions of the Northwest Hawaiian Ridge (NWHR), which consists of 51 volcanoes spanning 42 Ma between the bend in the chain and the Hawaiian Islands. This segment of the chain previously represented a significant data gap where Hawaiian plume geochemistry changed markedly, along with magmatic flux: only Kea compositions have been observed on Emperor seamounts (>50 Ma), whereas the Hawaiian Islands (<6 Ma) present both Kea and Loa compositions. A database of 700 Hawaiian Island shield basalts Pb-Hf-Nd-Sr isotopic compositions were compiled to construct a logistical regression model of Loa or Kea affinity that sorts data into a dichotomous category and provides insight into the relationship between independent variables. We use this model to predict whether newly analyzed NWHR samples are Loa or Kea composition based on their Pb-Sr-Nd-Hf isotopic compositions. The logistical regression model is significantly better at prediciting Loa or Kea affinity than the constant only model (χ2=263.3, df=4, p<0.0001), with Pb and Sr isotopes providing the most predicitive power. Daikakuji, West Nihoa, Nihoa, and Mokumanamana erupt Loa-type lavas, suggesting that the Loa source is sampled ephemerally during the NWHR and increases in presence and volume towards the younger section of the NWHR (younger than Midway 20-25 Ma). These results complete the picture of Hawaiian mantle plume geochemistry and geodynamics for 81 Myr, and show that the Hawaiian mantle plume has transitioned from a dominately Kea source during the Emperor seamounts and older NWHR to an increasingly enriched Loa source from the mid NWHR to Hawaiian Islands. We propose this is due to Hawaiian mantle plume drift through different lower mantle geohemical domains.

A comparative study of two prediction models for brain tumor progression

NASA Astrophysics Data System (ADS)

Zhou, Deqi; Tran, Loc; Wang, Jihong; Li, Jiang

2015-03-01

MR diffusion tensor imaging (DTI) technique together with traditional T1 or T2 weighted MRI scans supplies rich information sources for brain cancer diagnoses. These images form large-scale, high-dimensional data sets. Due to the fact that significant correlations exist among these images, we assume low-dimensional geometry data structures (manifolds) are embedded in the high-dimensional space. Those manifolds might be hidden from radiologists because it is challenging for human experts to interpret high-dimensional data. Identification of the manifold is a critical step for successfully analyzing multimodal MR images. We have developed various manifold learning algorithms (Tran et al. 2011; Tran et al. 2013) for medical image analysis. This paper presents a comparative study of an incremental manifold learning scheme (Tran. et al. 2013) versus the deep learning model (Hinton et al. 2006) in the application of brain tumor progression prediction. The incremental manifold learning is a variant of manifold learning algorithm to handle large-scale datasets in which a representative subset of original data is sampled first to construct a manifold skeleton and remaining data points are then inserted into the skeleton by following their local geometry. The incremental manifold learning algorithm aims at mitigating the computational burden associated with traditional manifold learning methods for large-scale datasets. Deep learning is a recently developed multilayer perceptron model that has achieved start-of-the-art performances in many applications. A recent technique named "Dropout" can further boost the deep model by preventing weight coadaptation to avoid over-fitting (Hinton et al. 2012). We applied the two models on multiple MRI scans from four brain tumor patients to predict tumor progression and compared the performances of the two models in terms of average prediction accuracy, sensitivity, specificity and precision. The quantitative performance metrics were calculated as average over the four patients. Experimental results show that both the manifold learning and deep neural network models produced better results compared to using raw data and principle component analysis (PCA), and the deep learning model is a better method than manifold learning on this data set. The averaged sensitivity and specificity by deep learning are comparable with these by the manifold learning approach while its precision is considerably higher. This means that the predicted abnormal points by deep learning are more likely to correspond to the actual progression region.

Revealing magma degassing below closed-conduit active volcanoes: Geochemical features of volcanic rocks versus fumarolic fluids at Vulcano (Aeolian Islands, Italy)

NASA Astrophysics Data System (ADS)

Mandarano, Michela; Paonita, Antonio; Martelli, Mauro; Viccaro, Marco; Nicotra, Eugenio; Millar, Ian L.

2016-04-01

The elemental and isotopic compositions of noble gases (He, Ne, and Ar) in olivine- and clinopyroxene-hosted fluid inclusions have been measured for rocks at various degrees of evolution and belonging to high-K calcalkaline-shoshonitic and shoshonitic-potassic series in order to cover the entire volcanological history of Vulcano Island (Italy). The major- and trace-element concentrations and the Sr- and Pb-isotope compositions for whole rocks were integrated with data obtained from the fluid inclusions. 3He/4He in fluid inclusions is within the range of 3.30 and 5.94 R/Ra, being lower than the theoretical value for the deep magmatic source expected for Vulcano Island (6.0-6.2 R/Ra). 3He/4He of the magmatic source is almost constant throughout the volcanic history of Vulcano. Integration of the He- and Sr-isotope systematics leads to the conclusion that a decrease in the He-isotope ratio of the rocks is mainly due to the assimilation of 10-25% of a crustal component similar to the Calabrian basement. 3He/4He shows a negative correlation with Sr isotopes except for the last-erupted Vulcanello latites (Punta del Roveto), which have anomalously high He isotope ratios. This anomaly has been attributed to a flushing process by fluids coming from the deepest reservoirs, since an input of deep magmatic volatiles with high 3He/4He values increases the He-isotope ratio without changing 87Sr/86Sr. A comparison of the He-isotope ratios between fluid inclusions and fumarolic gases shows that only the basalts of La Sommata and the latites of Vulcanello have comparable values. Taking into account that the latites of Vulcanello relate to one of the most-recent eruptions at Vulcano (in the 17th century), we infer that the most probable magma which actually feeds the fumarolic emissions is a latitic body that ponded at about 3-3.5 km of depth and is flushed by fluids coming from a deeper and basic magma.

Geochemical evolution of groundwater salinity at basin scale: a case study from Datong basin, Northern China.

PubMed

Wu, Ya; Wang, Yanxin

2014-05-01

A hydrogeochemical investigation using integrated methods of stable isotopes ((18)O, (2)H), (87)Sr/(86)Sr ratios, Cl/Br ratios, chloride-mass balance, mass balance and hydrogeochemical modeling was conducted to interpret the geochemical evolution of groundwater salinity in Datong basin, northern China. The δ(2)H, δ(18)O ratios in precipitation exhibited a local meteoric water line of δ(2)H = 6.4 δ(18)O -5 (R(2) = 0.94), while those in groundwater suggested their meteoric origin in a historically colder climatic regime with a speculated recharge rate of less than 20.5 mm overall per year, in addition to recharge from a component of deep residual ancient lake water enriched with Br. According to the Sr isotope binary mixing model, the mixing of recharges from the Shentou karst springs (24%), the western margins (11%) and the eastern margins (65%) accounts for the groundwater from the deep aquifers of the down-gradient parts in the central basin is a possible mixing mechanism. In Datong, hydrolysis of silicate minerals is the most important hydrogeochemical process responsible for groundwater chemistry, in addition to dissolution of carbonate and evaporites. In the recharge areas, silicate chemical weathering is typically at the bisiallitization stage, while that in the central basin is mostly at the monosiallitization stage with limited evidence of being in equilibrium with gibbsite. Na exchange with bound Ca, Mg prevails at basin scale, and intensifies with groundwater salinity, while Ca, Mg exchange with bound Na locally occurs in the east pluvial and alluvial plains. Although groundwater salinity increases with the progress of water-rock/sediment interactions along the flow path, as a result of carbonate solubility control and continuous evapotranspiration, Na-HCO3 and Na-Cl-SO4 types of water are usually characterized respectively in the deep and the shallow aquifers of an inland basin with a silicate terrain in an arid climatic regime.

Two novel nonlinear optical carbonates in the deep-ultraviolet region: KBeCO3F and RbAlCO3F2

PubMed Central

Kang, Lei; Lin, Zheshuai; Qin, Jingui; Chen, Chuangtian

2013-01-01

With the rapid developments of the all-solid-state deep-ultraviolet (deep-UV) lasers, the good nonlinear optical (NLO) crystal applied in this spectral region is currently lacking. Here, we design two novel NLO carbonates KBeCO3F and RbAlCO3F2 from the first-principles theory implemented in the molecular engineering expert system especially for NLO crystals. Both structurally stable crystals possess very large energy band gaps and optical anisotropy, so they would become the very promising deep-UV NLO crystals alternative to KBBF. Recent experimental results on MNCO3F (M = K, Rb, Cs; N = Ca, Sr, Ba) not only confirm our calculations, but also suggest that the synthesis of the KBeCO3F and RbAlCO3F2 crystals is feasible. PMID:23455618

Evaluation of Deep Learning Based Stereo Matching Methods: from Ground to Aerial Images

NASA Astrophysics Data System (ADS)

Liu, J.; Ji, S.; Zhang, C.; Qin, Z.

2018-05-01

Dense stereo matching has been extensively studied in photogrammetry and computer vision. In this paper we evaluate the application of deep learning based stereo methods, which were raised from 2016 and rapidly spread, on aerial stereos other than ground images that are commonly used in computer vision community. Two popular methods are evaluated. One learns matching cost with a convolutional neural network (known as MC-CNN); the other produces a disparity map in an end-to-end manner by utilizing both geometry and context (known as GC-net). First, we evaluate the performance of the deep learning based methods for aerial stereo images by a direct model reuse. The models pre-trained on KITTI 2012, KITTI 2015 and Driving datasets separately, are directly applied to three aerial datasets. We also give the results of direct training on target aerial datasets. Second, the deep learning based methods are compared to the classic stereo matching method, Semi-Global Matching(SGM), and a photogrammetric software, SURE, on the same aerial datasets. Third, transfer learning strategy is introduced to aerial image matching based on the assumption of a few target samples available for model fine tuning. It experimentally proved that the conventional methods and the deep learning based methods performed similarly, and the latter had greater potential to be explored.

Deep Learning for Automated Extraction of Primary Sites From Cancer Pathology Reports.

PubMed

Qiu, John X; Yoon, Hong-Jun; Fearn, Paul A; Tourassi, Georgia D

2018-01-01

Pathology reports are a primary source of information for cancer registries which process high volumes of free-text reports annually. Information extraction and coding is a manual, labor-intensive process. In this study, we investigated deep learning and a convolutional neural network (CNN), for extracting ICD-O-3 topographic codes from a corpus of breast and lung cancer pathology reports. We performed two experiments, using a CNN and a more conventional term frequency vector approach, to assess the effects of class prevalence and inter-class transfer learning. The experiments were based on a set of 942 pathology reports with human expert annotations as the gold standard. CNN performance was compared against a more conventional term frequency vector space approach. We observed that the deep learning models consistently outperformed the conventional approaches in the class prevalence experiment, resulting in micro- and macro-F score increases of up to 0.132 and 0.226, respectively, when class labels were well populated. Specifically, the best performing CNN achieved a micro-F score of 0.722 over 12 ICD-O-3 topography codes. Transfer learning provided a consistent but modest performance boost for the deep learning methods but trends were contingent on the CNN method and cancer site. These encouraging results demonstrate the potential of deep learning for automated abstraction of pathology reports.

A Deep Learning Approach for Fault Diagnosis of Induction Motors in Manufacturing

NASA Astrophysics Data System (ADS)

Shao, Si-Yu; Sun, Wen-Jun; Yan, Ru-Qiang; Wang, Peng; Gao, Robert X.

2017-11-01

Extracting features from original signals is a key procedure for traditional fault diagnosis of induction motors, as it directly influences the performance of fault recognition. However, high quality features need expert knowledge and human intervention. In this paper, a deep learning approach based on deep belief networks (DBN) is developed to learn features from frequency distribution of vibration signals with the purpose of characterizing working status of induction motors. It combines feature extraction procedure with classification task together to achieve automated and intelligent fault diagnosis. The DBN model is built by stacking multiple-units of restricted Boltzmann machine (RBM), and is trained using layer-by-layer pre-training algorithm. Compared with traditional diagnostic approaches where feature extraction is needed, the presented approach has the ability of learning hierarchical representations, which are suitable for fault classification, directly from frequency distribution of the measurement data. The structure of the DBN model is investigated as the scale and depth of the DBN architecture directly affect its classification performance. Experimental study conducted on a machine fault simulator verifies the effectiveness of the deep learning approach for fault diagnosis of induction motors. This research proposes an intelligent diagnosis method for induction motor which utilizes deep learning model to automatically learn features from sensor data and realize working status recognition.

The cerebellum: a neuronal learning machine?

NASA Technical Reports Server (NTRS)

Raymond, J. L.; Lisberger, S. G.; Mauk, M. D.

1996-01-01

Comparison of two seemingly quite different behaviors yields a surprisingly consistent picture of the role of the cerebellum in motor learning. Behavioral and physiological data about classical conditioning of the eyelid response and motor learning in the vestibulo-ocular reflex suggests that (i) plasticity is distributed between the cerebellar cortex and the deep cerebellar nuclei; (ii) the cerebellar cortex plays a special role in learning the timing of movement; and (iii) the cerebellar cortex guides learning in the deep nuclei, which may allow learning to be transferred from the cortex to the deep nuclei. Because many of the similarities in the data from the two systems typify general features of cerebellar organization, the cerebellar mechanisms of learning in these two systems may represent principles that apply to many motor systems.

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.

Sentiment analysis: a comparison of deep learning neural network algorithm with SVM and naϊve Bayes for Indonesian text

NASA Astrophysics Data System (ADS)

Calvin Frans Mariel, Wahyu; Mariyah, Siti; Pramana, Setia

2018-03-01

Deep learning is a new era of machine learning techniques that essentially imitate the structure and function of the human brain. It is a development of deeper Artificial Neural Network (ANN) that uses more than one hidden layer. Deep Learning Neural Network has a great ability on recognizing patterns from various data types such as picture, audio, text, and many more. In this paper, the authors tries to measure that algorithm’s ability by applying it into the text classification. The classification task herein is done by considering the content of sentiment in a text which is also called as sentiment analysis. By using several combinations of text preprocessing and feature extraction techniques, we aim to compare the precise modelling results of Deep Learning Neural Network with the other two commonly used algorithms, the Naϊve Bayes and Support Vector Machine (SVM). This algorithm comparison uses Indonesian text data with balanced and unbalanced sentiment composition. Based on the experimental simulation, Deep Learning Neural Network clearly outperforms the Naϊve Bayes and SVM and offers a better F-1 Score while for the best feature extraction technique which improves that modelling result is Bigram.

De novo peptide sequencing by deep learning

PubMed Central

Tran, Ngoc Hieu; Zhang, Xianglilan; Xin, Lei; Shan, Baozhen; Li, Ming

2017-01-01

De novo peptide sequencing from tandem MS data is the key technology in proteomics for the characterization of proteins, especially for new sequences, such as mAbs. In this study, we propose a deep neural network model, DeepNovo, for de novo peptide sequencing. DeepNovo architecture combines recent advances in convolutional neural networks and recurrent neural networks to learn features of tandem mass spectra, fragment ions, and sequence patterns of peptides. The networks are further integrated with local dynamic programming to solve the complex optimization task of de novo sequencing. We evaluated the method on a wide variety of species and found that DeepNovo considerably outperformed state of the art methods, achieving 7.7–22.9% higher accuracy at the amino acid level and 38.1–64.0% higher accuracy at the peptide level. We further used DeepNovo to automatically reconstruct the complete sequences of antibody light and heavy chains of mouse, achieving 97.5–100% coverage and 97.2–99.5% accuracy, without assisting databases. Moreover, DeepNovo is retrainable to adapt to any sources of data and provides a complete end-to-end training and prediction solution to the de novo sequencing problem. Not only does our study extend the deep learning revolution to a new field, but it also shows an innovative approach in solving optimization problems by using deep learning and dynamic programming. PMID:28720701

Four Major South Korea's Rivers Using Deep Learning Models.

PubMed

Lee, Sangmok; Lee, Donghyun

2018-06-24

Harmful algal blooms are an annual phenomenon that cause environmental damage, economic losses, and disease outbreaks. A fundamental solution to this problem is still lacking, thus, the best option for counteracting the effects of algal blooms is to improve advance warnings (predictions). However, existing physical prediction models have difficulties setting a clear coefficient indicating the relationship between each factor when predicting algal blooms, and many variable data sources are required for the analysis. These limitations are accompanied by high time and economic costs. Meanwhile, artificial intelligence and deep learning methods have become increasingly common in scientific research; attempts to apply the long short-term memory (LSTM) model to environmental research problems are increasing because the LSTM model exhibits good performance for time-series data prediction. However, few studies have applied deep learning models or LSTM to algal bloom prediction, especially in South Korea, where algal blooms occur annually. Therefore, we employed the LSTM model for algal bloom prediction in four major rivers of South Korea. We conducted short-term (one week) predictions by employing regression analysis and deep learning techniques on a newly constructed water quality and quantity dataset drawn from 16 dammed pools on the rivers. Three deep learning models (multilayer perceptron, MLP; recurrent neural network, RNN; and long short-term memory, LSTM) were used to predict chlorophyll-a, a recognized proxy for algal activity. The results were compared to those from OLS (ordinary least square) regression analysis and actual data based on the root mean square error (RSME). The LSTM model showed the highest prediction rate for harmful algal blooms and all deep learning models out-performed the OLS regression analysis. Our results reveal the potential for predicting algal blooms using LSTM and deep learning.

[Deep learning and neuronal networks in ophthalmology : Applications in the field of optical coherence tomography].

PubMed

Treder, M; Eter, N

2018-04-19

Deep learning is increasingly becoming the focus of various imaging methods in medicine. Due to the large number of different imaging modalities, ophthalmology is particularly suitable for this field of application. This article gives a general overview on the topic of deep learning and its current applications in the field of optical coherence tomography. For the benefit of the reader it focuses on the clinical rather than the technical aspects.

Saliency U-Net: A regional saliency map-driven hybrid deep learning network for anomaly segmentation

NASA Astrophysics Data System (ADS)

Karargyros, Alex; Syeda-Mahmood, Tanveer

2018-02-01

Deep learning networks are gaining popularity in many medical image analysis tasks due to their generalized ability to automatically extract relevant features from raw images. However, this can make the learning problem unnecessarily harder requiring network architectures of high complexity. In case of anomaly detection, in particular, there is often sufficient regional difference between the anomaly and the surrounding parenchyma that could be easily highlighted through bottom-up saliency operators. In this paper we propose a new hybrid deep learning network using a combination of raw image and such regional maps to more accurately learn the anomalies using simpler network architectures. Specifically, we modify a deep learning network called U-Net using both the raw and pre-segmented images as input to produce joint encoding (contraction) and expansion paths (decoding) in the U-Net. We present results of successfully delineating subdural and epidural hematomas in brain CT imaging and liver hemangioma in abdominal CT images using such network.

SchNet - A deep learning architecture for molecules and materials

NASA Astrophysics Data System (ADS)

Schütt, K. T.; Sauceda, H. E.; Kindermans, P.-J.; Tkatchenko, A.; Müller, K.-R.

2018-06-01

Deep learning has led to a paradigm shift in artificial intelligence, including web, text, and image search, speech recognition, as well as bioinformatics, with growing impact in chemical physics. Machine learning, in general, and deep learning, in particular, are ideally suitable for representing quantum-mechanical interactions, enabling us to model nonlinear potential-energy surfaces or enhancing the exploration of chemical compound space. Here we present the deep learning architecture SchNet that is specifically designed to model atomistic systems by making use of continuous-filter convolutional layers. We demonstrate the capabilities of SchNet by accurately predicting a range of properties across chemical space for molecules and materials, where our model learns chemically plausible embeddings of atom types across the periodic table. Finally, we employ SchNet to predict potential-energy surfaces and energy-conserving force fields for molecular dynamics simulations of small molecules and perform an exemplary study on the quantum-mechanical properties of C20-fullerene that would have been infeasible with regular ab initio molecular dynamics.

Towards Scalable Deep Learning via I/O Analysis and Optimization

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

Pumma, Sarunya; Si, Min; Feng, Wu-Chun

Deep learning systems have been growing in prominence as a way to automatically characterize objects, trends, and anomalies. Given the importance of deep learning systems, researchers have been investigating techniques to optimize such systems. An area of particular interest has been using large supercomputing systems to quickly generate effective deep learning networks: a phase often referred to as “training” of the deep learning neural network. As we scale existing deep learning frameworks—such as Caffe—on these large supercomputing systems, we notice that the parallelism can help improve the computation tremendously, leaving data I/O as the major bottleneck limiting the overall systemmore » scalability. In this paper, we first present a detailed analysis of the performance bottlenecks of Caffe on large supercomputing systems. Our analysis shows that the I/O subsystem of Caffe—LMDB—relies on memory-mapped I/O to access its database, which can be highly inefficient on large-scale systems because of its interaction with the process scheduling system and the network-based parallel filesystem. Based on this analysis, we then present LMDBIO, our optimized I/O plugin for Caffe that takes into account the data access pattern of Caffe in order to vastly improve I/O performance. Our experimental results show that LMDBIO can improve the overall execution time of Caffe by nearly 20-fold in some cases.« less

A theory of local learning, the learning channel, and the optimality of backpropagation.

PubMed

Baldi, Pierre; Sadowski, Peter

2016-11-01

In a physical neural system, where storage and processing are intimately intertwined, the rules for adjusting the synaptic weights can only depend on variables that are available locally, such as the activity of the pre- and post-synaptic neurons, resulting in local learning rules. A systematic framework for studying the space of local learning rules is obtained by first specifying the nature of the local variables, and then the functional form that ties them together into each learning rule. Such a framework enables also the systematic discovery of new learning rules and exploration of relationships between learning rules and group symmetries. We study polynomial local learning rules stratified by their degree and analyze their behavior and capabilities in both linear and non-linear units and networks. Stacking local learning rules in deep feedforward networks leads to deep local learning. While deep local learning can learn interesting representations, it cannot learn complex input-output functions, even when targets are available for the top layer. Learning complex input-output functions requires local deep learning where target information is communicated to the deep layers through a backward learning channel. The nature of the communicated information about the targets and the structure of the learning channel partition the space of learning algorithms. For any learning algorithm, the capacity of the learning channel can be defined as the number of bits provided about the error gradient per weight, divided by the number of required operations per weight. We estimate the capacity associated with several learning algorithms and show that backpropagation outperforms them by simultaneously maximizing the information rate and minimizing the computational cost. This result is also shown to be true for recurrent networks, by unfolding them in time. The theory clarifies the concept of Hebbian learning, establishes the power and limitations of local learning rules, introduces the learning channel which enables a formal analysis of the optimality of backpropagation, and explains the sparsity of the space of learning rules discovered so far. Copyright © 2016 Elsevier Ltd. All rights reserved.

Clay mineralogy, strontium and neodymium isotope ratios in the sediments of two High Arctic catchments (Svalbard)

NASA Astrophysics Data System (ADS)

Hindshaw, Ruth S.; Tosca, Nicholas J.; Piotrowski, Alexander M.; Tipper, Edward T.

2018-03-01

The identification of sediment sources to the ocean is a prerequisite to using marine sediment cores to extract information on past climate and ocean circulation. Sr and Nd isotopes are classical tools with which to trace source provenance. Despite considerable interest in the Arctic Ocean, the circum-Arctic source regions are poorly characterised in terms of their Sr and Nd isotopic compositions. In this study we present Sr and Nd isotope data from the Paleogene Central Basin sediments of Svalbard, including the first published data of stream suspended sediments from Svalbard. The stream suspended sediments exhibit considerable isotopic variation (ɛNd = -20.6 to -13.4; 87Sr / 86Sr = 0.73421 to 0.74704) which can be related to the depositional history of the sedimentary formations from which they are derived. In combination with analysis of the clay mineralogy of catchment rocks and sediments, we suggest that the Central Basin sedimentary rocks were derived from two sources. One source is Proterozoic sediments derived from Greenlandic basement rocks which are rich in illite and have high 87Sr / 86Sr and low ɛNd values. The second source is Carboniferous to Jurassic sediments derived from Siberian basalts which are rich in smectite and have low 87Sr / 86Sr and high ɛNd values. Due to a change in depositional conditions throughout the Paleogene (from deep sea to continental) the relative proportions of these two sources vary in the Central Basin formations. The modern stream suspended sediment isotopic composition is then controlled by modern processes, in particular glaciation, which determines the present-day exposure of the formations and therefore the relative contribution of each formation to the stream suspended sediment load. This study demonstrates that the Nd isotopic composition of stream suspended sediments exhibits seasonal variation, which likely mirrors longer-term hydrological changes, with implications for source provenance studies based on fixed end-members through time.

The Effects of Inquiry-Based Computer Simulation with Cooperative Learning on Scientific Thinking and Conceptual Understanding of Gas Laws

ERIC Educational Resources Information Center

Abdullah, Sopiah; Shariff, Adilah

2008-01-01

The purpose of the study was to investigate the effects of inquiry-based computer simulation with heterogeneous-ability cooperative learning (HACL) and inquiry-based computer simulation with friendship cooperative learning (FCL) on (a) scientific reasoning (SR) and (b) conceptual understanding (CU) among Form Four students in Malaysian Smart…

Deep learning for classification of islanding and grid disturbance based on multi-resolution singular spectrum entropy

NASA Astrophysics Data System (ADS)

Li, Tie; He, Xiaoyang; Tang, Junci; Zeng, Hui; Zhou, Chunying; Zhang, Nan; Liu, Hui; Lu, Zhuoxin; Kong, Xiangrui; Yan, Zheng

2018-02-01

Forasmuch as the distinguishment of islanding is easy to be interfered by grid disturbance, island detection device may make misjudgment thus causing the consequence of photovoltaic out of service. The detection device must provide with the ability to differ islanding from grid disturbance. In this paper, the concept of deep learning is introduced into classification of islanding and grid disturbance for the first time. A novel deep learning framework is proposed to detect and classify islanding or grid disturbance. The framework is a hybrid of wavelet transformation, multi-resolution singular spectrum entropy, and deep learning architecture. As a signal processing method after wavelet transformation, multi-resolution singular spectrum entropy combines multi-resolution analysis and spectrum analysis with entropy as output, from which we can extract the intrinsic different features between islanding and grid disturbance. With the features extracted, deep learning is utilized to classify islanding and grid disturbance. Simulation results indicate that the method can achieve its goal while being highly accurate, so the photovoltaic system mistakenly withdrawing from power grids can be avoided.

Discrimination of Breast Cancer with Microcalcifications on Mammography by Deep Learning.

PubMed

Wang, Jinhua; Yang, Xi; Cai, Hongmin; Tan, Wanchang; Jin, Cangzheng; Li, Li

2016-06-07

Microcalcification is an effective indicator of early breast cancer. To improve the diagnostic accuracy of microcalcifications, this study evaluates the performance of deep learning-based models on large datasets for its discrimination. A semi-automated segmentation method was used to characterize all microcalcifications. A discrimination classifier model was constructed to assess the accuracies of microcalcifications and breast masses, either in isolation or combination, for classifying breast lesions. Performances were compared to benchmark models. Our deep learning model achieved a discriminative accuracy of 87.3% if microcalcifications were characterized alone, compared to 85.8% with a support vector machine. The accuracies were 61.3% for both methods with masses alone and improved to 89.7% and 85.8% after the combined analysis with microcalcifications. Image segmentation with our deep learning model yielded 15, 26 and 41 features for the three scenarios, respectively. Overall, deep learning based on large datasets was superior to standard methods for the discrimination of microcalcifications. Accuracy was increased by adopting a combinatorial approach to detect microcalcifications and masses simultaneously. This may have clinical value for early detection and treatment of breast cancer.

Discrimination of Breast Cancer with Microcalcifications on Mammography by Deep Learning

PubMed Central

Wang, Jinhua; Yang, Xi; Cai, Hongmin; Tan, Wanchang; Jin, Cangzheng; Li, Li

2016-01-01

Microcalcification is an effective indicator of early breast cancer. To improve the diagnostic accuracy of microcalcifications, this study evaluates the performance of deep learning-based models on large datasets for its discrimination. A semi-automated segmentation method was used to characterize all microcalcifications. A discrimination classifier model was constructed to assess the accuracies of microcalcifications and breast masses, either in isolation or combination, for classifying breast lesions. Performances were compared to benchmark models. Our deep learning model achieved a discriminative accuracy of 87.3% if microcalcifications were characterized alone, compared to 85.8% with a support vector machine. The accuracies were 61.3% for both methods with masses alone and improved to 89.7% and 85.8% after the combined analysis with microcalcifications. Image segmentation with our deep learning model yielded 15, 26 and 41 features for the three scenarios, respectively. Overall, deep learning based on large datasets was superior to standard methods for the discrimination of microcalcifications. Accuracy was increased by adopting a combinatorial approach to detect microcalcifications and masses simultaneously. This may have clinical value for early detection and treatment of breast cancer. PMID:27273294

Context and Deep Learning Design

ERIC Educational Resources Information Center

Boyle, Tom; Ravenscroft, Andrew

2012-01-01

Conceptual clarification is essential if we are to establish a stable and deep discipline of technology enhanced learning. The technology is alluring; this can distract from deep design in a surface rush to exploit the affordances of the new technology. We need a basis for design, and a conceptual unit of organization, that are applicable across…

Quantification of flexoelectricity in PbTiO3/SrTiO3 superlattice polar vortices using machine learning and phase-field modeling.

PubMed

Li, Q; Nelson, C T; Hsu, S-L; Damodaran, A R; Li, L-L; Yadav, A K; McCarter, M; Martin, L W; Ramesh, R; Kalinin, S V

2017-11-13

Flexoelectricity refers to electric polarization generated by heterogeneous mechanical strains, namely strain gradients, in materials of arbitrary crystal symmetries. Despite more than 50 years of work on this effect, an accurate identification of its coupling strength remains an experimental challenge for most materials, which impedes its wide recognition. Here, we show the presence of flexoelectricity in the recently discovered polar vortices in PbTiO 3 /SrTiO 3 superlattices based on a combination of machine-learning analysis of the atomic-scale electron microscopy imaging data and phenomenological phase-field modeling. By scrutinizing the influence of flexocoupling on the global vortex structure, we match theory and experiment using computer vision methodologies to determine the flexoelectric coefficients for PbTiO 3 and SrTiO 3 . Our findings highlight the inherent, nontrivial role of flexoelectricity in the generation of emergent complex polarization morphologies and demonstrate a viable approach to delineating this effect, conducive to the deeper exploration of both topics.

Hydrous parental magmas of Early to Middle Permian gabbroic intrusions in western Inner Mongolia, North China: New constraints on deep-Earth fluid cycling in the Central Asian Orogenic Belt

NASA Astrophysics Data System (ADS)

Pang, Chong-Jin; Wang, Xuan-Ce; Xu, Bei; Luo, Zhi-Wen; Liu, Yi-Zhi

2017-08-01

The role of fluids in the formation of the Permian-aged Xigedan and Mandula gabbroic intrusions in western Inner Mongolia was significant to the evolution of the Xing'an Mongolia Orogenic Belt (XMOB), and the active northern margin of the North China Craton (NCC). Secondary Ion Mass Spectroscopy (SIMS) U-Pb zircon geochronology establishes that the Xigedan gabbroic intrusion in the northern NCC was emplaced at 266 Ma, and is therefore slightly younger than the ca 280 Ma Mandula gabbroic intrusion in the XMOB. Along with their felsic counterparts, the mafic igneous intrusions record extensive bimodal magmatism along the northern NCC and in the XMOB during the Early to Middle Permian. The Mandula gabbroic rocks have low initial 87Sr/86Sr ratios (0.7040-0.7043) and positive εNd(t) (+6.2 to +7.3) and εHf(t) values (+13.4 to +14.5), resembling to those of contemporaneous Mandula basalts. These features, together with the presence of amphibole and the enrichment of large ion lithophile elements (LILE, e.g., Rb, Ba, U and Sr) and depletion of Nb-Ta suggest that the parental magmas of the Mandula mafic igneous rocks were derived from a depleted mantle source metasomatized by water-rich fluids. In contrast, the Xigedan gabbroic rocks are characterised by high 87Sr/86Sr ratios (0.7078-0.7080) and zircon δ18O values (5.84-6.61‰), but low εNd(t) (-9.3 to -10.2) and εHf(t) values (-8.76 to -8.54), indicative of a long-term enriched subcontinental lithosphere mantle source that was metasomatized by recycled, high δ18O crustal materials prior to partial melting. The high water contents (4.6-6.9 wt%) and arc-like geochemical signature (enrichment of fluid-mobile elements and depletion of Nb-Ta) of the parental magmas of the Xigedan gabbroic rocks further establish the existence of a mantle hydration event caused by fluid/melts released from hydrated recycled oceanic crust. Incompatible element modelling shows that 5-10% partial melting of an enriched mantle source by adding respectively 0.5% and 2% sediment melts and fluids, could have produced the parental magmas of the Xigedan gabbroic rocks. A range of geological evidence establishes an intracontinental origin for Late Paleozoic mafic igneous rocks along the northern NCC and in the XMOB, rather than a subduction-related setting. We therefore propose a deep-Earth water cycling process to account for mantle hydration and subsequent Late Paleozoic magmatism, supporting a geodynamic link between deep-Earth water cycling, and intracontinental magmatism and lithospheric extension.

Creating the learning situation to promote student deep learning: Data analysis and application case

NASA Astrophysics Data System (ADS)

Guo, Yuanyuan; Wu, Shaoyan

2017-05-01

How to lead students to deeper learning and cultivate engineering innovative talents need to be studied for higher engineering education. In this study, through the survey data analysis and theoretical research, we discuss the correlation of teaching methods, learning motivation, and learning methods. In this research, we find that students have different motivation orientation according to the perception of teaching methods in the process of engineering education, and this affects their choice of learning methods. As a result, creating situations is critical to lead students to deeper learning. Finally, we analyze the process of learning situational creation in the teaching process of «bidding and contract management workshops». In this creation process, teachers use the student-centered teaching to lead students to deeper study. Through the study of influence factors of deep learning process, and building the teaching situation for the purpose of promoting deep learning, this thesis provide a meaningful reference for enhancing students' learning quality, teachers' teaching quality and the quality of innovation talent.

Deep machine learning provides state-of-the-art performance in image-based plant phenotyping.

PubMed

Pound, Michael P; Atkinson, Jonathan A; Townsend, Alexandra J; Wilson, Michael H; Griffiths, Marcus; Jackson, Aaron S; Bulat, Adrian; Tzimiropoulos, Georgios; Wells, Darren M; Murchie, Erik H; Pridmore, Tony P; French, Andrew P

2017-10-01

In plant phenotyping, it has become important to be able to measure many features on large image sets in order to aid genetic discovery. The size of the datasets, now often captured robotically, often precludes manual inspection, hence the motivation for finding a fully automated approach. Deep learning is an emerging field that promises unparalleled results on many data analysis problems. Building on artificial neural networks, deep approaches have many more hidden layers in the network, and hence have greater discriminative and predictive power. We demonstrate the use of such approaches as part of a plant phenotyping pipeline. We show the success offered by such techniques when applied to the challenging problem of image-based plant phenotyping and demonstrate state-of-the-art results (>97% accuracy) for root and shoot feature identification and localization. We use fully automated trait identification using deep learning to identify quantitative trait loci in root architecture datasets. The majority (12 out of 14) of manually identified quantitative trait loci were also discovered using our automated approach based on deep learning detection to locate plant features. We have shown deep learning-based phenotyping to have very good detection and localization accuracy in validation and testing image sets. We have shown that such features can be used to derive meaningful biological traits, which in turn can be used in quantitative trait loci discovery pipelines. This process can be completely automated. We predict a paradigm shift in image-based phenotyping bought about by such deep learning approaches, given sufficient training sets. © The Authors 2017. Published by Oxford University Press.

A Case Study on Sepsis Using PubMed and Deep Learning for Ontology Learning.

PubMed

Arguello Casteleiro, Mercedes; Maseda Fernandez, Diego; Demetriou, George; Read, Warren; Fernandez Prieto, Maria Jesus; Des Diz, Julio; Nenadic, Goran; Keane, John; Stevens, Robert

2017-01-01

We investigate the application of distributional semantics models for facilitating unsupervised extraction of biomedical terms from unannotated corpora. Term extraction is used as the first step of an ontology learning process that aims to (semi-)automatic annotation of biomedical concepts and relations from more than 300K PubMed titles and abstracts. We experimented with both traditional distributional semantics methods such as Latent Semantic Analysis (LSA) and Latent Dirichlet Allocation (LDA) as well as the neural language models CBOW and Skip-gram from Deep Learning. The evaluation conducted concentrates on sepsis, a major life-threatening condition, and shows that Deep Learning models outperform LSA and LDA with much higher precision.

Exploring the Function Space of Deep-Learning Machines

NASA Astrophysics Data System (ADS)

Li, Bo; Saad, David

2018-06-01

The function space of deep-learning machines is investigated by studying growth in the entropy of functions of a given error with respect to a reference function, realized by a deep-learning machine. Using physics-inspired methods we study both sparsely and densely connected architectures to discover a layerwise convergence of candidate functions, marked by a corresponding reduction in entropy when approaching the reference function, gain insight into the importance of having a large number of layers, and observe phase transitions as the error increases.

The Effects of Deep Approaches to Learning on Students' Need for Cognition over Four Years of College

ERIC Educational Resources Information Center

Wang, Jui-Sheng

2013-01-01

This study examines the effect of deep approaches to learning on development of the inclination to inquire and lifelong learning over four years, as an essential graduated outcome that helps students face the challenges of a complex and rapidly changing world. Despite the importance of the inclination to inquire and lifelong learning, some…

Accommodating Learning Styles in Prison Writing Classes.

ERIC Educational Resources Information Center

Glasgow, Jacqueline N.

1994-01-01

Describes a developmental writing course taught in prison. Describes how the teaching styles took into account the presumed learning preferences of the African Americans in this group of inmates and resulted in a boost of both their self-confidence and the level of their writing. (SR)

Exploring How Conversations Meet Teacher Learning Needs

ERIC Educational Resources Information Center

Rowland, Amber Heiserman

2012-01-01

This study identified the content of educator conversations and determined how social interactions contributed to participant learning. Data sources included videos from face-to-face conversational sessions and individual, video stimulated-recall (SR) interviews conducted virtually. Participants included fifth and sixth-grade teachers from five…

Plant Species Identification by Bi-channel Deep Convolutional Networks

NASA Astrophysics Data System (ADS)

He, Guiqing; Xia, Zhaoqiang; Zhang, Qiqi; Zhang, Haixi; Fan, Jianping

2018-04-01

Plant species identification achieves much attention recently as it has potential application in the environmental protection and human life. Although deep learning techniques can be directly applied for plant species identification, it still needs to be designed for this specific task to obtain the state-of-art performance. In this paper, a bi-channel deep learning framework is developed for identifying plant species. In the framework, two different sub-networks are fine-tuned over their pretrained models respectively. And then a stacking layer is used to fuse the output of two different sub-networks. We construct a plant dataset of Orchidaceae family for algorithm evaluation. Our experimental results have demonstrated that our bi-channel deep network can achieve very competitive performance on accuracy rates compared to the existing deep learning algorithm.

NiftyNet: a deep-learning platform for medical imaging.

PubMed

Gibson, Eli; Li, Wenqi; Sudre, Carole; Fidon, Lucas; Shakir, Dzhoshkun I; Wang, Guotai; Eaton-Rosen, Zach; Gray, Robert; Doel, Tom; Hu, Yipeng; Whyntie, Tom; Nachev, Parashkev; Modat, Marc; Barratt, Dean C; Ourselin, Sébastien; Cardoso, M Jorge; Vercauteren, Tom

2018-05-01

Medical image analysis and computer-assisted intervention problems are increasingly being addressed with deep-learning-based solutions. Established deep-learning platforms are flexible but do not provide specific functionality for medical image analysis and adapting them for this domain of application requires substantial implementation effort. Consequently, there has been substantial duplication of effort and incompatible infrastructure developed across many research groups. This work presents the open-source NiftyNet platform for deep learning in medical imaging. The ambition of NiftyNet is to accelerate and simplify the development of these solutions, and to provide a common mechanism for disseminating research outputs for the community to use, adapt and build upon. The NiftyNet infrastructure provides a modular deep-learning pipeline for a range of medical imaging applications including segmentation, regression, image generation and representation learning applications. Components of the NiftyNet pipeline including data loading, data augmentation, network architectures, loss functions and evaluation metrics are tailored to, and take advantage of, the idiosyncracies of medical image analysis and computer-assisted intervention. NiftyNet is built on the TensorFlow framework and supports features such as TensorBoard visualization of 2D and 3D images and computational graphs by default. We present three illustrative medical image analysis applications built using NiftyNet infrastructure: (1) segmentation of multiple abdominal organs from computed tomography; (2) image regression to predict computed tomography attenuation maps from brain magnetic resonance images; and (3) generation of simulated ultrasound images for specified anatomical poses. The NiftyNet infrastructure enables researchers to rapidly develop and distribute deep learning solutions for segmentation, regression, image generation and representation learning applications, or extend the platform to new applications. Copyright © 2018 The Authors. Published by Elsevier B.V. All rights reserved.

Goal Orientation, Deep Learning, and Sustainable Feedback in Higher Business Education

ERIC Educational Resources Information Center

Geitz, Gerry; Brinke, Desirée Joosten-ten; Kirschner, Paul A.

2015-01-01

Relations between and changeability of goal orientation and learning behavior have been studied in several domains and contexts. To alter the adopted goal orientation into a mastery orientation and increase a concomitant deep learning in international business students, a sustainable feedback intervention study was carried out. Sustainable…

Theoretical Explanation for Success of Deep-Level-Learning Study Tours

ERIC Educational Resources Information Center

Bergsteiner, Harald; Avery, Gayle C.

2008-01-01

Study tours can help internationalize curricula and prepare students for global workplaces. We examine benefits of tours providing deep-level learning experiences rather than industrial tourism using five main theoretical frameworks to highlight the diverse learning benefits associated with intensive study tours in particular. Relevant theoretical…

Defects, stoichiometry, and electronic transport in SrTiO{sub 3-δ} epilayers: A high pressure oxygen sputter deposition study

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

Ambwani, P.; Xu, P.; Jeong, J. S.

SrTiO{sub 3} is not only of enduring interest due to its unique dielectric, structural, and lattice dynamical properties, but is also the archetypal perovskite oxide semiconductor and a foundational material in oxide heterostructures and electronics. This has naturally focused attention on growth, stoichiometry, and defects in SrTiO{sub 3}, one exciting recent development being such precisely stoichiometric defect-managed thin films that electron mobilities have finally exceeded bulk crystals. This has been achieved only by molecular beam epitaxy, however (and to a somewhat lesser extent pulsed laser deposition (PLD)), and numerous open questions remain. Here, we present a study of the stoichiometry,more » defects, and structure in SrTiO{sub 3} synthesized by a different method, high pressure oxygen sputtering, relating the results to electronic transport. We find that this form of sputter deposition is also capable of homoepitaxy of precisely stoichiometric SrTiO{sub 3}, but only provided that substrate and target preparation, temperature, pressure, and deposition rate are carefully controlled. Even under these conditions, oxygen-vacancy-doped heteroepitaxial SrTiO{sub 3} films are found to have carrier density, mobility, and conductivity significantly lower than bulk. While surface depletion plays a role, it is argued from particle-induced X-ray emission (PIXE) measurements of trace impurities in commercial sputtering targets that this is also due to deep acceptors such as Fe at 100's of parts-per-million levels. Comparisons of PIXE from SrTiO{sub 3} crystals and polycrystalline targets are shown to be of general interest, with clear implications for sputter and PLD deposition of this important material.« less

ACTIVIS: Visual Exploration of Industry-Scale Deep Neural Network Models.

PubMed

Kahng, Minsuk; Andrews, Pierre Y; Kalro, Aditya; Polo Chau, Duen Horng

2017-08-30

While deep learning models have achieved state-of-the-art accuracies for many prediction tasks, understanding these models remains a challenge. Despite the recent interest in developing visual tools to help users interpret deep learning models, the complexity and wide variety of models deployed in industry, and the large-scale datasets that they used, pose unique design challenges that are inadequately addressed by existing work. Through participatory design sessions with over 15 researchers and engineers at Facebook, we have developed, deployed, and iteratively improved ACTIVIS, an interactive visualization system for interpreting large-scale deep learning models and results. By tightly integrating multiple coordinated views, such as a computation graph overview of the model architecture, and a neuron activation view for pattern discovery and comparison, users can explore complex deep neural network models at both the instance- and subset-level. ACTIVIS has been deployed on Facebook's machine learning platform. We present case studies with Facebook researchers and engineers, and usage scenarios of how ACTIVIS may work with different models.

Jet-images — deep learning edition

DOE PAGES

de Oliveira, Luke; Kagan, Michael; Mackey, Lester; ...

2016-07-13

Building on the notion of a particle physics detector as a camera and the collimated streams of high energy particles, or jets, it measures as an image, we investigate the potential of machine learning techniques based on deep learning architectures to identify highly boosted W bosons. Modern deep learning algorithms trained on jet images can out-perform standard physically-motivated feature driven approaches to jet tagging. We develop techniques for visualizing how these features are learned by the network and what additional information is used to improve performance. Finally, this interplay between physically-motivated feature driven tools and supervised learning algorithms is generalmore » and can be used to significantly increase the sensitivity to discover new particles and new forces, and gain a deeper understanding of the physics within jets.« less

Jet-images — deep learning edition

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

de Oliveira, Luke; Kagan, Michael; Mackey, Lester

Building on the notion of a particle physics detector as a camera and the collimated streams of high energy particles, or jets, it measures as an image, we investigate the potential of machine learning techniques based on deep learning architectures to identify highly boosted W bosons. Modern deep learning algorithms trained on jet images can out-perform standard physically-motivated feature driven approaches to jet tagging. We develop techniques for visualizing how these features are learned by the network and what additional information is used to improve performance. Finally, this interplay between physically-motivated feature driven tools and supervised learning algorithms is generalmore » and can be used to significantly increase the sensitivity to discover new particles and new forces, and gain a deeper understanding of the physics within jets.« less

Deep Processing Strategies and Critical Thinking: Developmental Trajectories Using Latent Growth Analyses

ERIC Educational Resources Information Center

Phan, Huy P.

2011-01-01

The author explored the developmental courses of deep learning approach and critical thinking over a 2-year period. Latent growth curve modeling (LGM) procedures were used to test and trace the trajectories of both theoretical frameworks over time. Participants were 264 (119 women, 145 men) university undergraduates. The Deep Learning subscale of…

Dahl (S × R) Rat Congenic Strain Analysis Confirms and Defines a Chromosome 17 Spatial Navigation Quantitative Trait Locus to <10 Mbp

PubMed Central

Herrera, Victoria L.; Pasion, Khristine A.; Tan, Glaiza A.; Ruiz-Opazo, Nelson

2013-01-01

A quantitative trait locus (QTL) linked with ability to find a platform in the Morris Water Maze (MWM) was located on chromosome 17 (Nav-5 QTL) using intercross between Dahl S and Dahl R rats. We developed two congenic strains, S.R17A and S.R17B introgressing Dahl R-chromosome 17 segments into Dahl S chromosome 17 region spanning putative Nav-5 QTL. Performance analysis of S.R17A, S.R17B and Dahl S rats in the Morris water maze (MWM) task showed a significantly decreased spatial navigation performance in S.R17B congenic rats when compared with Dahl S controls (P = 0.02). The S.R17A congenic segment did not affect MWM performance delimiting Nav-5 to the chromosome 17 65.02–74.66 Mbp region. Additional fine mapping is necessary to identify the specific gene variant accounting for Nav-5 effect on spatial learning and memory in Dahl rats. PMID:23469157

Enhancing Self-Practice/Self-Reflection (SP/SR) approach to cognitive behaviour training through the use of reflective blogs.

PubMed

Farrand, Paul; Perry, Jon; Linsley, Sue

2010-07-01

Self-Practice/Self-Reflection (SP/SR) is increasingly beginning to feature as a central component of CBT training programmes (Bennett-Levy et al., 2001). Programmes including a reflective element, however, are not unproblematic and it has been documented that simply setting time aside for reflection does not necessarily result in trainees using such time to reflect. Such limitations may be overcome by including a requirement to post reflections on reflective blogs. To examine the effect that a requirement to contribute to a reflective blog had upon a SP/SR approach to CBT training. A focus group methodology was adopted with data analyzed using a general inductive qualitative approach. The requirement to use blogs to reflect upon the self-practice of CBT techniques enhanced SP/SR, established a learning community, and improved course supervision, although some technical difficulties arose. Consideration should be given towards using reflective blogs to support a SP/SR approach to CBT training. Benefits afforded by the use of reflective blogs further establish SP/SR as a valid and flexible training approach.

DeepPicker: A deep learning approach for fully automated particle picking in cryo-EM.

PubMed

Wang, Feng; Gong, Huichao; Liu, Gaochao; Li, Meijing; Yan, Chuangye; Xia, Tian; Li, Xueming; Zeng, Jianyang

2016-09-01

Particle picking is a time-consuming step in single-particle analysis and often requires significant interventions from users, which has become a bottleneck for future automated electron cryo-microscopy (cryo-EM). Here we report a deep learning framework, called DeepPicker, to address this problem and fill the current gaps toward a fully automated cryo-EM pipeline. DeepPicker employs a novel cross-molecule training strategy to capture common features of particles from previously-analyzed micrographs, and thus does not require any human intervention during particle picking. Tests on the recently-published cryo-EM data of three complexes have demonstrated that our deep learning based scheme can successfully accomplish the human-level particle picking process and identify a sufficient number of particles that are comparable to those picked manually by human experts. These results indicate that DeepPicker can provide a practically useful tool to significantly reduce the time and manual effort spent in single-particle analysis and thus greatly facilitate high-resolution cryo-EM structure determination. DeepPicker is released as an open-source program, which can be downloaded from https://github.com/nejyeah/DeepPicker-python. Copyright © 2016 Elsevier Inc. All rights reserved.

Strontium isotopic signatures of the streams and lakes of Taylor Valley, Southern Victoria Land, Antarctica: Chemical weathering in a polar climate

USGS Publications Warehouse

Lyons, W.B.; Nezat, C.A.; Benson, L.V.; Bullen, T.D.; Graham, E.Y.; Kidd, J.; Welch, K.A.

2002-01-01

We have collected and analyzed a series of water samples from three closed-basin lakes (Lakes Bonney, Fryxell, and Hoare) in Taylor Valley, Antarctica, and the streams that flow into them. In all three lakes, the hypolimnetic waters have different 87Sr/86Sr ratios than the surface waters, with the deep water of Lakes Fryxell and Hoare being less radiogenic than the surface waters. The opposite occurs in Lake Bonney. The Lake Fryxell isotopic ratios are lower than modern-day ocean water and most of the whole-rock ratios of the surrounding geologic materials. A conceivable source of Sr to the system could be either the Cenozoic volcanic rocks that make up a small portion of the till deposited in the valley during the Last Glacial Maximum or from marble derived from the local basement rocks. The more radiogenic ratios from Lake Bonney originate from ancient salt deposits that flow into the lake from Taylor Glacier and the weathering of minerals with more radiogenic Sr isotopic ratios within the tills. The Sr isotopic data from the streams and lakes of Taylor Valley strongly support the notion documented by previous investigators that chemical weathering has been, and is currently, a major process in determining the overall aquatic chemistry of these lakes in this polar desert environment.

Zircon (U-Th)/He Thermochronometric Constraints on Himalayan Thrust Belt Exhumation, Bedrock Weathering, and Cenozoic Seawater Chemistry

NASA Astrophysics Data System (ADS)

Colleps, Cody L.; McKenzie, N. Ryan; Stockli, Daniel F.; Hughes, Nigel C.; Singh, Birendra P.; Webb, A. Alexander G.; Myrow, Paul M.; Planavsky, Noah J.; Horton, Brian K.

2018-01-01

Shifts in global seawater 187Os/188Os and 87Sr/86Sr are often utilized as proxies to track global weathering processes responsible for CO2 fluctuations in Earth history, particularly climatic cooling during the Cenozoic. It has been proposed, however, that these isotopic records instead reflect the weathering of chemically distinctive Himalayan lithologies exposed at the surface. We present new zircon (U-Th)/He thermochronometric and detrital zircon U-Pb geochronologic evidence from the Himalaya of northwest India to explore these contrasting interpretations concerning the driving mechanisms responsible for these seawater records. Our data demonstrate in-sequence southward thrust propagation with rapid exhumation of Lesser Himalayan strata enriched in labile 187Os and relatively less in radiogenic 87Sr at ˜16 Ma, which directly corresponds with coeval shifts in seawater 187Os/188Os and 87Sr/86Sr. Results presented here provide substantial evidence that the onset of exhumation of 187Os-enriched Lesser Himalayan strata could have significantly impacted the marine 187Os/188Os record at 16 Ma. These results support the hypothesis that regional weathering of isotopically unique source rocks can drive seawater records independently from shifts in global-scale weathering rates, hindering the utility of these records as reliable proxies to track global weathering processes and climate in deep geologic time.

Structural brain changes versus self-report: machine-learning classification of chronic fatigue syndrome patients.

PubMed

Sevel, Landrew S; Boissoneault, Jeff; Letzen, Janelle E; Robinson, Michael E; Staud, Roland

2018-05-30

Chronic fatigue syndrome (CFS) is a disorder associated with fatigue, pain, and structural/functional abnormalities seen during magnetic resonance brain imaging (MRI). Therefore, we evaluated the performance of structural MRI (sMRI) abnormalities in the classification of CFS patients versus healthy controls and compared it to machine learning (ML) classification based upon self-report (SR). Participants included 18 CFS patients and 15 healthy controls (HC). All subjects underwent T1-weighted sMRI and provided visual analogue-scale ratings of fatigue, pain intensity, anxiety, depression, anger, and sleep quality. sMRI data were segmented using FreeSurfer and 61 regions based on functional and structural abnormalities previously reported in patients with CFS. Classification was performed in RapidMiner using a linear support vector machine and bootstrap optimism correction. We compared ML classifiers based on (1) 61 a priori sMRI regional estimates and (2) SR ratings. The sMRI model achieved 79.58% classification accuracy. The SR (accuracy = 95.95%) outperformed both sMRI models. Estimates from multiple brain areas related to cognition, emotion, and memory contributed strongly to group classification. This is the first ML-based group classification of CFS. Our findings suggest that sMRI abnormalities are useful for discriminating CFS patients from HC, but SR ratings remain most effective in classification tasks.

Deep learning for healthcare: review, opportunities and challenges.

PubMed

Miotto, Riccardo; Wang, Fei; Wang, Shuang; Jiang, Xiaoqian; Dudley, Joel T

2017-05-06

Gaining knowledge and actionable insights from complex, high-dimensional and heterogeneous biomedical data remains a key challenge in transforming health care. Various types of data have been emerging in modern biomedical research, including electronic health records, imaging, -omics, sensor data and text, which are complex, heterogeneous, poorly annotated and generally unstructured. Traditional data mining and statistical learning approaches typically need to first perform feature engineering to obtain effective and more robust features from those data, and then build prediction or clustering models on top of them. There are lots of challenges on both steps in a scenario of complicated data and lacking of sufficient domain knowledge. The latest advances in deep learning technologies provide new effective paradigms to obtain end-to-end learning models from complex data. In this article, we review the recent literature on applying deep learning technologies to advance the health care domain. Based on the analyzed work, we suggest that deep learning approaches could be the vehicle for translating big biomedical data into improved human health. However, we also note limitations and needs for improved methods development and applications, especially in terms of ease-of-understanding for domain experts and citizen scientists. We discuss such challenges and suggest developing holistic and meaningful interpretable architectures to bridge deep learning models and human interpretability. © The Author 2017. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com.

Deep Learning for Automated Extraction of Primary Sites from Cancer Pathology Reports

DOE PAGES

Qiu, John; Yoon, Hong-Jun; Fearn, Paul A.; ...

2017-05-03

Pathology reports are a primary source of information for cancer registries which process high volumes of free-text reports annually. Information extraction and coding is a manual, labor-intensive process. Here in this study we investigated deep learning and a convolutional neural network (CNN), for extracting ICDO- 3 topographic codes from a corpus of breast and lung cancer pathology reports. We performed two experiments, using a CNN and a more conventional term frequency vector approach, to assess the effects of class prevalence and inter-class transfer learning. The experiments were based on a set of 942 pathology reports with human expert annotations asmore » the gold standard. CNN performance was compared against a more conventional term frequency vector space approach. We observed that the deep learning models consistently outperformed the conventional approaches in the class prevalence experiment, resulting in micro and macro-F score increases of up to 0.132 and 0.226 respectively when class labels were well populated. Specifically, the best performing CNN achieved a micro-F score of 0.722 over 12 ICD-O-3 topography codes. Transfer learning provided a consistent but modest performance boost for the deep learning methods but trends were contingent on CNN method and cancer site. Finally, these encouraging results demonstrate the potential of deep learning for automated abstraction of pathology reports.« less

Deep Learning for Automated Extraction of Primary Sites from Cancer Pathology Reports

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

Qiu, John; Yoon, Hong-Jun; Fearn, Paul A.

Pathology reports are a primary source of information for cancer registries which process high volumes of free-text reports annually. Information extraction and coding is a manual, labor-intensive process. Here in this study we investigated deep learning and a convolutional neural network (CNN), for extracting ICDO- 3 topographic codes from a corpus of breast and lung cancer pathology reports. We performed two experiments, using a CNN and a more conventional term frequency vector approach, to assess the effects of class prevalence and inter-class transfer learning. The experiments were based on a set of 942 pathology reports with human expert annotations asmore » the gold standard. CNN performance was compared against a more conventional term frequency vector space approach. We observed that the deep learning models consistently outperformed the conventional approaches in the class prevalence experiment, resulting in micro and macro-F score increases of up to 0.132 and 0.226 respectively when class labels were well populated. Specifically, the best performing CNN achieved a micro-F score of 0.722 over 12 ICD-O-3 topography codes. Transfer learning provided a consistent but modest performance boost for the deep learning methods but trends were contingent on CNN method and cancer site. Finally, these encouraging results demonstrate the potential of deep learning for automated abstraction of pathology reports.« less

A deep learning-based multi-model ensemble method for cancer prediction.

PubMed

Xiao, Yawen; Wu, Jun; Lin, Zongli; Zhao, Xiaodong

2018-01-01

Cancer is a complex worldwide health problem associated with high mortality. With the rapid development of the high-throughput sequencing technology and the application of various machine learning methods that have emerged in recent years, progress in cancer prediction has been increasingly made based on gene expression, providing insight into effective and accurate treatment decision making. Thus, developing machine learning methods, which can successfully distinguish cancer patients from healthy persons, is of great current interest. However, among the classification methods applied to cancer prediction so far, no one method outperforms all the others. In this paper, we demonstrate a new strategy, which applies deep learning to an ensemble approach that incorporates multiple different machine learning models. We supply informative gene data selected by differential gene expression analysis to five different classification models. Then, a deep learning method is employed to ensemble the outputs of the five classifiers. The proposed deep learning-based multi-model ensemble method was tested on three public RNA-seq data sets of three kinds of cancers, Lung Adenocarcinoma, Stomach Adenocarcinoma and Breast Invasive Carcinoma. The test results indicate that it increases the prediction accuracy of cancer for all the tested RNA-seq data sets as compared to using a single classifier or the majority voting algorithm. By taking full advantage of different classifiers, the proposed deep learning-based multi-model ensemble method is shown to be accurate and effective for cancer prediction. Copyright © 2017 Elsevier B.V. All rights reserved.

Broad Learning System: An Effective and Efficient Incremental Learning System Without the Need for Deep Architecture.

PubMed

Chen, C L Philip; Liu, Zhulin

2018-01-01

Broad Learning System (BLS) that aims to offer an alternative way of learning in deep structure is proposed in this paper. Deep structure and learning suffer from a time-consuming training process because of a large number of connecting parameters in filters and layers. Moreover, it encounters a complete retraining process if the structure is not sufficient to model the system. The BLS is established in the form of a flat network, where the original inputs are transferred and placed as "mapped features" in feature nodes and the structure is expanded in wide sense in the "enhancement nodes." The incremental learning algorithms are developed for fast remodeling in broad expansion without a retraining process if the network deems to be expanded. Two incremental learning algorithms are given for both the increment of the feature nodes (or filters in deep structure) and the increment of the enhancement nodes. The designed model and algorithms are very versatile for selecting a model rapidly. In addition, another incremental learning is developed for a system that has been modeled encounters a new incoming input. Specifically, the system can be remodeled in an incremental way without the entire retraining from the beginning. Satisfactory result for model reduction using singular value decomposition is conducted to simplify the final structure. Compared with existing deep neural networks, experimental results on the Modified National Institute of Standards and Technology database and NYU NORB object recognition dataset benchmark data demonstrate the effectiveness of the proposed BLS.

Deep learning with non-medical training used for chest pathology identification

NASA Astrophysics Data System (ADS)

Bar, Yaniv; Diamant, Idit; Wolf, Lior; Greenspan, Hayit

2015-03-01

In this work, we examine the strength of deep learning approaches for pathology detection in chest radiograph data. Convolutional neural networks (CNN) deep architecture classification approaches have gained popularity due to their ability to learn mid and high level image representations. We explore the ability of a CNN to identify different types of pathologies in chest x-ray images. Moreover, since very large training sets are generally not available in the medical domain, we explore the feasibility of using a deep learning approach based on non-medical learning. We tested our algorithm on a dataset of 93 images. We use a CNN that was trained with ImageNet, a well-known large scale nonmedical image database. The best performance was achieved using a combination of features extracted from the CNN and a set of low-level features. We obtained an area under curve (AUC) of 0.93 for Right Pleural Effusion detection, 0.89 for Enlarged heart detection and 0.79 for classification between healthy and abnormal chest x-ray, where all pathologies are combined into one large class. This is a first-of-its-kind experiment that shows that deep learning with large scale non-medical image databases may be sufficient for general medical image recognition tasks.

An Advanced Deep Learning Approach for Ki-67 Stained Hotspot Detection and Proliferation Rate Scoring for Prognostic Evaluation of Breast Cancer.

PubMed

Saha, Monjoy; Chakraborty, Chandan; Arun, Indu; Ahmed, Rosina; Chatterjee, Sanjoy

2017-06-12

Being a non-histone protein, Ki-67 is one of the essential biomarkers for the immunohistochemical assessment of proliferation rate in breast cancer screening and grading. The Ki-67 signature is always sensitive to radiotherapy and chemotherapy. Due to random morphological, color and intensity variations of cell nuclei (immunopositive and immunonegative), manual/subjective assessment of Ki-67 scoring is error-prone and time-consuming. Hence, several machine learning approaches have been reported; nevertheless, none of them had worked on deep learning based hotspots detection and proliferation scoring. In this article, we suggest an advanced deep learning model for computerized recognition of candidate hotspots and subsequent proliferation rate scoring by quantifying Ki-67 appearance in breast cancer immunohistochemical images. Unlike existing Ki-67 scoring techniques, our methodology uses Gamma mixture model (GMM) with Expectation-Maximization for seed point detection and patch selection and deep learning, comprises with decision layer, for hotspots detection and proliferation scoring. Experimental results provide 93% precision, 0.88% recall and 0.91% F-score value. The model performance has also been compared with the pathologists' manual annotations and recently published articles. In future, the proposed deep learning framework will be highly reliable and beneficial to the junior and senior pathologists for fast and efficient Ki-67 scoring.

Deep learning for tumor classification in imaging mass spectrometry.

PubMed

Behrmann, Jens; Etmann, Christian; Boskamp, Tobias; Casadonte, Rita; Kriegsmann, Jörg; Maaß, Peter

2018-04-01

Tumor classification using imaging mass spectrometry (IMS) data has a high potential for future applications in pathology. Due to the complexity and size of the data, automated feature extraction and classification steps are required to fully process the data. Since mass spectra exhibit certain structural similarities to image data, deep learning may offer a promising strategy for classification of IMS data as it has been successfully applied to image classification. Methodologically, we propose an adapted architecture based on deep convolutional networks to handle the characteristics of mass spectrometry data, as well as a strategy to interpret the learned model in the spectral domain based on a sensitivity analysis. The proposed methods are evaluated on two algorithmically challenging tumor classification tasks and compared to a baseline approach. Competitiveness of the proposed methods is shown on both tasks by studying the performance via cross-validation. Moreover, the learned models are analyzed by the proposed sensitivity analysis revealing biologically plausible effects as well as confounding factors of the considered tasks. Thus, this study may serve as a starting point for further development of deep learning approaches in IMS classification tasks. https://gitlab.informatik.uni-bremen.de/digipath/Deep_Learning_for_Tumor_Classification_in_IMS. jbehrmann@uni-bremen.de or christianetmann@uni-bremen.de. Supplementary data are available at Bioinformatics online.

Deep learning methods to guide CT image reconstruction and reduce metal artifacts

NASA Astrophysics Data System (ADS)

Gjesteby, Lars; Yang, Qingsong; Xi, Yan; Zhou, Ye; Zhang, Junping; Wang, Ge

2017-03-01

The rapidly-rising field of machine learning, including deep learning, has inspired applications across many disciplines. In medical imaging, deep learning has been primarily used for image processing and analysis. In this paper, we integrate a convolutional neural network (CNN) into the computed tomography (CT) image reconstruction process. Our first task is to monitor the quality of CT images during iterative reconstruction and decide when to stop the process according to an intelligent numerical observer instead of using a traditional stopping rule, such as a fixed error threshold or a maximum number of iterations. After training on ground truth images, the CNN was successful in guiding an iterative reconstruction process to yield high-quality images. Our second task is to improve a sinogram to correct for artifacts caused by metal objects. A large number of interpolation and normalization-based schemes were introduced for metal artifact reduction (MAR) over the past four decades. The NMAR algorithm is considered a state-of-the-art method, although residual errors often remain in the reconstructed images, especially in cases of multiple metal objects. Here we merge NMAR with deep learning in the projection domain to achieve additional correction in critical image regions. Our results indicate that deep learning can be a viable tool to address CT reconstruction challenges.

A novel application of deep learning for single-lead ECG classification.

PubMed

Mathews, Sherin M; Kambhamettu, Chandra; Barner, Kenneth E

2018-06-04

Detecting and classifying cardiac arrhythmias is critical to the diagnosis of patients with cardiac abnormalities. In this paper, a novel approach based on deep learning methodology is proposed for the classification of single-lead electrocardiogram (ECG) signals. We demonstrate the application of the Restricted Boltzmann Machine (RBM) and deep belief networks (DBN) for ECG classification following detection of ventricular and supraventricular heartbeats using single-lead ECG. The effectiveness of this proposed algorithm is illustrated using real ECG signals from the widely-used MIT-BIH database. Simulation results demonstrate that with a suitable choice of parameters, RBM and DBN can achieve high average recognition accuracies of ventricular ectopic beats (93.63%) and of supraventricular ectopic beats (95.57%) at a low sampling rate of 114 Hz. Experimental results indicate that classifiers built into this deep learning-based framework achieved state-of-the art performance models at lower sampling rates and simple features when compared to traditional methods. Further, employing features extracted at a sampling rate of 114 Hz when combined with deep learning provided enough discriminatory power for the classification task. This performance is comparable to that of traditional methods and uses a much lower sampling rate and simpler features. Thus, our proposed deep neural network algorithm demonstrates that deep learning-based methods offer accurate ECG classification and could potentially be extended to other physiological signal classifications, such as those in arterial blood pressure (ABP), nerve conduction (EMG), and heart rate variability (HRV) studies. Copyright © 2018. Published by Elsevier Ltd.

Automatic detection of hemorrhagic pericardial effusion on PMCT using deep learning - a feasibility study.

PubMed

Ebert, Lars C; Heimer, Jakob; Schweitzer, Wolf; Sieberth, Till; Leipner, Anja; Thali, Michael; Ampanozi, Garyfalia

2017-12-01

Post mortem computed tomography (PMCT) can be used as a triage tool to better identify cases with a possibly non-natural cause of death, especially when high caseloads make it impossible to perform autopsies on all cases. Substantial data can be generated by modern medical scanners, especially in a forensic setting where the entire body is documented at high resolution. A solution for the resulting issues could be the use of deep learning techniques for automatic analysis of radiological images. In this article, we wanted to test the feasibility of such methods for forensic imaging by hypothesizing that deep learning methods can detect and segment a hemopericardium in PMCT. For deep learning image analysis software, we used the ViDi Suite 2.0. We retrospectively selected 28 cases with, and 24 cases without, hemopericardium. Based on these data, we trained two separate deep learning networks. The first one classified images into hemopericardium/not hemopericardium, and the second one segmented the blood content. We randomly selected 50% of the data for training and 50% for validation. This process was repeated 20 times. The best performing classification network classified all cases of hemopericardium from the validation images correctly with only a few false positives. The best performing segmentation network would tend to underestimate the amount of blood in the pericardium, which is the case for most networks. This is the first study that shows that deep learning has potential for automated image analysis of radiological images in forensic medicine.

Learned Helplessness: A Reply and an Alternative S-R Interpretation

ERIC Educational Resources Information Center

Levis, Donald J.

1976-01-01

Author attempts to provide a careful analysis of Maier and Seligman's manuscript (AA 522 796) on learned helplessness with the hope that such a critique will produce a positive effect by clarifying issues of contention and pinpointing weaknesses in need of correction. (Author/RK)

Hormone-Driven Kids: A Call for Social and Emotional Learning in the Middle School Years.

ERIC Educational Resources Information Center

Stern, Robin

1999-01-01

Describes the basic tenets of social and emotional learning; discusses the emotional and social challenges that middle school children face; and provides suggestions for parents and teachers to help promote smooth sailing through these rocky middle school years. (SR)

Dated eclogitic diamond growth zones reveal variable recycling of crustal carbon through time

NASA Astrophysics Data System (ADS)

Timmerman, S.; Koornneef, J. M.; Chinn, I. L.; Davies, G. R.

2017-04-01

Monocrystalline diamonds commonly record complex internal structures reflecting episodic growth linked to changing carbon-bearing fluids in the mantle. Using diamonds to trace the evolution of the deep carbon cycle therefore requires dating of individual diamond growth zones. To this end Rb-Sr and Sm-Nd isotope data are presented from individual eclogitic silicate inclusions from the Orapa and Letlhakane diamond mines, Botswana. δ13 C values are reported from the host diamond growth zones. Heterogeneous 87Sr/86Sr ratios (0.7033-0.7097) suggest inclusion formation in multiple and distinct tectono-magmatic environments. Sm-Nd isochron ages were determined based on groups of inclusions with similar trace element chemistry, Sr isotope ratios, and nitrogen aggregation of the host diamond growth zone. Diamond growth events at 0.14 ± 0.09, 0.25 ± 0.04, 1.1 ± 0.09, 1.70 ± 0.34 and 2.33 ± 0.02 Ga can be directly related to regional tectono-magmatic events. Individual diamonds record episodic growth with age differences of up to 2 Ga. Dated diamond zones have variable δ13 C values (-5.0 to -33.6‰ vs PDB) and appear to imply changes in subducted material over time. The studied Botswanan diamonds are interpreted to have formed in different tectono-magmatic environments that involve mixing of carbon from three sources that represent: i) subducted biogenic sediments (lightest δ13 C, low 87Sr/86Sr); ii) subducted carbonate-rich sediments (heavy δ13 C, high 87Sr/86Sr) and iii) depleted upper mantle (heavy δ13 C, low 87Sr/86Sr). We infer that older diamonds from these two localities are more likely to have light δ13 C due to greater subduction of biogenic sediments that may be related to hotter and more reduced conditions in the Archaean before the Great Oxidation Event at 2.3 Ga. These findings imply a marked temporal change in the nature of subducted carbon beneath Botswana and warrant further study to establish if this is a global phenomenon.

Spatio-temporal variability of modern sedimentation rates in Lake Nam Co, central Tibetan Plateau, China -- the first results from sediment traps

NASA Astrophysics Data System (ADS)

Wang, J.; Ju, J.; Daut, G.; Wang, Y.; Maeusbacher, R.; Zhu, L.

2013-12-01

As a big and deep lake in high altitude environment, Nam Co has played an important role in the past decade concerning paleoenvironmental change study. However, the modern process monitoring research is still insufficient in this lake to understand the variations in the modern sedimentation patterns. Sediment traps are widely used in lakes monitoring and research, providing the modern sedimentation rates (SR) and flux information as well as the materials for multidisciplinary studies. Here we present the first and preliminary result of spatio-temporal variability of SR in Nam Co based on one-year sediment traps data. Three integrated self-made traps mooring were deployed in different areas in Nam Co, which were eastern area (T1, ~57m depth), middle area (T2, ~93m depth) and western area (T3, ~62m depth). There were three layers traps in T1 and T3 station while four layers in T2 station. Additionally, a time-series automatic samples changing trap (Technicap PPS 3/3, France) was set up in the bottom (~90m depth) of T2 station with a sampling interval of two weeks. All traps were established in late May, 2012 and collected in Mid-September, 2012 for the first time. Then after winter time, samples were again collected in late May, 2013. Therefore, we got results for two periods, namely summer half year (May-September) and winter half year (September-next May). The results showed remarkable variation of SR vertically in all three stations, the bottom layers received much more materials than the up and middle layers. This fact could be attributed to the distinct influence of high density flows occurring at the lake bottom. This is also supported by multiprobe measurements showing high turbidity in the water body close to the bottom. In shallow areas (T1 and T3) the SR were higher than that of deep area (T2), which could probably reflect the different distance from the terrestrial source to the sites where the traps were deployed. In T1 and T2 stations, SR of winter half year (calculated as mg/cm2/day) was much higher than summer half year and this trend was also partly detected in the time-series sediment trap (T2), which showed higher SR in October, November and early June (no data from December to May). From early June to mid-November, the average SR of T2 station (~90m depth) ranged 0.09-0.95 mg/cm2/day, showed a remarkable temporal variation. More data and detailed analysis are still needed to elucidate the variability of modern SR in Nam Co and the influencing factors, including some internal mechanisms and outside driving related to climate change.

Deep Learning Accurately Predicts Estrogen Receptor Status in Breast Cancer Metabolomics Data.

PubMed

Alakwaa, Fadhl M; Chaudhary, Kumardeep; Garmire, Lana X

2018-01-05

Metabolomics holds the promise as a new technology to diagnose highly heterogeneous diseases. Conventionally, metabolomics data analysis for diagnosis is done using various statistical and machine learning based classification methods. However, it remains unknown if deep neural network, a class of increasingly popular machine learning methods, is suitable to classify metabolomics data. Here we use a cohort of 271 breast cancer tissues, 204 positive estrogen receptor (ER+), and 67 negative estrogen receptor (ER-) to test the accuracies of feed-forward networks, a deep learning (DL) framework, as well as six widely used machine learning models, namely random forest (RF), support vector machines (SVM), recursive partitioning and regression trees (RPART), linear discriminant analysis (LDA), prediction analysis for microarrays (PAM), and generalized boosted models (GBM). DL framework has the highest area under the curve (AUC) of 0.93 in classifying ER+/ER- patients, compared to the other six machine learning algorithms. Furthermore, the biological interpretation of the first hidden layer reveals eight commonly enriched significant metabolomics pathways (adjusted P-value <0.05) that cannot be discovered by other machine learning methods. Among them, protein digestion and absorption and ATP-binding cassette (ABC) transporters pathways are also confirmed in integrated analysis between metabolomics and gene expression data in these samples. In summary, deep learning method shows advantages for metabolomics based breast cancer ER status classification, with both the highest prediction accuracy (AUC = 0.93) and better revelation of disease biology. We encourage the adoption of feed-forward networks based deep learning method in the metabolomics research community for classification.

Robust stochastic resonance: Signal detection and adaptation in impulsive noise

NASA Astrophysics Data System (ADS)

Kosko, Bart; Mitaim, Sanya

2001-11-01

Stochastic resonance (SR) occurs when noise improves a system performance measure such as a spectral signal-to-noise ratio or a cross-correlation measure. All SR studies have assumed that the forcing noise has finite variance. Most have further assumed that the noise is Gaussian. We show that SR still occurs for the more general case of impulsive or infinite-variance noise. The SR effect fades as the noise grows more impulsive. We study this fading effect on the family of symmetric α-stable bell curves that includes the Gaussian bell curve as a special case. These bell curves have thicker tails as the parameter α falls from 2 (the Gaussian case) to 1 (the Cauchy case) to even lower values. Thicker tails create more frequent and more violent noise impulses. The main feedback and feedforward models in the SR literature show this fading SR effect for periodic forcing signals when we plot either the signal-to-noise ratio or a signal correlation measure against the dispersion of the α-stable noise. Linear regression shows that an exponential law γopt(α)=cAα describes this relation between the impulsive index α and the SR-optimal noise dispersion γopt. The results show that SR is robust against noise ``outliers.'' So SR may be more widespread in nature than previously believed. Such robustness also favors the use of SR in engineering systems. We further show that an adaptive system can learn the optimal noise dispersion for two standard SR models (the quartic bistable model and the FitzHugh-Nagumo neuron model) for the signal-to-noise ratio performance measure. This also favors practical applications of SR and suggests that evolution may have tuned the noise-sensitive parameters of biological systems.

Deep Restricted Kernel Machines Using Conjugate Feature Duality.

PubMed

Suykens, Johan A K

2017-08-01

The aim of this letter is to propose a theory of deep restricted kernel machines offering new foundations for deep learning with kernel machines. From the viewpoint of deep learning, it is partially related to restricted Boltzmann machines, which are characterized by visible and hidden units in a bipartite graph without hidden-to-hidden connections and deep learning extensions as deep belief networks and deep Boltzmann machines. From the viewpoint of kernel machines, it includes least squares support vector machines for classification and regression, kernel principal component analysis (PCA), matrix singular value decomposition, and Parzen-type models. A key element is to first characterize these kernel machines in terms of so-called conjugate feature duality, yielding a representation with visible and hidden units. It is shown how this is related to the energy form in restricted Boltzmann machines, with continuous variables in a nonprobabilistic setting. In this new framework of so-called restricted kernel machine (RKM) representations, the dual variables correspond to hidden features. Deep RKM are obtained by coupling the RKMs. The method is illustrated for deep RKM, consisting of three levels with a least squares support vector machine regression level and two kernel PCA levels. In its primal form also deep feedforward neural networks can be trained within this framework.

Distributed Cerebellar Motor Learning: A Spike-Timing-Dependent Plasticity Model

PubMed Central

Luque, Niceto R.; Garrido, Jesús A.; Naveros, Francisco; Carrillo, Richard R.; D'Angelo, Egidio; Ros, Eduardo

2016-01-01

Deep cerebellar nuclei neurons receive both inhibitory (GABAergic) synaptic currents from Purkinje cells (within the cerebellar cortex) and excitatory (glutamatergic) synaptic currents from mossy fibers. Those two deep cerebellar nucleus inputs are thought to be also adaptive, embedding interesting properties in the framework of accurate movements. We show that distributed spike-timing-dependent plasticity mechanisms (STDP) located at different cerebellar sites (parallel fibers to Purkinje cells, mossy fibers to deep cerebellar nucleus cells, and Purkinje cells to deep cerebellar nucleus cells) in close-loop simulations provide an explanation for the complex learning properties of the cerebellum in motor learning. Concretely, we propose a new mechanistic cerebellar spiking model. In this new model, deep cerebellar nuclei embed a dual functionality: deep cerebellar nuclei acting as a gain adaptation mechanism and as a facilitator for the slow memory consolidation at mossy fibers to deep cerebellar nucleus synapses. Equipping the cerebellum with excitatory (e-STDP) and inhibitory (i-STDP) mechanisms at deep cerebellar nuclei afferents allows the accommodation of synaptic memories that were formed at parallel fibers to Purkinje cells synapses and then transferred to mossy fibers to deep cerebellar nucleus synapses. These adaptive mechanisms also contribute to modulate the deep-cerebellar-nucleus-output firing rate (output gain modulation toward optimizing its working range). PMID:26973504

Research on Daily Objects Detection Based on Deep Neural Network

NASA Astrophysics Data System (ADS)

Ding, Sheng; Zhao, Kun

2018-03-01

With the rapid development of deep learning, great breakthroughs have been made in the field of object detection. In this article, the deep learning algorithm is applied to the detection of daily objects, and some progress has been made in this direction. Compared with traditional object detection methods, the daily objects detection method based on deep learning is faster and more accurate. The main research work of this article: 1. collect a small data set of daily objects; 2. in the TensorFlow framework to build different models of object detection, and use this data set training model; 3. the training process and effect of the model are improved by fine-tuning the model parameters.

Surface, Deep, and Transfer? Considering the Role of Content Literacy Instructional Strategies

ERIC Educational Resources Information Center

Frey, Nancy; Fisher, Douglas; Hattie, John

2017-01-01

This article provides an organizational review of content literacy instructional strategies to forward a claim that some strategies work better for surface learning, whereas others are more effective for deep learning and still others for transfer learning. The authors argue that the failure to adopt content literacy strategies by disciplinary…

The Experience of Deep Learning by Accounting Students

ERIC Educational Resources Information Center

Turner, Martin; Baskerville, Rachel

2013-01-01

This study examines how to support accounting students to experience deep learning. A sample of 81 students in a third-year undergraduate accounting course was studied employing a phenomenographic research approach, using ten assessed learning tasks for each student (as well as a focus group and student surveys) to measure their experience of how…

Measuring Deep, Reflective Comprehension and Learning Strategies: Challenges and Successes

ERIC Educational Resources Information Center

McNamara, Danielle S.

2011-01-01

There is a heightened understanding that metacognition and strategy use are crucial to deep, long-lasting comprehension and learning, but their assessment is challenging. First, students' judgments of what their abilities and habits and measurements of their performance often do not match. Second, students tend to learn and comprehend differently…

What Can Be Learned from a Laboratory Model of Conceptual Change? Descriptive Findings and Methodological Issues

ERIC Educational Resources Information Center

Ohlsson, Stellan; Cosejo, David G.

2014-01-01

The problem of how people process novel and unexpected information--"deep learning" (Ohlsson in "Deep learning: how the mind overrides experience." Cambridge University Press, New York, 2011)--is central to several fields of research, including creativity, belief revision, and conceptual change. Researchers have not converged…

Getting Inside Knowledge: The Application of Entwistle's Model of Surface/Deep Processing in Producing Open Learning Materials.

ERIC Educational Resources Information Center

Evans, Barbara; Honour, Leslie

1997-01-01

Reports on a study that required student teachers training in business education to produce open learning materials on intercultural communication. Analysis of stages and responses to this assignment revealed a distinction between "deep" and "surface" learning. Includes charts delineating the characteristics of these two types…

A Critical Comparison of Transformation and Deep Approach Theories of Learning

ERIC Educational Resources Information Center

Howie, Peter; Bagnall, Richard

2015-01-01

This paper reports a critical comparative analysis of two popular and significant theories of adult learning: the transformation and the deep approach theories of learning. These theories are operative in different educational sectors, are significant, respectively, in each, and they may be seen as both touching on similar concerns with learning…

Who Benefits from a Low versus High Guidance CSCL Script and Why?

ERIC Educational Resources Information Center

Mende, Stephan; Proske, Antje; Körndle, Hermann; Narciss, Susanne

2017-01-01

Computer-supported collaborative learning (CSCL) scripts can foster learners' deep text comprehension. However, this depends on (a) the extent to which the learning activities targeted by a script promote deep text comprehension and (b) whether the guidance level provided by the script is adequate to induce the targeted learning activities…

Accumulation of artificial radionuclides in deep sediments of the Mediterranean Sea

NASA Astrophysics Data System (ADS)

Garcia-Orellana, J.; Sanchez-Cabeza, J. A.; Masque, P.; Costa, E.; Bruach, J. M.; Morist, A.; Luna, J. A.

2003-04-01

Concentrations and inventories of artificial radionuclides (90Sr, 137Cs and 239,40Pu) were determined in deep sediment cores (3.000 m) collected in the western and eastern basins of the Mediterranean Sea in the frame of the ADIOS project. Artificial radionuclides enter the Mediterranean Sea mainly though atmospheric deposition after nuclear weapons tests and the Chernobyl accident, but also through the river discharge of effluents of nuclear facilities (e.g. Rhone and Ebro rivers). The aim of this work is to investigate the degree by which pollutants are transferred to the deep environment of the Mediterranean Sea as a basis to elucidate their effects on benthic organisms. The mean inventories of 239+240Pu, 137Cs and 90Sr in the Western basin are 2.77 ± 0.26, 68 ± 12 and < 7 Bq\\cdotm-2 respectively and 3.29 ± 0.60, 115 ± 33 and 249±154 Bq\\cdotm-2 in the Eastern basin. The activity - depth profiles of 210Pb, together with 14C dating, indicate that sediment mixing redistributes the artificial radionuclides within the first 2 cm of the sedimentary column. Artificial radionuclides inventories in the deep-sea sediments were used to calculate the fraction of the total inventory of artificial radionuclides that is accumulated in the deep sea sediments after scavenging from the water column. Indeed, a balance of the radionuclide distributions in the water column allows evaluating the importance of lateral transport of particulate matter from the continental margins on the accumulation of artificial radionuclides in the deep, open Mediterranean Sea. This is achieved in i) comparison with reported data from coastal areas at different locations in the Mediterranean Sea, and ii) balance of the distribution of the natural radionuclide 210Pb in studied areas (vertical profiles of dissolved and particulate activities, fluxes determined by using sediment trap deployed at different depths and inventories in the bottom sediments). The results, taking into account radioactive decay and exchange fluxes through the Gibraltar Strait, permit to estimate the residence times of pollutants in the water column and predict future evolution of their distributions.

Machine Learning in Ultrasound Computer-Aided Diagnostic Systems: A Survey

PubMed Central

Zhang, Fan; Li, Xuelong

2018-01-01

The ultrasound imaging is one of the most common schemes to detect diseases in the clinical practice. There are many advantages of ultrasound imaging such as safety, convenience, and low cost. However, reading ultrasound imaging is not easy. To support the diagnosis of clinicians and reduce the load of doctors, many ultrasound computer-aided diagnosis (CAD) systems are proposed. In recent years, the success of deep learning in the image classification and segmentation led to more and more scholars realizing the potential of performance improvement brought by utilizing the deep learning in the ultrasound CAD system. This paper summarized the research which focuses on the ultrasound CAD system utilizing machine learning technology in recent years. This study divided the ultrasound CAD system into two categories. One is the traditional ultrasound CAD system which employed the manmade feature and the other is the deep learning ultrasound CAD system. The major feature and the classifier employed by the traditional ultrasound CAD system are introduced. As for the deep learning ultrasound CAD, newest applications are summarized. This paper will be useful for researchers who focus on the ultrasound CAD system. PMID:29687000

Geographical topic learning for social images with a deep neural network

NASA Astrophysics Data System (ADS)

Feng, Jiangfan; Xu, Xin

2017-03-01

The use of geographical tagging in social-media images is becoming a part of image metadata and a great interest for geographical information science. It is well recognized that geographical topic learning is crucial for geographical annotation. Existing methods usually exploit geographical characteristics using image preprocessing, pixel-based classification, and feature recognition. How to effectively exploit the high-level semantic feature and underlying correlation among different types of contents is a crucial task for geographical topic learning. Deep learning (DL) has recently demonstrated robust capabilities for image tagging and has been introduced into geoscience. It extracts high-level features computed from a whole image component, where the cluttered background may dominate spatial features in the deep representation. Therefore, a method of spatial-attentional DL for geographical topic learning is provided and we can regard it as a special case of DL combined with various deep networks and tuning tricks. Results demonstrated that the method is discriminative for different types of geographical topic learning. In addition, it outperforms other sequential processing models in a tagging task for a geographical image dataset.

Machine Learning in Ultrasound Computer-Aided Diagnostic Systems: A Survey.

PubMed

Huang, Qinghua; Zhang, Fan; Li, Xuelong

2018-01-01

The ultrasound imaging is one of the most common schemes to detect diseases in the clinical practice. There are many advantages of ultrasound imaging such as safety, convenience, and low cost. However, reading ultrasound imaging is not easy. To support the diagnosis of clinicians and reduce the load of doctors, many ultrasound computer-aided diagnosis (CAD) systems are proposed. In recent years, the success of deep learning in the image classification and segmentation led to more and more scholars realizing the potential of performance improvement brought by utilizing the deep learning in the ultrasound CAD system. This paper summarized the research which focuses on the ultrasound CAD system utilizing machine learning technology in recent years. This study divided the ultrasound CAD system into two categories. One is the traditional ultrasound CAD system which employed the manmade feature and the other is the deep learning ultrasound CAD system. The major feature and the classifier employed by the traditional ultrasound CAD system are introduced. As for the deep learning ultrasound CAD, newest applications are summarized. This paper will be useful for researchers who focus on the ultrasound CAD system.

Quantum-assisted Helmholtz machines: A quantum–classical deep learning framework for industrial datasets in near-term devices

NASA Astrophysics Data System (ADS)

Benedetti, Marcello; Realpe-Gómez, John; Perdomo-Ortiz, Alejandro

2018-07-01

Machine learning has been presented as one of the key applications for near-term quantum technologies, given its high commercial value and wide range of applicability. In this work, we introduce the quantum-assisted Helmholtz machine:a hybrid quantum–classical framework with the potential of tackling high-dimensional real-world machine learning datasets on continuous variables. Instead of using quantum computers only to assist deep learning, as previous approaches have suggested, we use deep learning to extract a low-dimensional binary representation of data, suitable for processing on relatively small quantum computers. Then, the quantum hardware and deep learning architecture work together to train an unsupervised generative model. We demonstrate this concept using 1644 quantum bits of a D-Wave 2000Q quantum device to model a sub-sampled version of the MNIST handwritten digit dataset with 16 × 16 continuous valued pixels. Although we illustrate this concept on a quantum annealer, adaptations to other quantum platforms, such as ion-trap technologies or superconducting gate-model architectures, could be explored within this flexible framework.

Application of deep learning to the classification of images from colposcopy.

PubMed

Sato, Masakazu; Horie, Koji; Hara, Aki; Miyamoto, Yuichiro; Kurihara, Kazuko; Tomio, Kensuke; Yokota, Harushige

2018-03-01

The objective of the present study was to investigate whether deep learning could be applied successfully to the classification of images from colposcopy. For this purpose, a total of 158 patients who underwent conization were enrolled, and medical records and data from the gynecological oncology database were retrospectively reviewed. Deep learning was performed with the Keras neural network and TensorFlow libraries. Using preoperative images from colposcopy as the input data and deep learning technology, the patients were classified into three groups [severe dysplasia, carcinoma in situ (CIS) and invasive cancer (IC)]. A total of 485 images were obtained for the analysis, of which 142 images were of severe dysplasia (2.9 images/patient), 257 were of CIS (3.3 images/patient), and 86 were of IC (4.1 images/patient). Of these, 233 images were captured with a green filter, and the remaining 252 were captured without a green filter. Following the application of L2 regularization, L1 regularization, dropout and data augmentation, the accuracy of the validation dataset was ~50%. Although the present study is preliminary, the results indicated that deep learning may be applied to classify colposcopy images.

Fault Diagnosis for Rotating Machinery Using Vibration Measurement Deep Statistical Feature Learning.

PubMed

Li, Chuan; Sánchez, René-Vinicio; Zurita, Grover; Cerrada, Mariela; Cabrera, Diego

2016-06-17

Fault diagnosis is important for the maintenance of rotating machinery. The detection of faults and fault patterns is a challenging part of machinery fault diagnosis. To tackle this problem, a model for deep statistical feature learning from vibration measurements of rotating machinery is presented in this paper. Vibration sensor signals collected from rotating mechanical systems are represented in the time, frequency, and time-frequency domains, each of which is then used to produce a statistical feature set. For learning statistical features, real-value Gaussian-Bernoulli restricted Boltzmann machines (GRBMs) are stacked to develop a Gaussian-Bernoulli deep Boltzmann machine (GDBM). The suggested approach is applied as a deep statistical feature learning tool for both gearbox and bearing systems. The fault classification performances in experiments using this approach are 95.17% for the gearbox, and 91.75% for the bearing system. The proposed approach is compared to such standard methods as a support vector machine, GRBM and a combination model. In experiments, the best fault classification rate was detected using the proposed model. The results show that deep learning with statistical feature extraction has an essential improvement potential for diagnosing rotating machinery faults.

Fault Diagnosis for Rotating Machinery Using Vibration Measurement Deep Statistical Feature Learning

PubMed Central

Li, Chuan; Sánchez, René-Vinicio; Zurita, Grover; Cerrada, Mariela; Cabrera, Diego

2016-01-01

Fault diagnosis is important for the maintenance of rotating machinery. The detection of faults and fault patterns is a challenging part of machinery fault diagnosis. To tackle this problem, a model for deep statistical feature learning from vibration measurements of rotating machinery is presented in this paper. Vibration sensor signals collected from rotating mechanical systems are represented in the time, frequency, and time-frequency domains, each of which is then used to produce a statistical feature set. For learning statistical features, real-value Gaussian-Bernoulli restricted Boltzmann machines (GRBMs) are stacked to develop a Gaussian-Bernoulli deep Boltzmann machine (GDBM). The suggested approach is applied as a deep statistical feature learning tool for both gearbox and bearing systems. The fault classification performances in experiments using this approach are 95.17% for the gearbox, and 91.75% for the bearing system. The proposed approach is compared to such standard methods as a support vector machine, GRBM and a combination model. In experiments, the best fault classification rate was detected using the proposed model. The results show that deep learning with statistical feature extraction has an essential improvement potential for diagnosing rotating machinery faults. PMID:27322273

Application of deep learning to the classification of images from colposcopy

PubMed Central

Sato, Masakazu; Horie, Koji; Hara, Aki; Miyamoto, Yuichiro; Kurihara, Kazuko; Tomio, Kensuke; Yokota, Harushige

2018-01-01

The objective of the present study was to investigate whether deep learning could be applied successfully to the classification of images from colposcopy. For this purpose, a total of 158 patients who underwent conization were enrolled, and medical records and data from the gynecological oncology database were retrospectively reviewed. Deep learning was performed with the Keras neural network and TensorFlow libraries. Using preoperative images from colposcopy as the input data and deep learning technology, the patients were classified into three groups [severe dysplasia, carcinoma in situ (CIS) and invasive cancer (IC)]. A total of 485 images were obtained for the analysis, of which 142 images were of severe dysplasia (2.9 images/patient), 257 were of CIS (3.3 images/patient), and 86 were of IC (4.1 images/patient). Of these, 233 images were captured with a green filter, and the remaining 252 were captured without a green filter. Following the application of L2 regularization, L1 regularization, dropout and data augmentation, the accuracy of the validation dataset was ~50%. Although the present study is preliminary, the results indicated that deep learning may be applied to classify colposcopy images. PMID:29456725

Deep learning based tissue analysis predicts outcome in colorectal cancer.

PubMed

Bychkov, Dmitrii; Linder, Nina; Turkki, Riku; Nordling, Stig; Kovanen, Panu E; Verrill, Clare; Walliander, Margarita; Lundin, Mikael; Haglund, Caj; Lundin, Johan

2018-02-21

Image-based machine learning and deep learning in particular has recently shown expert-level accuracy in medical image classification. In this study, we combine convolutional and recurrent architectures to train a deep network to predict colorectal cancer outcome based on images of tumour tissue samples. The novelty of our approach is that we directly predict patient outcome, without any intermediate tissue classification. We evaluate a set of digitized haematoxylin-eosin-stained tumour tissue microarray (TMA) samples from 420 colorectal cancer patients with clinicopathological and outcome data available. The results show that deep learning-based outcome prediction with only small tissue areas as input outperforms (hazard ratio 2.3; CI 95% 1.79-3.03; AUC 0.69) visual histological assessment performed by human experts on both TMA spot (HR 1.67; CI 95% 1.28-2.19; AUC 0.58) and whole-slide level (HR 1.65; CI 95% 1.30-2.15; AUC 0.57) in the stratification into low- and high-risk patients. Our results suggest that state-of-the-art deep learning techniques can extract more prognostic information from the tissue morphology of colorectal cancer than an experienced human observer.

Automatic segmentation of the prostate on CT images using deep learning and multi-atlas fusion

NASA Astrophysics Data System (ADS)

Ma, Ling; Guo, Rongrong; Zhang, Guoyi; Tade, Funmilayo; Schuster, David M.; Nieh, Peter; Master, Viraj; Fei, Baowei

2017-02-01

Automatic segmentation of the prostate on CT images has many applications in prostate cancer diagnosis and therapy. However, prostate CT image segmentation is challenging because of the low contrast of soft tissue on CT images. In this paper, we propose an automatic segmentation method by combining a deep learning method and multi-atlas refinement. First, instead of segmenting the whole image, we extract the region of interesting (ROI) to delete irrelevant regions. Then, we use the convolutional neural networks (CNN) to learn the deep features for distinguishing the prostate pixels from the non-prostate pixels in order to obtain the preliminary segmentation results. CNN can automatically learn the deep features adapting to the data, which are different from some handcrafted features. Finally, we select some similar atlases to refine the initial segmentation results. The proposed method has been evaluated on a dataset of 92 prostate CT images. Experimental results show that our method achieved a Dice similarity coefficient of 86.80% as compared to the manual segmentation. The deep learning based method can provide a useful tool for automatic segmentation of the prostate on CT images and thus can have a variety of clinical applications.

Applications of Deep Learning and Reinforcement Learning to Biological Data.

PubMed

Mahmud, Mufti; Kaiser, Mohammed Shamim; Hussain, Amir; Vassanelli, Stefano

2018-06-01

Rapid advances in hardware-based technologies during the past decades have opened up new possibilities for life scientists to gather multimodal data in various application domains, such as omics, bioimaging, medical imaging, and (brain/body)-machine interfaces. These have generated novel opportunities for development of dedicated data-intensive machine learning techniques. In particular, recent research in deep learning (DL), reinforcement learning (RL), and their combination (deep RL) promise to revolutionize the future of artificial intelligence. The growth in computational power accompanied by faster and increased data storage, and declining computing costs have already allowed scientists in various fields to apply these techniques on data sets that were previously intractable owing to their size and complexity. This paper provides a comprehensive survey on the application of DL, RL, and deep RL techniques in mining biological data. In addition, we compare the performances of DL techniques when applied to different data sets across various application domains. Finally, we outline open issues in this challenging research area and discuss future development perspectives.

Deep Learning for Flow Sculpting: Insights into Efficient Learning using Scientific Simulation Data

NASA Astrophysics Data System (ADS)

Stoecklein, Daniel; Lore, Kin Gwn; Davies, Michael; Sarkar, Soumik; Ganapathysubramanian, Baskar

2017-04-01

A new technique for shaping microfluid flow, known as flow sculpting, offers an unprecedented level of passive fluid flow control, with potential breakthrough applications in advancing manufacturing, biology, and chemistry research at the microscale. However, efficiently solving the inverse problem of designing a flow sculpting device for a desired fluid flow shape remains a challenge. Current approaches struggle with the many-to-one design space, requiring substantial user interaction and the necessity of building intuition, all of which are time and resource intensive. Deep learning has emerged as an efficient function approximation technique for high-dimensional spaces, and presents a fast solution to the inverse problem, yet the science of its implementation in similarly defined problems remains largely unexplored. We propose that deep learning methods can completely outpace current approaches for scientific inverse problems while delivering comparable designs. To this end, we show how intelligent sampling of the design space inputs can make deep learning methods more competitive in accuracy, while illustrating their generalization capability to out-of-sample predictions.

Phenotypic Antimicrobial Susceptibility Testing with Deep Learning Video Microscopy.

PubMed

Yu, Hui; Jing, Wenwen; Iriya, Rafael; Yang, Yunze; Syal, Karan; Mo, Manni; Grys, Thomas E; Haydel, Shelley E; Wang, Shaopeng; Tao, Nongjian

2018-05-15

Timely determination of antimicrobial susceptibility for a bacterial infection enables precision prescription, shortens treatment time, and helps minimize the spread of antibiotic resistant infections. Current antimicrobial susceptibility testing (AST) methods often take several days and thus impede these clinical and health benefits. Here, we present an AST method by imaging freely moving bacterial cells in urine in real time and analyzing the videos with a deep learning algorithm. The deep learning algorithm determines if an antibiotic inhibits a bacterial cell by learning multiple phenotypic features of the cell without the need for defining and quantifying each feature. We apply the method to urinary tract infection, a common infection that affects millions of people, to determine the minimum inhibitory concentration of pathogens from both bacteria spiked urine and clinical infected urine samples for different antibiotics within 30 min and validate the results with the gold standard broth macrodilution method. The deep learning video microscopy-based AST holds great potential to contribute to the solution of increasing drug-resistant infections.

Geochemical modeling of subsurface fluid generation in the Gulf of Cadiz

NASA Astrophysics Data System (ADS)

Schmidt, Christopher; Hensen, Christian; Wallmann, Klaus

2016-04-01

During RV METEOR cruise M86/5 in 2012 a number of deep-sea mud volcanoes were discovered at about 4500 m water depth west of the deformation front of the accretionary wedge in the Gulf of Cadiz (NE Atlantic). Fluid flow at these locations is mediated by an active strike-slip fault marking the transcurrent plate boundary between Africa and Eurasia. Geochemical signals of emanating fluids have been interpreted as being a mixture of various deep-sourced processes such as the alteration of oceanic crust, clay-mineral dehydration, and recrystallization of carbonaceous, Upper Jurassic sediments (Hensen et al. 2015). In the current study we present results of a geochemical reactive-transport model that was designed to simulate major fluid-affecting processes, such as the smectite to illite transformation or recrystallization of carbonates in order to provide a proof of concept. Preliminary results show that the model is able to reproduce pore water signatures (e.g. for chloride, strontium, 87Sr/86Sr) in subsurface sediments that are similar to those of MV fluids. Hensen, C., Scholz, F., Nuzzo, M., Valadares, V., Gràcia, E., Terrinha, P., Liebetrau, V., Kaul, N., Silva, S., Martínez-Loriente, S., Bartolome, R., Piñero, E., Magalhães, V.H., Schmidt, M., Weise, S.M., Cunha, M., Hilario, A., Perea, H., Rovelli, L. and Lackschewitz, K. (2015) Strike-slip faults mediate the rise of crustal-derived fluids and mud volcanism in the deep sea. Geology 43, 339-342.

Sex-Specific Effects on Spatial Learning and Memory, and Sex-Independent Effects on Blood Pressure of a <3.3 Mbp Rat Chromosome 2 QTL Region in Dahl Salt-Sensitive Rats

PubMed Central

Herrera, Victoria L.; Pasion, Khristine A.; Tan, Glaiza A.; Moran, Ann Marie; Ruiz-Opazo, Nelson

2013-01-01

Epidemiological studies have consistently found that hypertension is associated with poor cognitive performance. We hypothesize that a putative causal mechanism underlying this association is due to genetic loci affecting both blood pressure and cognition. Consistent with this notion, we reported several blood pressure (BP) quantitative trait loci (QTLs) that co-localized with navigational performance (Nav)-QTLs influencing spatial learning and memory in Dahl rats. The present study investigates a chromosome 2 region harboring BP-f4 and Nav-8 QTLs. We developed two congenic strains, S.R2A and S.R2B introgressing Dahl R-chromosome 2 segments into Dahl S chromosome 2 region spanning BP-f4 and Nav-8 QTLs. Radiotelemetric blood pressure analysis identified only S.R2A congenic rats with lower systolic blood pressure (females: −26.0 mmHg, P = 0.003; males: −30.9 mmHg, P<1×10−5), diastolic blood pressure (females: −21.2 mmHg, P = 0.01; males: −25.7 mmHg, P<1×10−5), and mean arterial pressure (females: −23.9 mmHg, P = 0.004; males: −28.0 mmHg, P<1×10−5) compared with corresponding Dahl S controls, confirming the presence of BP-f4 QTL on rat chromosome 2. The S.R2B congenic segment did not affect blood pressure. Testing of S.R2A, S.R2B, and Dahl S male rats in the Morris water maze (MWM) task revealed significantly decreased spatial navigation performance in S.R2A male congenic rats when compared with Dahl S male controls (P<0.05). The S.R2B congenic segment did not affect performance of the MWM task in males. The S.R2A female rats did not differ in spatial navigation when compared with Dahl S female controls, indicating that the Nav-8 effect on spatial navigation is male-specific. Our results suggest the existence of a single QTL on chromosome 2 176.6–179.9 Mbp region which affects blood pressure in both males and females and cognition solely in males. PMID:23861781

Sex-specific effects on spatial learning and memory, and sex-independent effects on blood pressure of a <3.3 Mbp rat chromosome 2 QTL region in Dahl salt-sensitive rats.

PubMed

Herrera, Victoria L; Pasion, Khristine A; Tan, Glaiza A; Moran, Ann Marie; Ruiz-Opazo, Nelson

2013-01-01

Epidemiological studies have consistently found that hypertension is associated with poor cognitive performance. We hypothesize that a putative causal mechanism underlying this association is due to genetic loci affecting both blood pressure and cognition. Consistent with this notion, we reported several blood pressure (BP) quantitative trait loci (QTLs) that co-localized with navigational performance (Nav)-QTLs influencing spatial learning and memory in Dahl rats. The present study investigates a chromosome 2 region harboring BP-f4 and Nav-8 QTLs. We developed two congenic strains, S.R2A and S.R2B introgressing Dahl R-chromosome 2 segments into Dahl S chromosome 2 region spanning BP-f4 and Nav-8 QTLs. Radiotelemetric blood pressure analysis identified only S.R2A congenic rats with lower systolic blood pressure (females: -26.0 mmHg, P = 0.003; males: -30.9 mmHg, P<1×10(-5)), diastolic blood pressure (females: -21.2 mmHg, P = 0.01; males: -25.7 mmHg, P<1×10(-5)), and mean arterial pressure (females: -23.9 mmHg, P = 0.004; males: -28.0 mmHg, P<1×10(-5)) compared with corresponding Dahl S controls, confirming the presence of BP-f4 QTL on rat chromosome 2. The S.R2B congenic segment did not affect blood pressure. Testing of S.R2A, S.R2B, and Dahl S male rats in the Morris water maze (MWM) task revealed significantly decreased spatial navigation performance in S.R2A male congenic rats when compared with Dahl S male controls (P<0.05). The S.R2B congenic segment did not affect performance of the MWM task in males. The S.R2A female rats did not differ in spatial navigation when compared with Dahl S female controls, indicating that the Nav-8 effect on spatial navigation is male-specific. Our results suggest the existence of a single QTL on chromosome 2 176.6-179.9 Mbp region which affects blood pressure in both males and females and cognition solely in males.

Developments in μSR and β NMR: Beyond a Muon Lifetime

NASA Astrophysics Data System (ADS)

Kiefl, Robert F.

Advances in the use of μSR and β-NMR are driven by technical developments. New methods were developed which allowed us to learn surprising things about muonium in semiconductors, its electronic structure, its relationship to hydrogen, its ability to diffuse via quantum tunneling, and its metastability. Similarly in the area of high Tc superconductors new capabilities in spectrometer design led to new information on the properties of superconducting vortices and how they interact. The development of low energy β-NMR at TRIUMF and LE-μSR at PSI has made it possible to study electronic and magnetic properties of thin films and interfaces where conventional NMR lacks the required sensitivity. Low energy β-NMR is almost identical to μSR in principle, but the longer lifetime of 8Li allows one to probe the system on a very different time scale. In this sense β-NMR can be viewed as a complement or extension of μSR.

Large-scale Exploration of Neuronal Morphologies Using Deep Learning and Augmented Reality.

PubMed

Li, Zhongyu; Butler, Erik; Li, Kang; Lu, Aidong; Ji, Shuiwang; Zhang, Shaoting

2018-02-12

Recently released large-scale neuron morphological data has greatly facilitated the research in neuroinformatics. However, the sheer volume and complexity of these data pose significant challenges for efficient and accurate neuron exploration. In this paper, we propose an effective retrieval framework to address these problems, based on frontier techniques of deep learning and binary coding. For the first time, we develop a deep learning based feature representation method for the neuron morphological data, where the 3D neurons are first projected into binary images and then learned features using an unsupervised deep neural network, i.e., stacked convolutional autoencoders (SCAEs). The deep features are subsequently fused with the hand-crafted features for more accurate representation. Considering the exhaustive search is usually very time-consuming in large-scale databases, we employ a novel binary coding method to compress feature vectors into short binary codes. Our framework is validated on a public data set including 58,000 neurons, showing promising retrieval precision and efficiency compared with state-of-the-art methods. In addition, we develop a novel neuron visualization program based on the techniques of augmented reality (AR), which can help users take a deep exploration of neuron morphologies in an interactive and immersive manner.

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.

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.

Generation and Evolution of Quaternary Magmas Beneath Tengchong: Sr-Nd-Pb-Hf Isotope and Zircon U-series Age Constraints

NASA Astrophysics Data System (ADS)

Zou, H.; Ma, M.; Fan, Q.; Xu, B.; Li, S. Q.; Zhao, Y.; King, D. T., Jr.

2017-12-01

The Tengchong volcanic field on the southeastern margin of the Tibetan Plateau represents rare Quaternary volcanic eruptions on the plateau. The Quaternary Tengchong volcanic field formed high-potassium calc-alkaline volcanic rocks that include trachybasalts, basaltic trachyandesites, trachyandesites, and dacites. Herein, we present comprehensive Nd-Sr-Pb-Hf isotopic and elemental data for trachybasalts, basaltic trachyandesites, and trachyandesites from four young Tengchong volcanoes at Maanshan, Dayingshan, Heikongshan, and Laoguipo, in order to understand their magma genesis and evolution. Nd-Sr-Pb-Hf isotopes for the primitive Tengchong magma (trachybasalts with SiO2 <52.5 wt. % and MgO >5.5% wt. %) reflect a heterogeneous enriched mantle source. High Th/U, Th/Ta, and Rb/Nb ratios and Nd-Sr-Pb-Hf isotope characteristics of the primitive magmas suggest that the enriched mantle beneath Tengchong formed as a result of subduction of clay-rich sediments, which probably came from the Indian continental plate. Partial melting of the enriched mantle was generated by deep continental subduction coupled with recent regional extension in the Tengchong area. With regard to the evolved magmas (basaltic trachyandesites and trachyandesites), good correlations between SiO2 content and the ratios 87Sr/86Sr, 143Nd/144Nd, 206Pb/204Pb, and 177Hf/176Hf strongly suggest that the combined assimilation and fractional crystallization (AFC) was an important process during magma evolution to form these basaltic trachyandesites and trachyandesites. Uranium-series zircon dating on these evolved lavas from Tengchong is used to constrain their magma evolution and residence timescales.

Deep Hashing for Scalable Image Search.

PubMed

Lu, Jiwen; Liong, Venice Erin; Zhou, Jie

2017-05-01

In this paper, we propose a new deep hashing (DH) approach to learn compact binary codes for scalable image search. Unlike most existing binary codes learning methods, which usually seek a single linear projection to map each sample into a binary feature vector, we develop a deep neural network to seek multiple hierarchical non-linear transformations to learn these binary codes, so that the non-linear relationship of samples can be well exploited. Our model is learned under three constraints at the top layer of the developed deep network: 1) the loss between the compact real-valued code and the learned binary vector is minimized, 2) the binary codes distribute evenly on each bit, and 3) different bits are as independent as possible. To further improve the discriminative power of the learned binary codes, we extend DH into supervised DH (SDH) and multi-label SDH by including a discriminative term into the objective function of DH, which simultaneously maximizes the inter-class variations and minimizes the intra-class variations of the learned binary codes with the single-label and multi-label settings, respectively. Extensive experimental results on eight widely used image search data sets show that our proposed methods achieve very competitive results with the state-of-the-arts.

ROOFN3D: Deep Learning Training Data for 3d Building Reconstruction

NASA Astrophysics Data System (ADS)

Wichmann, A.; Agoub, A.; Kada, M.

2018-05-01

Machine learning methods have gained in importance through the latest development of artificial intelligence and computer hardware. Particularly approaches based on deep learning have shown that they are able to provide state-of-the-art results for various tasks. However, the direct application of deep learning methods to improve the results of 3D building reconstruction is often not possible due, for example, to the lack of suitable training data. To address this issue, we present RoofN3D which provides a new 3D point cloud training dataset that can be used to train machine learning models for different tasks in the context of 3D building reconstruction. It can be used, among others, to train semantic segmentation networks or to learn the structure of buildings and the geometric model construction. Further details about RoofN3D and the developed data preparation framework, which enables the automatic derivation of training data, are described in this paper. Furthermore, we provide an overview of other available 3D point cloud training data and approaches from current literature in which solutions for the application of deep learning to unstructured and not gridded 3D point cloud data are presented.

DeepSynergy: predicting anti-cancer drug synergy with Deep Learning

PubMed Central

Preuer, Kristina; Lewis, Richard P I; Hochreiter, Sepp; Bender, Andreas; Bulusu, Krishna C; Klambauer, Günter

2018-01-01

Abstract Motivation While drug combination therapies are a well-established concept in cancer treatment, identifying novel synergistic combinations is challenging due to the size of combinatorial space. However, computational approaches have emerged as a time- and cost-efficient way to prioritize combinations to test, based on recently available large-scale combination screening data. Recently, Deep Learning has had an impact in many research areas by achieving new state-of-the-art model performance. However, Deep Learning has not yet been applied to drug synergy prediction, which is the approach we present here, termed DeepSynergy. DeepSynergy uses chemical and genomic information as input information, a normalization strategy to account for input data heterogeneity, and conical layers to model drug synergies. Results DeepSynergy was compared to other machine learning methods such as Gradient Boosting Machines, Random Forests, Support Vector Machines and Elastic Nets on the largest publicly available synergy dataset with respect to mean squared error. DeepSynergy significantly outperformed the other methods with an improvement of 7.2% over the second best method at the prediction of novel drug combinations within the space of explored drugs and cell lines. At this task, the mean Pearson correlation coefficient between the measured and the predicted values of DeepSynergy was 0.73. Applying DeepSynergy for classification of these novel drug combinations resulted in a high predictive performance of an AUC of 0.90. Furthermore, we found that all compared methods exhibit low predictive performance when extrapolating to unexplored drugs or cell lines, which we suggest is due to limitations in the size and diversity of the dataset. We envision that DeepSynergy could be a valuable tool for selecting novel synergistic drug combinations. Availability and implementation DeepSynergy is available via www.bioinf.jku.at/software/DeepSynergy. Contact klambauer@bioinf.jku.at Supplementary information Supplementary data are available at Bioinformatics online. PMID:29253077

Deep learning based syndrome diagnosis of chronic gastritis.

PubMed

Liu, Guo-Ping; Yan, Jian-Jun; Wang, Yi-Qin; Zheng, Wu; Zhong, Tao; Lu, Xiong; Qian, Peng

2014-01-01

In Traditional Chinese Medicine (TCM), most of the algorithms used to solve problems of syndrome diagnosis are superficial structure algorithms and not considering the cognitive perspective from the brain. However, in clinical practice, there is complex and nonlinear relationship between symptoms (signs) and syndrome. So we employed deep leaning and multilabel learning to construct the syndrome diagnostic model for chronic gastritis (CG) in TCM. The results showed that deep learning could improve the accuracy of syndrome recognition. Moreover, the studies will provide a reference for constructing syndrome diagnostic models and guide clinical practice.

Deep Learning Based Syndrome Diagnosis of Chronic Gastritis

PubMed Central

Liu, Guo-Ping; Wang, Yi-Qin; Zheng, Wu; Zhong, Tao; Lu, Xiong; Qian, Peng

2014-01-01

In Traditional Chinese Medicine (TCM), most of the algorithms used to solve problems of syndrome diagnosis are superficial structure algorithms and not considering the cognitive perspective from the brain. However, in clinical practice, there is complex and nonlinear relationship between symptoms (signs) and syndrome. So we employed deep leaning and multilabel learning to construct the syndrome diagnostic model for chronic gastritis (CG) in TCM. The results showed that deep learning could improve the accuracy of syndrome recognition. Moreover, the studies will provide a reference for constructing syndrome diagnostic models and guide clinical practice. PMID:24734118

Impact of Scots pine (Pinus sylvestris L.) plantings on long term (137)Cs and (90)Sr recycling from a waste burial site in the Chernobyl Red Forest.

PubMed

Thiry, Yves; Colle, Claude; Yoschenko, Vasyl; Levchuk, Svjatoslav; Van Hees, May; Hurtevent, Pierre; Kashparov, Valery

2009-12-01

Plantings of Scots pine (Pinus sylvestris L.) on a waste burial site in the Chernobyl Red Forest was shown to greatly influence the long term redistribution of radioactivity contained in sub-surfaces trenches. After 15 years of growth, aboveground biomass of the average tree growing on waste trench no.22 had accumulated 1.7 times more (137)Cs than that of trees growing off the trench, and 5.4 times more (90)Sr. At the scale of the trench and according to an average tree density of 3300 trees/ha for the study zone, tree contamination would correspond to 0.024% of the (137)Cs and 2.52% of the (90)Sr contained in the buried waste material. A quantitative description of the radionuclide cycling showed a potential for trees to annually extract up to 0.82% of the (90)Sr pool in the trench and 0.0038% of the (137)Cs. A preferential (90)Sr uptake from the deep soil is envisioned while pine roots would take up (137)Cs mostly from less contaminated shallow soil layers. The current upward flux of (90)Sr through vegetation appeared at least equal to downward loss in waste material leaching as reported by Dewiere et al. (2004, Journal of Environmental Radioactivity 74, 139-150). Using a prospective calculation model, we estimated that maximum (90)Sr cycling can be expected to occur at 40 years post-planting, resulting in 12% of the current (90)Sr content in the trench transferred to surface soils through biomass turnover and 7% stored in tree biomass. These results are preliminary, although based on accurate methodology. A more integrated ecosystem study leading to the coupling between biological and geochemical models of radionuclide cycling within the Red Forest seems opportune. Such a study would help in the adequate management of that new forest and the waste trenches upon which they reside.

The Biological Nature of Geochemical Proxies: algal symbionts affect coral skeletal chemistry

NASA Astrophysics Data System (ADS)

Owens, K.; Cohen, A. L.; Shimizu, N.

2001-12-01

The strontium-calcium ratio (Sr/Ca) of reef coral skeleton is an important ocean temperature proxy that has been used to address some particularly controversial climate change issues. However, the paleothermometer has sometimes proven unreliable and there are indications that the temperature-dependence of Sr/Ca in coral aragonite is linked to the photosynthetic activity of algal symbionts (zooxanthellae) in coral tissue. We examined the effect of algal symbiosis on skeletal chemistry using Astrangia danae, a small colonial temperate scleractinian that occurs naturally with and without zooxanthellae. Live symbiotic (deep brown) and asymbiotic (white) colonies of similar size were collected in Woods Hole where water temperatures fluctuate seasonally between -2oC and 23oC. We used a microbeam technique (Secondary Ion Mass Spectrometry) and a 30 micron diameter sampling beam to construct high-resolution Sr/Ca profiles, 2500 microns long, down the growth axes of the outer calical (thecal) walls. Profiles generated from co-occuring symbiotic and asymbiotic colonies are remarkably different despite their exposure to identical water temperatures. Symbiotic coral Sr/Ca displays four large-amplitude annual cycles with high values in the winter, low values in the summer and a temperature dependence similar to that of tropical reef corals. By comparison, Sr/Ca profiles constructed from asymbiotic coral skeleton display little variability over the same time period. Asymbiont Sr/Ca is relatively insensitive to the enormous temperature changes experienced over the year; the temperature dependence is similar to that of nighttime skeletal deposits in tropical reef corals and non-biological aragonite precipitates. We propose that the large variations in skeletal Sr/Ca observed in all symbiont-hosting coral species are not related to SST variability per se but are driven primarily by large seasonal variations in skeletal calcification rate associated with symbiont photosynthesis. Our model provides a framework for understanding the role of biology in determining coral skeletal chemistry and an explanation for anomalous Sr/Ca-based paleotemperature derivations.

Trans-species learning of cellular signaling systems with bimodal deep belief networks.

PubMed

Chen, Lujia; Cai, Chunhui; Chen, Vicky; Lu, Xinghua

2015-09-15

Model organisms play critical roles in biomedical research of human diseases and drug development. An imperative task is to translate information/knowledge acquired from model organisms to humans. In this study, we address a trans-species learning problem: predicting human cell responses to diverse stimuli, based on the responses of rat cells treated with the same stimuli. We hypothesized that rat and human cells share a common signal-encoding mechanism but employ different proteins to transmit signals, and we developed a bimodal deep belief network and a semi-restricted bimodal deep belief network to represent the common encoding mechanism and perform trans-species learning. These 'deep learning' models include hierarchically organized latent variables capable of capturing the statistical structures in the observed proteomic data in a distributed fashion. The results show that the models significantly outperform two current state-of-the-art classification algorithms. Our study demonstrated the potential of using deep hierarchical models to simulate cellular signaling systems. The software is available at the following URL: http://pubreview.dbmi.pitt.edu/TransSpeciesDeepLearning/. The data are available through SBV IMPROVER website, https://www.sbvimprover.com/challenge-2/overview, upon publication of the report by the organizers. xinghua@pitt.edu Supplementary data are available at Bioinformatics online. © The Author 2015. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com.

A novel deep learning approach for classification of EEG motor imagery signals.

PubMed

Tabar, Yousef Rezaei; Halici, Ugur

2017-02-01

Signal classification is an important issue in brain computer interface (BCI) systems. Deep learning approaches have been used successfully in many recent studies to learn features and classify different types of data. However, the number of studies that employ these approaches on BCI applications is very limited. In this study we aim to use deep learning methods to improve classification performance of EEG motor imagery signals. In this study we investigate convolutional neural networks (CNN) and stacked autoencoders (SAE) to classify EEG Motor Imagery signals. A new form of input is introduced to combine time, frequency and location information extracted from EEG signal and it is used in CNN having one 1D convolutional and one max-pooling layers. We also proposed a new deep network by combining CNN and SAE. In this network, the features that are extracted in CNN are classified through the deep network SAE. The classification performance obtained by the proposed method on BCI competition IV dataset 2b in terms of kappa value is 0.547. Our approach yields 9% improvement over the winner algorithm of the competition. Our results show that deep learning methods provide better classification performance compared to other state of art approaches. These methods can be applied successfully to BCI systems where the amount of data is large due to daily recording.

Breaking the Silence (Teaching and Learning about Cultural Diversity).

ERIC Educational Resources Information Center

Miller, Howard M., Ed.

1997-01-01

Discusses the policy of silence (and its complex reasons) that often rules when it comes to teaching and learning about race, religion, ethnicity, and sexuality. Discusses briefly five books and articles that deal with breaking this silence, and offers observations about effective multiculturalism in the classroom. (SR)

Feedback and Sentence Learning.

ERIC Educational Resources Information Center

Guthrie, John T.

The theoretical functions of external feedback in SR and closed loop models of verbal learning are presented. Contradictory predictions from the models are tested with a three by three factorial experiment including three types of feedback and three amounts of rehearsal. There were 90 adult students run individually and they were required to learn…

Teaching Deanna to Read: A Case Study.

ERIC Educational Resources Information Center

Tiwald, Jeanette M.

1995-01-01

Describes a Reading Recovery case study involving a first-grade student who was at risk for learning how to read and write. Notes that this student learned to read strategically and was accelerated to the average band in her classroom after 81 Reading Recovery lessons, without first knowing the alphabet. (SR)

Teaching neuroanatomy using computer-aided learning: What makes for successful outcomes?

PubMed

Svirko, Elena; Mellanby, Jane

2017-11-01

Computer-aided learning (CAL) is an integral part of many medical courses. The neuroscience course at Oxford University for medical students includes CAL course of neuroanatomy. CAL is particularly suited to this since neuroanatomy requires much detailed three-dimensional visualization, which can be presented on screen. The CAL course was evaluated using the concept of approach to learning. The aims of university teaching are congruent with the deep approach-seeking meaning and relating new information to previous knowledge-rather than to the surface approach of concentrating on rote learning of detail. Seven cohorts of medical students (N = 869) filled in approach to learning scale and a questionnaire investigating their engagement with the CAL course. The students' scores on CAL-course-based neuroanatomy assessment and later university examinations were obtained. Although the students reported less use of the deep approach for the neuroanatomy CAL course than for the rest of their neuroanatomy course (mean = 24.99 vs. 31.49, P < 0.001), deep approach for CAL was positively correlated with neuroanatomy assessment performance (r = 0.12, P < 0.001). Time spent on the CAL course, enjoyment of it, the amount of CAL videos watched and quizzes completed were each significantly positively related to deep approach. The relationship between deep approach and enjoyment was particularly notable (25.5% shared variance). Reported relationships between deep approach and academic performance support the desirability of deep approach in university students. It is proposed that enjoyment of the course and the deep approach could be increased by incorporation of more clinical material which is what the students liked most. Anat Sci Educ 10: 560-569. © 2017 American Association of Anatomists. © 2017 American Association of Anatomists.

Exploring the relationships between epistemic beliefs about medicine and approaches to learning medicine: a structural equation modeling analysis.

PubMed

Chiu, Yen-Lin; Liang, Jyh-Chong; Hou, Cheng-Yen; Tsai, Chin-Chung

2016-07-18

Students' epistemic beliefs may vary in different domains; therefore, it may be beneficial for medical educators to better understand medical students' epistemic beliefs regarding medicine. Understanding how medical students are aware of medical knowledge and how they learn medicine is a critical issue of medical education. The main purposes of this study were to investigate medical students' epistemic beliefs relating to medical knowledge, and to examine their relationships with students' approaches to learning medicine. A total of 340 undergraduate medical students from 9 medical colleges in Taiwan were surveyed with the Medical-Specific Epistemic Beliefs (MSEB) questionnaire (i.e., multi-source, uncertainty, development, justification) and the Approach to Learning Medicine (ALM) questionnaire (i.e., surface motive, surface strategy, deep motive, and deep strategy). By employing the structural equation modeling technique, the confirmatory factor analysis and path analysis were conducted to validate the questionnaires and explore the structural relations between these two constructs. It was indicated that medical students with multi-source beliefs who were suspicious of medical knowledge transmitted from authorities were less likely to possess a surface motive and deep strategies. Students with beliefs regarding uncertain medical knowledge tended to utilize flexible approaches, that is, they were inclined to possess a surface motive but adopt deep strategies. Students with beliefs relating to justifying medical knowledge were more likely to have mixed motives (both surface and deep motives) and mixed strategies (both surface and deep strategies). However, epistemic beliefs regarding development did not have significant relations with approaches to learning. Unexpectedly, it was found that medical students with sophisticated epistemic beliefs (e.g., suspecting knowledge from medical experts) did not necessarily engage in deep approaches to learning medicine. Instead of a deep approach, medical students with sophisticated epistemic beliefs in uncertain and justifying medical knowledge intended to employ a flexible approach and a mixed approach, respectively.

Using Computer Technology to Foster Learning for Understanding

PubMed Central

VAN MELLE, ELAINE; TOMALTY, LEWIS

2000-01-01

The literature shows that students typically use either a surface approach to learning, in which the emphasis is on memorization of facts, or a deep approach to learning, in which learning for understanding is the primary focus. This paper describes how computer technology, specifically the use of a multimedia CD-ROM, was integrated into a microbiology curriculum as part of the transition from focusing on facts to fostering learning for understanding. Evaluation of the changes in approaches to learning over the course of the term showed a statistically significant shift in a deep approach to learning, as measured by the Study Process Questionnaire. Additional data collected showed that the use of computer technology supported this shift by providing students with the opportunity to apply what they had learned in class to order tests and interpret the test results in relation to specific patient-focused case studies. The extent of the impact, however, varied among different groups of students in the class. For example, students who were recent high school graduates did not show a statistically significant increase in deep learning scores over the course of the term and did not perform as well in the course. The results also showed that a surface approach to learning was an important aspect of learning for understanding, although only those students who were able to combine a surface with a deep approach to learning were successfully able to learn for understanding. Implications of this finding for the future use of computer technology and learning for understanding are considered. PMID:23653533

Efficacy of a Deep Learning System for Detecting Glaucomatous Optic Neuropathy Based on Color Fundus Photographs.

PubMed

Li, Zhixi; He, Yifan; Keel, Stuart; Meng, Wei; Chang, Robert T; He, Mingguang

2018-03-02

To assess the performance of a deep learning algorithm for detecting referable glaucomatous optic neuropathy (GON) based on color fundus photographs. A deep learning system for the classification of GON was developed for automated classification of GON on color fundus photographs. We retrospectively included 48 116 fundus photographs for the development and validation of a deep learning algorithm. This study recruited 21 trained ophthalmologists to classify the photographs. Referable GON was defined as vertical cup-to-disc ratio of 0.7 or more and other typical changes of GON. The reference standard was made until 3 graders achieved agreement. A separate validation dataset of 8000 fully gradable fundus photographs was used to assess the performance of this algorithm. The area under receiver operator characteristic curve (AUC) with sensitivity and specificity was applied to evaluate the efficacy of the deep learning algorithm detecting referable GON. In the validation dataset, this deep learning system achieved an AUC of 0.986 with sensitivity of 95.6% and specificity of 92.0%. The most common reasons for false-negative grading (n = 87) were GON with coexisting eye conditions (n = 44 [50.6%]), including pathologic or high myopia (n = 37 [42.6%]), diabetic retinopathy (n = 4 [4.6%]), and age-related macular degeneration (n = 3 [3.4%]). The leading reason for false-positive results (n = 480) was having other eye conditions (n = 458 [95.4%]), mainly including physiologic cupping (n = 267 [55.6%]). Misclassification as false-positive results amidst a normal-appearing fundus occurred in only 22 eyes (4.6%). A deep learning system can detect referable GON with high sensitivity and specificity. Coexistence of high or pathologic myopia is the most common cause resulting in false-negative results. Physiologic cupping and pathologic myopia were the most common reasons for false-positive results. Copyright © 2018 American Academy of Ophthalmology. Published by Elsevier Inc. All rights reserved.

A deep learning approach for pose estimation from volumetric OCT data.

PubMed

Gessert, Nils; Schlüter, Matthias; Schlaefer, Alexander

2018-05-01

Tracking the pose of instruments is a central problem in image-guided surgery. For microscopic scenarios, optical coherence tomography (OCT) is increasingly used as an imaging modality. OCT is suitable for accurate pose estimation due to its micrometer range resolution and volumetric field of view. However, OCT image processing is challenging due to speckle noise and reflection artifacts in addition to the images' 3D nature. We address pose estimation from OCT volume data with a new deep learning-based tracking framework. For this purpose, we design a new 3D convolutional neural network (CNN) architecture to directly predict the 6D pose of a small marker geometry from OCT volumes. We use a hexapod robot to automatically acquire labeled data points which we use to train 3D CNN architectures for multi-output regression. We use this setup to provide an in-depth analysis on deep learning-based pose estimation from volumes. Specifically, we demonstrate that exploiting volume information for pose estimation yields higher accuracy than relying on 2D representations with depth information. Supporting this observation, we provide quantitative and qualitative results that 3D CNNs effectively exploit the depth structure of marker objects. Regarding the deep learning aspect, we present efficient design principles for 3D CNNs, making use of insights from the 2D deep learning community. In particular, we present Inception3D as a new architecture which performs best for our application. We show that our deep learning approach reaches errors at our ground-truth label's resolution. We achieve a mean average error of 14.89 ± 9.3 µm and 0.096 ± 0.072° for position and orientation learning, respectively. Copyright © 2018 Elsevier B.V. All rights reserved.

Deep Learning Method for Denial of Service Attack Detection Based on Restricted Boltzmann Machine.

PubMed

Imamverdiyev, Yadigar; Abdullayeva, Fargana

2018-06-01

In this article, the application of the deep learning method based on Gaussian-Bernoulli type restricted Boltzmann machine (RBM) to the detection of denial of service (DoS) attacks is considered. To increase the DoS attack detection accuracy, seven additional layers are added between the visible and the hidden layers of the RBM. Accurate results in DoS attack detection are obtained by optimization of the hyperparameters of the proposed deep RBM model. The form of the RBM that allows application of the continuous data is used. In this type of RBM, the probability distribution of the visible layer is replaced by a Gaussian distribution. Comparative analysis of the accuracy of the proposed method with Bernoulli-Bernoulli RBM, Gaussian-Bernoulli RBM, deep belief network type deep learning methods on DoS attack detection is provided. Detection accuracy of the methods is verified on the NSL-KDD data set. Higher accuracy from the proposed multilayer deep Gaussian-Bernoulli type RBM is obtained.

Deep learning architecture for iris recognition based on optimal Gabor filters and deep belief network

NASA Astrophysics Data System (ADS)

He, Fei; Han, Ye; Wang, Han; Ji, Jinchao; Liu, Yuanning; Ma, Zhiqiang

2017-03-01

Gabor filters are widely utilized to detect iris texture information in several state-of-the-art iris recognition systems. However, the proper Gabor kernels and the generative pattern of iris Gabor features need to be predetermined in application. The traditional empirical Gabor filters and shallow iris encoding ways are incapable of dealing with such complex variations in iris imaging including illumination, aging, deformation, and device variations. Thereby, an adaptive Gabor filter selection strategy and deep learning architecture are presented. We first employ particle swarm optimization approach and its binary version to define a set of data-driven Gabor kernels for fitting the most informative filtering bands, and then capture complex pattern from the optimal Gabor filtered coefficients by a trained deep belief network. A succession of comparative experiments validate that our optimal Gabor filters may produce more distinctive Gabor coefficients and our iris deep representations be more robust and stable than traditional iris Gabor codes. Furthermore, the depth and scales of the deep learning architecture are also discussed.

Deep learning in mammography and breast histology, an overview and future trends.

PubMed

Hamidinekoo, Azam; Denton, Erika; Rampun, Andrik; Honnor, Kate; Zwiggelaar, Reyer

2018-07-01

Recent improvements in biomedical image analysis using deep learning based neural networks could be exploited to enhance the performance of Computer Aided Diagnosis (CAD) systems. Considering the importance of breast cancer worldwide and the promising results reported by deep learning based methods in breast imaging, an overview of the recent state-of-the-art deep learning based CAD systems developed for mammography and breast histopathology images is presented. In this study, the relationship between mammography and histopathology phenotypes is described, which takes biological aspects into account. We propose a computer based breast cancer modelling approach: the Mammography-Histology-Phenotype-Linking-Model, which develops a mapping of features/phenotypes between mammographic abnormalities and their histopathological representation. Challenges are discussed along with the potential contribution of such a system to clinical decision making and treatment management. Crown Copyright © 2018. Published by Elsevier B.V. All rights reserved.

Comparison of hand-craft feature based SVM and CNN based deep learning framework for automatic polyp classification.

PubMed

Younghak Shin; Balasingham, Ilangko

2017-07-01

Colonoscopy is a standard method for screening polyps by highly trained physicians. Miss-detected polyps in colonoscopy are potential risk factor for colorectal cancer. In this study, we investigate an automatic polyp classification framework. We aim to compare two different approaches named hand-craft feature method and convolutional neural network (CNN) based deep learning method. Combined shape and color features are used for hand craft feature extraction and support vector machine (SVM) method is adopted for classification. For CNN approach, three convolution and pooling based deep learning framework is used for classification purpose. The proposed framework is evaluated using three public polyp databases. From the experimental results, we have shown that the CNN based deep learning framework shows better classification performance than the hand-craft feature based methods. It achieves over 90% of classification accuracy, sensitivity, specificity and precision.

Interpretable Deep Models for ICU Outcome Prediction

PubMed Central

Che, Zhengping; Purushotham, Sanjay; Khemani, Robinder; Liu, Yan

2016-01-01

Exponential surge in health care data, such as longitudinal data from electronic health records (EHR), sensor data from intensive care unit (ICU), etc., is providing new opportunities to discover meaningful data-driven characteristics and patterns ofdiseases. Recently, deep learning models have been employedfor many computational phenotyping and healthcare prediction tasks to achieve state-of-the-art performance. However, deep models lack interpretability which is crucial for wide adoption in medical research and clinical decision-making. In this paper, we introduce a simple yet powerful knowledge-distillation approach called interpretable mimic learning, which uses gradient boosting trees to learn interpretable models and at the same time achieves strong prediction performance as deep learning models. Experiment results on Pediatric ICU dataset for acute lung injury (ALI) show that our proposed method not only outperforms state-of-the-art approaches for morality and ventilator free days prediction tasks but can also provide interpretable models to clinicians. PMID:28269832

Constraining the origin of the Messinian gypsum deposits using coupled measurement of δ^{18}O$/δD in gypsum hydration water and salinity of fluid inclusions

NASA Astrophysics Data System (ADS)

Evans, Nicholas P.; Gázquez, Fernando; McKenzie, Judith A.; Chapman, Hazel J.; Hodell, David A.

2016-04-01

We used oxygen and hydrogen isotopes of gypsum hydration water (GHW) coupled with salinity deduced from ice melting temperatures of primary fluid inclusions in the same samples (in tandem with 87Sr/86Sr, δ34S and other isotopic measurements) to determine the composition of the mother fluids that formed the gypsum deposits of the Messinian Salinity Crisis from shallow and intermediate-depth basins. Using this method, we constrain the origin of the Messinian Primary Lower Gypsum (PLG) of the Sorbas basin (Betic foreland) and both the Upper Gypsum (UG) and the Lower Gypsum of the Sicilian basin. We then compare these results to measurements made on UG recovered from the deep Ionian and Balearic basins drilled during DSDP Leg 42A. The evolution of GHW δ18O/δD vs. salinity is controlled by mixing processes between fresh and seawater, coupled with the degree of evaporation. Evaporation and subsequent precipitation of gypsum from fluids dominated by freshwater will result in a depressed 87Sr/86Sr values and different trajectory in δ18O/δD vs. salinity space compared to fluids dominated by seawater. The slopes of these regression equations help to define the end-members from which the fluid originated. For example, salinity estimates from PLG cycle 6 in the Sorbas basin range from 18 to 51ppt, and after correction for fractionation factors, estimated δ18O and δD values of the mother water are low (-2.6 < δ18O < 2.7‰ ; -16.2 < δD < 15.8‰). The intercepts of the regression equations (i.e. at zero salinity) are within error of the average isotope composition of the modern precipitation and groundwater in this region of SE Spain. This indicates there was a significant contribution of meteoric water during gypsum deposition, while 87Sr/86Sr (0.708942 < 87Sr/86Sr < 0.708971) indicate the ions originated from the dissolution of previously marine evaporites. Gypsum from cycle 2 displays similar mother water values (-2.4 < δ18O < 2.4‰ ; -13.2 < δD < 17.0‰) to cycle 6, but salinities of fluid inclusions are higher averaging ˜100ppt. In contrast to cycle 6, the intercepts of the regression equations of cycle 2 display more positive δ18O/δD values. While the estimated range in δ18O and δD of the mother water and salinities fall below those expected from the evaporation of seawater alone, the slope of the regression equation is similar to that of seawater evaporation. This implies that there is a change up-section from a dominantly marine environment in cycle 2 to a greater influence of meteoric water in cycle 6. The UG from the Sicilian basin display greater δ18O/δD values (2.9 < δ18O < 6.0‰ ; 16.6 < δD < 38.3‰) compared to the PLG of Sorbas, with average salinities of ˜90ppt. The intercept of the regression equations are similar to those of Sorbas cycle 6, indicating the mother fluid was composed of a large percentage of meteoric water that subsequently underwent intense evaporation. This observation concurs with the low values of 87Sr/86Sr from the same UG samples (0.708745 < 87Sr/86Sr < 0.708810) that have been interpreted previously to reflect a substantial dilution of Mediterranean surface water during this period, and with brackish to fresh-water fauna described from the associated marl of the UG in other studies. Ongoing analyses will test if this pattern of intense evaporation of a predominately meteoric mother fluid is reflected in the isotopic composition of the UG deposited in the deep Ionian and Balearic basins.

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.

Can Creative Podcasting Promote Deep Learning? The Use of Podcasting for Learning Content in an Undergraduate Science Unit

ERIC Educational Resources Information Center

Pegrum, Mark; Bartle, Emma; Longnecker, Nancy

2015-01-01

This paper examines the effect of a podcasting task on the examination performance of several hundred first-year chemistry undergraduate students. Educational researchers have established that a deep approach to learning that promotes active understanding of meaning can lead to better student outcomes, higher grades and superior retention of…

Transforming Passive Receptivity of Knowledge into Deep Learning Experiences at the Undergraduate Level: An Example from Music Theory

ERIC Educational Resources Information Center

Ferenc, Anna

2015-01-01

This article discusses transformation of passive knowledge receptivity into experiences of deep learning in a lecture-based music theory course at the second-year undergraduate level through implementation of collaborative projects that evoke natural critical learning environments. It presents an example of such a project, addresses key features…

Using Flipped Classroom Approach to Explore Deep Learning in Large Classrooms

ERIC Educational Resources Information Center

Danker, Brenda

2015-01-01

This project used two Flipped Classroom approaches to stimulate deep learning in large classrooms during the teaching of a film module as part of a Diploma in Performing Arts course at Sunway University, Malaysia. The flipped classes utilized either a blended learning approach where students first watched online lectures as homework, and then…

Aligning Seminars with Bologna Requirements: Reciprocal Peer Tutoring, the Solo Taxonomy and Deep Learning

ERIC Educational Resources Information Center

Lueg, Rainer; Lueg, Klarissa; Lauridsen, Ole

2016-01-01

Changes in public policy, such as the Bologna Process, require students to be equipped with multifunctional competencies to master relevant tasks in unfamiliar situations. Achieving this goal might imply a change in many curricula toward deeper learning. As a didactical means to achieve deep learning results, the authors suggest reciprocal peer…

Student Engagement for Effective Teaching and Deep Learning

ERIC Educational Resources Information Center

Dunleavy, Jodene; Milton, Penny

2008-01-01

Today, all young people need to learn to "use their minds well" through deep engagement in learning that reflects skills, knowledge, and dispositions fit for their present lives as well as the ones they aspire to in the future. More than ever, their health and well being, success in the workplace, ability to construct identities and…

Are Deep Strategic Learners Better Suited to PBL? A Preliminary Study

ERIC Educational Resources Information Center

Papinczak, Tracey

2009-01-01

The aim of this study was to determine if medical students categorised as having deep and strategic approaches to their learning find problem-based learning (PBL) enjoyable and supportive of their learning, and achieve well in the first-year course. Quantitative and qualitative data were gathered from first-year medical students (N = 213). All…

Deep Learning in Distance Education: Are We Achieving the Goal?

ERIC Educational Resources Information Center

Shearer, Rick L.; Gregg, Andrea; Joo, K. P.

2015-01-01

As educators, one of our goals is to help students arrive at deeper levels of learning. However, how is this accomplished, especially in online courses? This design-based research study explored the concept of deep learning through a series of design changes in a graduate education course. A key question that emerged was through what learning…

Pleasure, Learning, Video Games, and Life: The Projective Stance

ERIC Educational Resources Information Center

Gee, James Paul

2005-01-01

This article addresses three questions. First, what is the deep pleasure that humans take from video games? Second, what is the relationship between video games and real life? Third, what do the answers to these questions have to do with learning? Good commercial video games are deep technologies for recruiting learning as a form of profound…

Multi-Site Diagnostic Classification of Schizophrenia Using Discriminant Deep Learning with Functional Connectivity MRI.

PubMed

Zeng, Ling-Li; Wang, Huaning; Hu, Panpan; Yang, Bo; Pu, Weidan; Shen, Hui; Chen, Xingui; Liu, Zhening; Yin, Hong; Tan, Qingrong; Wang, Kai; Hu, Dewen

2018-04-01

A lack of a sufficiently large sample at single sites causes poor generalizability in automatic diagnosis classification of heterogeneous psychiatric disorders such as schizophrenia based on brain imaging scans. Advanced deep learning methods may be capable of learning subtle hidden patterns from high dimensional imaging data, overcome potential site-related variation, and achieve reproducible cross-site classification. However, deep learning-based cross-site transfer classification, despite less imaging site-specificity and more generalizability of diagnostic models, has not been investigated in schizophrenia. A large multi-site functional MRI sample (n = 734, including 357 schizophrenic patients from seven imaging resources) was collected, and a deep discriminant autoencoder network, aimed at learning imaging site-shared functional connectivity features, was developed to discriminate schizophrenic individuals from healthy controls. Accuracies of approximately 85·0% and 81·0% were obtained in multi-site pooling classification and leave-site-out transfer classification, respectively. The learned functional connectivity features revealed dysregulation of the cortical-striatal-cerebellar circuit in schizophrenia, and the most discriminating functional connections were primarily located within and across the default, salience, and control networks. The findings imply that dysfunctional integration of the cortical-striatal-cerebellar circuit across the default, salience, and control networks may play an important role in the "disconnectivity" model underlying the pathophysiology of schizophrenia. The proposed discriminant deep learning method may be capable of learning reliable connectome patterns and help in understanding the pathophysiology and achieving accurate prediction of schizophrenia across multiple independent imaging sites. Copyright © 2018 German Center for Neurodegenerative Diseases (DZNE). Published by Elsevier B.V. All rights reserved.

The rise of deep learning in drug discovery.

PubMed

Chen, Hongming; Engkvist, Ola; Wang, Yinhai; Olivecrona, Marcus; Blaschke, Thomas

2018-06-01

Over the past decade, deep learning has achieved remarkable success in various artificial intelligence research areas. Evolved from the previous research on artificial neural networks, this technology has shown superior performance to other machine learning algorithms in areas such as image and voice recognition, natural language processing, among others. The first wave of applications of deep learning in pharmaceutical research has emerged in recent years, and its utility has gone beyond bioactivity predictions and has shown promise in addressing diverse problems in drug discovery. Examples will be discussed covering bioactivity prediction, de novo molecular design, synthesis prediction and biological image analysis. Copyright © 2018 The Authors. Published by Elsevier Ltd.. All rights reserved.

Learning approaches as predictors of academic performance in first year health and science students.

PubMed

Salamonson, Yenna; Weaver, Roslyn; Chang, Sungwon; Koch, Jane; Bhathal, Ragbir; Khoo, Cheang; Wilson, Ian

2013-07-01

To compare health and science students' demographic characteristics and learning approaches across different disciplines, and to examine the relationship between learning approaches and academic performance. While there is increasing recognition of a need to foster learning approaches that improve the quality of student learning, little is known about students' learning approaches across different disciplines, and their relationships with academic performance. Prospective, correlational design. Using a survey design, a total of 919 first year health and science students studying in a university located in the western region of Sydney from the following disciplines were recruited to participate in the study - i) Nursing: n = 476, ii) Engineering: n = 75, iii) Medicine: n = 77, iv) Health Sciences: n = 204, and v) Medicinal Chemistry: n = 87. Although there was no statistically significant difference in the use of surface learning among the five discipline groups, there were wide variations in the use of deep learning approach. Furthermore, older students and those with English as an additional language were more likely to use deep learning approach. Controlling for hours spent in paid work during term-time and English language usage, both surface learning approach (β = -0.13, p = 0.001) and deep learning approach (β = 0.11, p = 0.009) emerged as independent and significant predictors of academic performance. Findings from this study provide further empirical evidence that underscore the importance for faculty to use teaching methods that foster deep instead of surface learning approaches, to improve the quality of student learning and academic performance. Copyright © 2013 Elsevier Ltd. All rights reserved.

Deep transfer learning for automatic target classification: MWIR to LWIR

NASA Astrophysics Data System (ADS)

Ding, Zhengming; Nasrabadi, Nasser; Fu, Yun

2016-05-01

Publisher's Note: This paper, originally published on 5/12/2016, was replaced with a corrected/revised version on 5/18/2016. If you downloaded the original PDF but are unable to access the revision, please contact SPIE Digital Library Customer Service for assistance. When dealing with sparse or no labeled data in the target domain, transfer learning shows its appealing performance by borrowing the supervised knowledge from external domains. Recently deep structure learning has been exploited in transfer learning due to its attractive power in extracting effective knowledge through multi-layer strategy, so that deep transfer learning is promising to address the cross-domain mismatch. In general, cross-domain disparity can be resulted from the difference between source and target distributions or different modalities, e.g., Midwave IR (MWIR) and Longwave IR (LWIR). In this paper, we propose a Weighted Deep Transfer Learning framework for automatic target classification through a task-driven fashion. Specifically, deep features and classifier parameters are obtained simultaneously for optimal classification performance. In this way, the proposed deep structures can extract more effective features with the guidance of the classifier performance; on the other hand, the classifier performance is further improved since it is optimized on more discriminative features. Furthermore, we build a weighted scheme to couple source and target output by assigning pseudo labels to target data, therefore we can transfer knowledge from source (i.e., MWIR) to target (i.e., LWIR). Experimental results on real databases demonstrate the superiority of the proposed algorithm by comparing with others.

Interaction between different groundwaters in brittany catchments (france): characterizing multiple sources through Sr- and S isotope tracing

NASA Astrophysics Data System (ADS)

Negrel, Ph; Pauwels, H.

2003-04-01

Water resources in hard-rocks commonly involve different hydrogeological compartments such as overlying sediments, weathered rock, the weathered-fissured zone, and fractured bedrock. Streams, lakes and wetlands that drain such environments can drain groundwater, recharge groundwater, or do both. Groundwater resources in many countries are increasingly threatened by growing demand, wasteful use, and contamination. Surface water and shallow groundwater are particularly vulnerable to pollution, while deeper resources are more protected from contamination. Sr- and S-isotope data as well as major ions, from shallow and deep groundwater in three granite and Brioverian "schist" areas of the Armorican Massif (NW France) with intensive agriculture covering large parts are presented. The stable-isotope signatures of the waters plot close to the general meteoric-water line, reflecting a meteoric origin and the lack of significant evaporation or water-rock interaction. The water chemistry from the different catchments shows large variation in the major-element contents. Plotting Na, Mg, NO_3, K, SO_4 and Sr vs. Cl contents concentrations reflect agricultural input from hog and livestock farming and fertilizer applications, with local sewage-effluent influence, although some water samples are clearly unpolluted. The δ34S(SO_4) is controlled by several potential sources (atmospheric sulphate, pyrite-derived sulphates, fertilizer sulphates). Some δ18O and δ34S values are expected to increase through sulphate reduction, with higher effect on δ34S for the dissimilatory processes and on δ18O for assimilatory processes. The range in Sr contents in groundwater from different catchments agrees with previous work on groundwater sampled from granites in France. The Sr content is well correlated with Mg and both are related to agricultural practises. As in granite-gneiss watersheds in France, 87Sr/86Sr ratios range from 0.71265 to 0.72009. The relationship between 87Sr/86Sr and Mg/Sr ratios defines the different end-members (rain, agricultural practise, water-rock interaction) both in the three Brittany catchments and elsewhere in France such as the Margeride mountains (S Massif Central), the Hérault watershed (S France), the Morvan (SE Paris Basin), the Cantal (E Massif Central) and the Vosges massif (NE France). Sr-isotope tracing defines and identifies the relative signature of groundwater circulation in alterite and underlying weathered-fissured and fractured bedrock.

Deep convolutional neural networks for automatic classification of gastric carcinoma using whole slide images in digital histopathology.

PubMed

Sharma, Harshita; Zerbe, Norman; Klempert, Iris; Hellwich, Olaf; Hufnagl, Peter

2017-11-01

Deep learning using convolutional neural networks is an actively emerging field in histological image analysis. This study explores deep learning methods for computer-aided classification in H&E stained histopathological whole slide images of gastric carcinoma. An introductory convolutional neural network architecture is proposed for two computerized applications, namely, cancer classification based on immunohistochemical response and necrosis detection based on the existence of tumor necrosis in the tissue. Classification performance of the developed deep learning approach is quantitatively compared with traditional image analysis methods in digital histopathology requiring prior computation of handcrafted features, such as statistical measures using gray level co-occurrence matrix, Gabor filter-bank responses, LBP histograms, gray histograms, HSV histograms and RGB histograms, followed by random forest machine learning. Additionally, the widely known AlexNet deep convolutional framework is comparatively analyzed for the corresponding classification problems. The proposed convolutional neural network architecture reports favorable results, with an overall classification accuracy of 0.6990 for cancer classification and 0.8144 for necrosis detection. Copyright © 2017 Elsevier Ltd. All rights reserved.

Deep learning for studies of galaxy morphology

NASA Astrophysics Data System (ADS)

Tuccillo, D.; Huertas-Company, M.; Decencière, E.; Velasco-Forero, S.

2017-06-01

Establishing accurate morphological measurements of galaxies in a reasonable amount of time for future big-data surveys such as EUCLID, the Large Synoptic Survey Telescope or the Wide Field Infrared Survey Telescope is a challenge. Because of its high level of abstraction with little human intervention, deep learning appears to be a promising approach. Deep learning is a rapidly growing discipline that models high-level patterns in data as complex multilayered networks. In this work we test the ability of deep convolutional networks to provide parametric properties of Hubble Space Telescope like galaxies (half-light radii, Sérsic indices, total flux etc..). We simulate a set of galaxies including point spread function and realistic noise from the CANDELS survey and try to recover the main galaxy parameters using deep-learning. We compare the results with the ones obtained with the commonly used profile fitting based software GALFIT. This way showing that with our method we obtain results at least equally good as the ones obtained with GALFIT but, once trained, with a factor 5 hundred time faster.

Nonlinear Deep Kernel Learning for Image Annotation.

PubMed

Jiu, Mingyuan; Sahbi, Hichem

2017-02-08

Multiple kernel learning (MKL) is a widely used technique for kernel design. Its principle consists in learning, for a given support vector classifier, the most suitable convex (or sparse) linear combination of standard elementary kernels. However, these combinations are shallow and often powerless to capture the actual similarity between highly semantic data, especially for challenging classification tasks such as image annotation. In this paper, we redefine multiple kernels using deep multi-layer networks. In this new contribution, a deep multiple kernel is recursively defined as a multi-layered combination of nonlinear activation functions, each one involves a combination of several elementary or intermediate kernels, and results into a positive semi-definite deep kernel. We propose four different frameworks in order to learn the weights of these networks: supervised, unsupervised, kernel-based semisupervised and Laplacian-based semi-supervised. When plugged into support vector machines (SVMs), the resulting deep kernel networks show clear gain, compared to several shallow kernels for the task of image annotation. Extensive experiments and analysis on the challenging ImageCLEF photo annotation benchmark, the COREL5k database and the Banana dataset validate the effectiveness of the proposed method.

A visual tracking method based on deep learning without online model updating

NASA Astrophysics Data System (ADS)

Tang, Cong; Wang, Yicheng; Feng, Yunsong; Zheng, Chao; Jin, Wei

2018-02-01

The paper proposes a visual tracking method based on deep learning without online model updating. In consideration of the advantages of deep learning in feature representation, deep model SSD (Single Shot Multibox Detector) is used as the object extractor in the tracking model. Simultaneously, the color histogram feature and HOG (Histogram of Oriented Gradient) feature are combined to select the tracking object. In the process of tracking, multi-scale object searching map is built to improve the detection performance of deep detection model and the tracking efficiency. In the experiment of eight respective tracking video sequences in the baseline dataset, compared with six state-of-the-art methods, the method in the paper has better robustness in the tracking challenging factors, such as deformation, scale variation, rotation variation, illumination variation, and background clutters, moreover, its general performance is better than other six tracking methods.

Automated Whole-Body Bone Lesion Detection for Multiple Myeloma on 68Ga-Pentixafor PET/CT Imaging Using Deep Learning Methods.

PubMed

Xu, Lina; Tetteh, Giles; Lipkova, Jana; Zhao, Yu; Li, Hongwei; Christ, Patrick; Piraud, Marie; Buck, Andreas; Shi, Kuangyu; Menze, Bjoern H

2018-01-01

The identification of bone lesions is crucial in the diagnostic assessment of multiple myeloma (MM). 68 Ga-Pentixafor PET/CT can capture the abnormal molecular expression of CXCR-4 in addition to anatomical changes. However, whole-body detection of dozens of lesions on hybrid imaging is tedious and error prone. It is even more difficult to identify lesions with a large heterogeneity. This study employed deep learning methods to automatically combine characteristics of PET and CT for whole-body MM bone lesion detection in a 3D manner. Two convolutional neural networks (CNNs), V-Net and W-Net, were adopted to segment and detect the lesions. The feasibility of deep learning for lesion detection on 68 Ga-Pentixafor PET/CT was first verified on digital phantoms generated using realistic PET simulation methods. Then the proposed methods were evaluated on real 68 Ga-Pentixafor PET/CT scans of MM patients. The preliminary results showed that deep learning method can leverage multimodal information for spatial feature representation, and W-Net obtained the best result for segmentation and lesion detection. It also outperformed traditional machine learning methods such as random forest classifier (RF), k -Nearest Neighbors ( k -NN), and support vector machine (SVM). The proof-of-concept study encourages further development of deep learning approach for MM lesion detection in population study.

Automated Whole-Body Bone Lesion Detection for Multiple Myeloma on 68Ga-Pentixafor PET/CT Imaging Using Deep Learning Methods

PubMed Central

Tetteh, Giles; Lipkova, Jana; Zhao, Yu; Li, Hongwei; Christ, Patrick; Buck, Andreas; Menze, Bjoern H.

2018-01-01

The identification of bone lesions is crucial in the diagnostic assessment of multiple myeloma (MM). 68Ga-Pentixafor PET/CT can capture the abnormal molecular expression of CXCR-4 in addition to anatomical changes. However, whole-body detection of dozens of lesions on hybrid imaging is tedious and error prone. It is even more difficult to identify lesions with a large heterogeneity. This study employed deep learning methods to automatically combine characteristics of PET and CT for whole-body MM bone lesion detection in a 3D manner. Two convolutional neural networks (CNNs), V-Net and W-Net, were adopted to segment and detect the lesions. The feasibility of deep learning for lesion detection on 68Ga-Pentixafor PET/CT was first verified on digital phantoms generated using realistic PET simulation methods. Then the proposed methods were evaluated on real 68Ga-Pentixafor PET/CT scans of MM patients. The preliminary results showed that deep learning method can leverage multimodal information for spatial feature representation, and W-Net obtained the best result for segmentation and lesion detection. It also outperformed traditional machine learning methods such as random forest classifier (RF), k-Nearest Neighbors (k-NN), and support vector machine (SVM). The proof-of-concept study encourages further development of deep learning approach for MM lesion detection in population study. PMID:29531504

Digging deeper on "deep" learning: A computational ecology approach.

PubMed

Buscema, Massimo; Sacco, Pier Luigi

2017-01-01

We propose an alternative approach to "deep" learning that is based on computational ecologies of structurally diverse artificial neural networks, and on dynamic associative memory responses to stimuli. Rather than focusing on massive computation of many different examples of a single situation, we opt for model-based learning and adaptive flexibility. Cross-fertilization of learning processes across multiple domains is the fundamental feature of human intelligence that must inform "new" artificial intelligence.

ML-o-Scope: A Diagnostic Visualization System for Deep Machine Learning Pipelines

DTIC Science & Technology

2014-05-16

ML-o-scope: a diagnostic visualization system for deep machine learning pipelines Daniel Bruckner Electrical Engineering and Computer Sciences... machine learning pipelines 5a. CONTRACT NUMBER 5b. GRANT NUMBER 5c. PROGRAM ELEMENT NUMBER 6. AUTHOR(S) 5d. PROJECT NUMBER 5e. TASK NUMBER 5f...the system as a support for tuning large scale object-classification pipelines. 1 Introduction A new generation of pipelined machine learning models

Multiagent cooperation and competition with deep reinforcement learning.

PubMed

Tampuu, Ardi; Matiisen, Tambet; Kodelja, Dorian; Kuzovkin, Ilya; Korjus, Kristjan; Aru, Juhan; Aru, Jaan; Vicente, Raul

2017-01-01

Evolution of cooperation and competition can appear when multiple adaptive agents share a biological, social, or technological niche. In the present work we study how cooperation and competition emerge between autonomous agents that learn by reinforcement while using only their raw visual input as the state representation. In particular, we extend the Deep Q-Learning framework to multiagent environments to investigate the interaction between two learning agents in the well-known video game Pong. By manipulating the classical rewarding scheme of Pong we show how competitive and collaborative behaviors emerge. We also describe the progression from competitive to collaborative behavior when the incentive to cooperate is increased. Finally we show how learning by playing against another adaptive agent, instead of against a hard-wired algorithm, results in more robust strategies. The present work shows that Deep Q-Networks can become a useful tool for studying decentralized learning of multiagent systems coping with high-dimensional environments.

Multiagent cooperation and competition with deep reinforcement learning

PubMed Central

Kodelja, Dorian; Kuzovkin, Ilya; Korjus, Kristjan; Aru, Juhan; Aru, Jaan; Vicente, Raul

2017-01-01

Evolution of cooperation and competition can appear when multiple adaptive agents share a biological, social, or technological niche. In the present work we study how cooperation and competition emerge between autonomous agents that learn by reinforcement while using only their raw visual input as the state representation. In particular, we extend the Deep Q-Learning framework to multiagent environments to investigate the interaction between two learning agents in the well-known video game Pong. By manipulating the classical rewarding scheme of Pong we show how competitive and collaborative behaviors emerge. We also describe the progression from competitive to collaborative behavior when the incentive to cooperate is increased. Finally we show how learning by playing against another adaptive agent, instead of against a hard-wired algorithm, results in more robust strategies. The present work shows that Deep Q-Networks can become a useful tool for studying decentralized learning of multiagent systems coping with high-dimensional environments. PMID:28380078

Quantification of flexoelectricity in PbTiO 3/SrTiO 3 superlattice polar vortices using machine learning and phase-field modeling

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

Li, Q.; Nelson, C. T.; Hsu, S. -L.

Flexoelectricity refers to electric polarization generated by heterogeneous mechanical strains, namely strain gradients, in materials of arbitrary crystal symmetries. Despite more than 50 years of work on this effect, an accurate identification of its coupling strength remains an experimental challenge for most materials, which impedes its wide recognition. Here, we show the presence of flexoelectricity in the recently discovered polar vortices in PbTiO 3 /SrTiO 3 superlattices based on a combination of machine-learning analysis of the atomic-scale electron microscopy imaging data and phenomenological phase-field modeling. By scrutinizing the influence of flexocoupling on the global vortex structure, we match theory andmore » experiment using computer vision methodologies to determine the flexoelectric coefficients for PbTiO 3 and SrTiO 3. Here, our findings highlight the inherent, nontrivial role of flexoelectricity in the generation of emergent complex polarization morphologies and demonstrate a viable approach to delineating this effect, conducive to the deeper exploration of both topics.« less

Quantification of flexoelectricity in PbTiO 3/SrTiO 3 superlattice polar vortices using machine learning and phase-field modeling

DOE PAGES

Li, Q.; Nelson, C. T.; Hsu, S. -L.; ...

2017-11-13

Flexoelectricity refers to electric polarization generated by heterogeneous mechanical strains, namely strain gradients, in materials of arbitrary crystal symmetries. Despite more than 50 years of work on this effect, an accurate identification of its coupling strength remains an experimental challenge for most materials, which impedes its wide recognition. Here, we show the presence of flexoelectricity in the recently discovered polar vortices in PbTiO 3 /SrTiO 3 superlattices based on a combination of machine-learning analysis of the atomic-scale electron microscopy imaging data and phenomenological phase-field modeling. By scrutinizing the influence of flexocoupling on the global vortex structure, we match theory andmore » experiment using computer vision methodologies to determine the flexoelectric coefficients for PbTiO 3 and SrTiO 3. Here, our findings highlight the inherent, nontrivial role of flexoelectricity in the generation of emergent complex polarization morphologies and demonstrate a viable approach to delineating this effect, conducive to the deeper exploration of both topics.« less

Deglacial Dust Provenance Changes in the Eastern Equatorial Pacific and Implications for ITCZ Movement

NASA Astrophysics Data System (ADS)

Xie, R.; Marcantonio, F.

2009-12-01

The provenance of the eolian dust component of deep-sea sediments has the potential to offer insights into changes in past atmospheric circulation patterns. Measuring temporal changes in dust provenance can shed more light on changes in the position of the Intertropical Convergence Zone (ITCZ), a region in which dust is removed from the upper troposphere by deep convection and scavenged by precipitation. The ITCZ should, therefore, act as a barrier separating wind-blown northern versus southern sources. Perhaps the best way to trace for provenance of dust sources is through the measurement of radiogenic isotope ratios. Here, we have analyzed Nd, Sr, and Pb isotope ratios in the detrital components extracted from deep-sea sediments in the eastern equatorial Pacific (EEP) along a meridional transect at 110W from 3S to 7N (ODP Leg 138, Sites 848 - 853). At each site, the sediments ranged in age from 0 to 25 ka. Detrital component extraction involved the chemical removal of the biogenic and authigenic sedimentary fractions. Preliminary detrital Nd isotope ratios show a range of 2.4 ɛNd units (from -5.7 to -3.3). There are distinct latitudinal trends in the ɛNd values, with more radiogenic values further south and less radiogenic values further north. This distinction holds true for both Holocene and glacial time. The difference in Nd isotope ratios at any one site between Holocene and glacial is smaller for the sites furthest North. The greatest Holocene-glacial differences in ɛNd occur at sites south of 3N, suggesting a distinct detrital component boundary at this latitude. The sites furthest north (7N and 5.29N) show the greatest variability in detrital 87Sr/86Sr isotope ratios, while sites furthest south (equator and 1.5N) show negligible variability. The detrital component of sediment at Site 851 (2.77N) has a Sr isotope variability that is intermediate between the northern and southern values, again suggesting a detrital boundary of some sort. We interpreted these preliminary results to suggest that the ITCZ position was displaced toward the south, with a paleo-position located between Site 851 (2.77N) and Site 852 (5.29N), from its average current location, which is approximately 7N. This supports one hypothesis put forward by McGee et al. [1], based on changes in dust flux for the same transect, that the ITCZ was displaced toward the south during the last glacial period. [1] McGee et al., 2007, EPSL 257, 215-230 .

Characterization of Carbonate Crust from Deep-sea Methane Seeps on the Northern US Atlantic Margin.

NASA Astrophysics Data System (ADS)

Gabitov, R. I.; Borrelli, C.; Buettner, J.; Testa, M.; Garner, B.; Weremeichik, J.; Thomas, J. B.; Wahidi, M.; Thirumalai, R. V. K. G.; Kirkland, B. L.; Skarke, A. D.

2017-12-01

Authigenic carbonate minerals widely occur at the seafloor as carbonate crusts and are often directly linked to microbial activity, about which promotion of carbonate crystal growth and geochemistry are not entirely understood. To evaluate a potential metabolic contribution, studies were conducted on carbonate crust collected from a methane seep and on precipitation experiments which produced inorganic aragonite crystallized at high pressure. Among the samples collected during a NSF sponsored cruise to the North Atlantic Continental Margin of the United States (off of New England) in July-August 2016, we analyzed one carbonate crust sample (AD4835 BB-4522) collected at 39.805860; -69.592593 and at a depth of 1419.6 m. In this crust sample, two textural types of aragonite were identified: 1) groundmass consisting of fine grey crystals (<1 µm in size); 2) veins consisting of white acicular crystals (up to 100 µm in width). In addition, large equant quartz crystals (>100 µm, 24.9 wt%), feldspar (5.6 wt%), and dolomite (3.6 wt%), and trace amount of troilite were identified using XRD, SEM, and optical microscopy. The sample was cut into slabs parallel to crust growth assuming the crust grew in a downward direction. Concentrations of Na, Mg, Al, Si, S, K, Ca, Mn, Fe, Sr, Zr, Ba, and U were measured in the direction parallel to growth of the crust using LA-ICP-MS. Proportions of Si, Al, (Na+K), Mg, S, and Fe in the groundmass suggest the occurrence of sub-micron inclusions of alkali feldspar, and potentially pyroxene, Fe oxide, and Fe sulfide, which were impossible to avoid with the instrument's spatial resolution. The occurrence of micro non-carbonate inclusions causes high elemental concentrations compared to the values expected for aragonite crystallized from seawater. White aragonite acicular crystals were free of silicate and sulfide inclusions, and therefore, yielded lower concentrations of all measured elements except Sr compared to the groundmass. Analyzed Mg and Sr are consistent with published data for deep-sea corals. Also, Sr is similar to experimental data on inorganic aragonite. Mg/Ca, Sr/Ca, Ba/Ca, and U/Ca of the fluid from which acicular aragonite grew were calculated based on partition coefficients from inorganic aragonite precipitated at 100 bars.

The Next Era: Deep Learning in Pharmaceutical Research

PubMed Central

Ekins, Sean

2016-01-01

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. PMID:27599991

Hyperspectral imagery super-resolution by compressive sensing inspired dictionary learning and spatial-spectral regularization.

PubMed

Huang, Wei; Xiao, Liang; Liu, Hongyi; Wei, Zhihui

2015-01-19

Due to the instrumental and imaging optics limitations, it is difficult to acquire high spatial resolution hyperspectral imagery (HSI). Super-resolution (SR) imagery aims at inferring high quality images of a given scene from degraded versions of the same scene. This paper proposes a novel hyperspectral imagery super-resolution (HSI-SR) method via dictionary learning and spatial-spectral regularization. The main contributions of this paper are twofold. First, inspired by the compressive sensing (CS) framework, for learning the high resolution dictionary, we encourage stronger sparsity on image patches and promote smaller coherence between the learned dictionary and sensing matrix. Thus, a sparsity and incoherence restricted dictionary learning method is proposed to achieve higher efficiency sparse representation. Second, a variational regularization model combing a spatial sparsity regularization term and a new local spectral similarity preserving term is proposed to integrate the spectral and spatial-contextual information of the HSI. Experimental results show that the proposed method can effectively recover spatial information and better preserve spectral information. The high spatial resolution HSI reconstructed by the proposed method outperforms reconstructed results by other well-known methods in terms of both objective measurements and visual evaluation.

DeepSig: deep learning improves signal peptide detection in proteins.

PubMed

Savojardo, Castrense; Martelli, Pier Luigi; Fariselli, Piero; Casadio, Rita

2018-05-15

The identification of signal peptides in protein sequences is an important step toward protein localization and function characterization. Here, we present DeepSig, an improved approach for signal peptide detection and cleavage-site prediction based on deep learning methods. Comparative benchmarks performed on an updated independent dataset of proteins show that DeepSig is the current best performing method, scoring better than other available state-of-the-art approaches on both signal peptide detection and precise cleavage-site identification. DeepSig is available as both standalone program and web server at https://deepsig.biocomp.unibo.it. All datasets used in this study can be obtained from the same website. pierluigi.martelli@unibo.it. Supplementary data are available at Bioinformatics online.

Deep learning on temporal-spectral data for anomaly detection

NASA Astrophysics Data System (ADS)

Ma, King; Leung, Henry; Jalilian, Ehsan; Huang, Daniel

2017-05-01

Detecting anomalies is important for continuous monitoring of sensor systems. One significant challenge is to use sensor data and autonomously detect changes that cause different conditions to occur. Using deep learning methods, we are able to monitor and detect changes as a result of some disturbance in the system. We utilize deep neural networks for sequence analysis of time series. We use a multi-step method for anomaly detection. We train the network to learn spectral and temporal features from the acoustic time series. We test our method using fiber-optic acoustic data from a pipeline.

Clinical evaluation of atlas and deep learning based automatic contouring for lung cancer.

PubMed

Lustberg, Tim; van Soest, Johan; Gooding, Mark; Peressutti, Devis; Aljabar, Paul; van der Stoep, Judith; van Elmpt, Wouter; Dekker, Andre

2018-02-01

Contouring of organs at risk (OARs) is an important but time consuming part of radiotherapy treatment planning. The aim of this study was to investigate whether using institutional created software-generated contouring will save time if used as a starting point for manual OAR contouring for lung cancer patients. Twenty CT scans of stage I-III NSCLC patients were used to compare user adjusted contours after an atlas-based and deep learning contour, against manual delineation. The lungs, esophagus, spinal cord, heart and mediastinum were contoured for this study. The time to perform the manual tasks was recorded. With a median time of 20 min for manual contouring, the total median time saved was 7.8 min when using atlas-based contouring and 10 min for deep learning contouring. Both atlas based and deep learning adjustment times were significantly lower than manual contouring time for all OARs except for the left lung and esophagus of the atlas based contouring. User adjustment of software generated contours is a viable strategy to reduce contouring time of OARs for lung radiotherapy while conforming to local clinical standards. In addition, deep learning contouring shows promising results compared to existing solutions. Copyright © 2017 The Authors. Published by Elsevier B.V. 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.

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.

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

On the Role of Discipline-Related Self-Concept in Deep and Surface Approaches to Learning among University Students

ERIC Educational Resources Information Center

Platow, Michael J.; Mavor, Kenneth I.; Grace, Diana M.

2013-01-01

The current research examined the role that students' discipline-related self-concepts may play in their deep and surface approaches to learning, their overall learning outcomes, and continued engagement in the discipline itself. Using a cross-lagged panel design of first-year university psychology students, a causal path was observed in which…

Evaluating Primary School Student's Deep Learning Approach to Science Lessons

ERIC Educational Resources Information Center

Ilkörücü Göçmençelebi, Sirin; Özkan, Muhlis; Bayram, Nuran

2012-01-01

This study examines the variables which help direct students to a deep learning approach to science lessons, with the aim of guiding programmers and teachers in primary education. The sample was composed of a total of 164 primary school students. The Learning Approaches to Science Scale developed by Ünal (2005) for Science and Technology lessons…

Deep Knowledge: Learning to Teach Science for Understanding and Equity. Teaching for Social Justice

ERIC Educational Resources Information Center

Larkin, Douglas B.

2013-01-01

"Deep Knowledge" is a book about how people's ideas change as they learn to teach. Using the experiences of six middle and high school student teachers as they learn to teach science in diverse classrooms, Larkin explores how their work changes the way they think about students, society, schools, and science itself. Through engaging case stories,…

The Effect of Peer Feedback for Blogging on College Students' Reflective Learning Processes

ERIC Educational Resources Information Center

Xie, Ying; Ke, Fengfeng; Sharma, Priya

2008-01-01

Reflection is an important prerequisite to making meaning of new information, and to advance from surface to deep learning. Strategies such as journal writing and peer feedback have been found to promote reflection as well as deep thinking and learning. This study used an empirical design to investigate the interaction effects of peer feedback and…

Examining Learning Approaches of Science Student Teachers According to the Class Level and Gender

ERIC Educational Resources Information Center

Tural Dincer, Guner; Akdeniz, Ali Riza

2008-01-01

There are many factors influence the level of students' achievement in education. Studies show that one of these factors is "learning approach of a student". Research findings generally have identified two approaches of learning: deep and surface. When a student uses the deep approach, he/she has an intrinsic interest in subject matter and is…

Using an In-Class Simulation in the First Accounting Class: Moving from Surface to Deep Learning

ERIC Educational Resources Information Center

Phillips, Mary E.; Graeff, Timothy R.

2014-01-01

As students often find the first accounting class to be abstract and difficult to understand, the authors designed an in-class simulation as an intervention to move students toward deep learning and away from surface learning. The simulation consists of buying and selling merchandise and accounting for transactions. The simulation is an effective…

Self-Regulated Learning and Perceived Health among University Students Participating in Physical Activity Classes

ERIC Educational Resources Information Center

McBride, Ron E.; Altunsöz, Irmak Hürmeriç; Su, Xiaoxia; Xiang, Ping; Demirhan, Giyasettin

2016-01-01

The purpose of this study was to explore motivational indicators of self-regulated learning (SRL) and the relationship between self-regulation (SR) and perceived health among university students enrolled in physical activity (PA) classes. One hundred thirty-one Turkish students participating in physical education activity classes at two…

A Read-Aloud for Foreign Languages: Becoming a Language Master (Read It Aloud).

ERIC Educational Resources Information Center

Richardson, Judy S.

1998-01-01

Describes the use of some read-alouds from Alexandre Dumas'"The Count of Monte Cristo" which helped to demonstrate some principles of learning foreign languages. Describes briefly the read aloud selection, discusses some specific activities that relate to foreign language learning, and discusses specific language arts activities. (SR)

What Whole Language in the Mainstream Means for Children with Learning Disabilities.

ERIC Educational Resources Information Center

Scala, Marilyn A.

1993-01-01

Describes how a teacher of children with learning disabilities worked with three regular classroom teachers to teach mainstreamed children in whole-language classrooms. Shows how students' reading abilities, self-esteem, and motivation improved as the lines were blurred between abled and disabled, teacher and specialist, and right and wrong. (SR)

Capital Equipment: How to Get It and Manage It.

ERIC Educational Resources Information Center

Caernarven-Smith, Patricia

1994-01-01

Describes how technical communicators can learn to write capital equipment requests that specify the needed new equipment, show the benefits to the company from buying it, and recommend the way to account for it. Shows that to do so, they must learn some business basics and make friends outside their own departments. (SR)

Learning representations for the early detection of sepsis with deep neural networks.

PubMed

Kam, Hye Jin; Kim, Ha Young

2017-10-01

Sepsis is one of the leading causes of death in intensive care unit patients. Early detection of sepsis is vital because mortality increases as the sepsis stage worsens. This study aimed to develop detection models for the early stage of sepsis using deep learning methodologies, and to compare the feasibility and performance of the new deep learning methodology with those of the regression method with conventional temporal feature extraction. Study group selection adhered to the InSight model. The results of the deep learning-based models and the InSight model were compared. With deep feedforward networks, the area under the ROC curve (AUC) of the models were 0.887 and 0.915 for the InSight and the new feature sets, respectively. For the model with the combined feature set, the AUC was the same as that of the basic feature set (0.915). For the long short-term memory model, only the basic feature set was applied and the AUC improved to 0.929 compared with the existing 0.887 of the InSight model. The contributions of this paper can be summarized in three ways: (i) improved performance without feature extraction using domain knowledge, (ii) verification of feature extraction capability of deep neural networks through comparison with reference features, and (iii) improved performance with feedforward neural networks using long short-term memory, a neural network architecture that can learn sequential patterns. Copyright © 2017 Elsevier Ltd. All rights reserved.

Surface and deep structures in graphics comprehension.

PubMed

Schnotz, Wolfgang; Baadte, Christiane

2015-05-01

Comprehension of graphics can be considered as a process of schema-mediated structure mapping from external graphics on internal mental models. Two experiments were conducted to test the hypothesis that graphics possess a perceptible surface structure as well as a semantic deep structure both of which affect mental model construction. The same content was presented to different groups of learners by graphics from different perspectives with different surface structures but the same deep structure. Deep structures were complementary: major features of the learning content in one experiment became minor features in the other experiment, and vice versa. Text was held constant. Participants were asked to read, understand, and memorize the learning material. Furthermore, they were either instructed to process the material from the perspective supported by the graphic or from an alternative perspective, or they received no further instruction. After learning, they were asked to recall the learning content from different perspectives by completing graphs of different formats as accurately as possible. Learners' recall was more accurate if the format of recall was the same as the learning format which indicates surface structure influences. However, participants also showed more accurate recall when they remembered the content from a perspective emphasizing the deep structure, regardless of the graphics format presented before. This included better recall of what they had not seen than of what they really had seen before. That is, deep structure effects overrode surface effects. Depending on context conditions, stimulation of additional cognitive processing by instruction had partially positive and partially negative effects.

Deep Learning for Flow Sculpting: Insights into Efficient Learning using Scientific Simulation Data

PubMed Central

Stoecklein, Daniel; Lore, Kin Gwn; Davies, Michael; Sarkar, Soumik; Ganapathysubramanian, Baskar

2017-01-01

A new technique for shaping microfluid flow, known as flow sculpting, offers an unprecedented level of passive fluid flow control, with potential breakthrough applications in advancing manufacturing, biology, and chemistry research at the microscale. However, efficiently solving the inverse problem of designing a flow sculpting device for a desired fluid flow shape remains a challenge. Current approaches struggle with the many-to-one design space, requiring substantial user interaction and the necessity of building intuition, all of which are time and resource intensive. Deep learning has emerged as an efficient function approximation technique for high-dimensional spaces, and presents a fast solution to the inverse problem, yet the science of its implementation in similarly defined problems remains largely unexplored. We propose that deep learning methods can completely outpace current approaches for scientific inverse problems while delivering comparable designs. To this end, we show how intelligent sampling of the design space inputs can make deep learning methods more competitive in accuracy, while illustrating their generalization capability to out-of-sample predictions. PMID:28402332

Holography as deep learning

NASA Astrophysics Data System (ADS)

Gan, Wen-Cong; Shu, Fu-Wen

Quantum many-body problem with exponentially large degrees of freedom can be reduced to a tractable computational form by neural network method [G. Carleo and M. Troyer, Science 355 (2017) 602, arXiv:1606.02318.] The power of deep neural network (DNN) based on deep learning is clarified by mapping it to renormalization group (RG), which may shed lights on holographic principle by identifying a sequence of RG transformations to the AdS geometry. In this paper, we show that any network which reflects RG process has intrinsic hyperbolic geometry, and discuss the structure of entanglement encoded in the graph of DNN. We find the entanglement structure of DNN is of Ryu-Takayanagi form. Based on these facts, we argue that the emergence of holographic gravitational theory is related to deep learning process of the quantum-field theory.

Multi-categorical deep learning neural network to classify retinal images: A pilot study employing small database.

PubMed

Choi, Joon Yul; Yoo, Tae Keun; Seo, Jeong Gi; Kwak, Jiyong; Um, Terry Taewoong; Rim, Tyler Hyungtaek

2017-01-01

Deep learning emerges as a powerful tool for analyzing medical images. Retinal disease detection by using computer-aided diagnosis from fundus image has emerged as a new method. We applied deep learning convolutional neural network by using MatConvNet for an automated detection of multiple retinal diseases with fundus photographs involved in STructured Analysis of the REtina (STARE) database. Dataset was built by expanding data on 10 categories, including normal retina and nine retinal diseases. The optimal outcomes were acquired by using a random forest transfer learning based on VGG-19 architecture. The classification results depended greatly on the number of categories. As the number of categories increased, the performance of deep learning models was diminished. When all 10 categories were included, we obtained results with an accuracy of 30.5%, relative classifier information (RCI) of 0.052, and Cohen's kappa of 0.224. Considering three integrated normal, background diabetic retinopathy, and dry age-related macular degeneration, the multi-categorical classifier showed accuracy of 72.8%, 0.283 RCI, and 0.577 kappa. In addition, several ensemble classifiers enhanced the multi-categorical classification performance. The transfer learning incorporated with ensemble classifier of clustering and voting approach presented the best performance with accuracy of 36.7%, 0.053 RCI, and 0.225 kappa in the 10 retinal diseases classification problem. First, due to the small size of datasets, the deep learning techniques in this study were ineffective to be applied in clinics where numerous patients suffering from various types of retinal disorders visit for diagnosis and treatment. Second, we found that the transfer learning incorporated with ensemble classifiers can improve the classification performance in order to detect multi-categorical retinal diseases. Further studies should confirm the effectiveness of algorithms with large datasets obtained from hospitals.

Fault Diagnosis Based on Chemical Sensor Data with an Active Deep Neural Network

PubMed Central

Jiang, Peng; Hu, Zhixin; Liu, Jun; Yu, Shanen; Wu, Feng

2016-01-01

Big sensor data provide significant potential for chemical fault diagnosis, which involves the baseline values of security, stability and reliability in chemical processes. A deep neural network (DNN) with novel active learning for inducing chemical fault diagnosis is presented in this study. It is a method using large amount of chemical sensor data, which is a combination of deep learning and active learning criterion to target the difficulty of consecutive fault diagnosis. DNN with deep architectures, instead of shallow ones, could be developed through deep learning to learn a suitable feature representation from raw sensor data in an unsupervised manner using stacked denoising auto-encoder (SDAE) and work through a layer-by-layer successive learning process. The features are added to the top Softmax regression layer to construct the discriminative fault characteristics for diagnosis in a supervised manner. Considering the expensive and time consuming labeling of sensor data in chemical applications, in contrast to the available methods, we employ a novel active learning criterion for the particularity of chemical processes, which is a combination of Best vs. Second Best criterion (BvSB) and a Lowest False Positive criterion (LFP), for further fine-tuning of diagnosis model in an active manner rather than passive manner. That is, we allow models to rank the most informative sensor data to be labeled for updating the DNN parameters during the interaction phase. The effectiveness of the proposed method is validated in two well-known industrial datasets. Results indicate that the proposed method can obtain superior diagnosis accuracy and provide significant performance improvement in accuracy and false positive rate with less labeled chemical sensor data by further active learning compared with existing methods. PMID:27754386

Fault Diagnosis Based on Chemical Sensor Data with an Active Deep Neural Network.

PubMed

Jiang, Peng; Hu, Zhixin; Liu, Jun; Yu, Shanen; Wu, Feng

2016-10-13

Big sensor data provide significant potential for chemical fault diagnosis, which involves the baseline values of security, stability and reliability in chemical processes. A deep neural network (DNN) with novel active learning for inducing chemical fault diagnosis is presented in this study. It is a method using large amount of chemical sensor data, which is a combination of deep learning and active learning criterion to target the difficulty of consecutive fault diagnosis. DNN with deep architectures, instead of shallow ones, could be developed through deep learning to learn a suitable feature representation from raw sensor data in an unsupervised manner using stacked denoising auto-encoder (SDAE) and work through a layer-by-layer successive learning process. The features are added to the top Softmax regression layer to construct the discriminative fault characteristics for diagnosis in a supervised manner. Considering the expensive and time consuming labeling of sensor data in chemical applications, in contrast to the available methods, we employ a novel active learning criterion for the particularity of chemical processes, which is a combination of Best vs. Second Best criterion (BvSB) and a Lowest False Positive criterion (LFP), for further fine-tuning of diagnosis model in an active manner rather than passive manner. That is, we allow models to rank the most informative sensor data to be labeled for updating the DNN parameters during the interaction phase. The effectiveness of the proposed method is validated in two well-known industrial datasets. Results indicate that the proposed method can obtain superior diagnosis accuracy and provide significant performance improvement in accuracy and false positive rate with less labeled chemical sensor data by further active learning compared with existing methods.

3-D Art Tasks.

ERIC Educational Resources Information Center

Niswander, Virginia

1983-01-01

Perceptual motor dysfunctions may not allow children with learning and behavior problems to perform as most children do. A successful art activity for these children is construction using wood scraps. (SR)

Studying the potential impact of automated document classification on scheduling a systematic review update.

PubMed

Cohen, Aaron M; Ambert, Kyle; McDonagh, Marian

2012-04-19

Systematic Reviews (SRs) are an essential part of evidence-based medicine, providing support for clinical practice and policy on a wide range of medical topics. However, producing SRs is resource-intensive, and progress in the research they review leads to SRs becoming outdated, requiring updates. Although the question of how and when to update SRs has been studied, the best method for determining when to update is still unclear, necessitating further research. In this work we study the potential impact of a machine learning-based automated system for providing alerts when new publications become available within an SR topic. Some of these new publications are especially important, as they report findings that are more likely to initiate a review update. To this end, we have designed a classification algorithm to identify articles that are likely to be included in an SR update, along with an annotation scheme designed to identify the most important publications in a topic area. Using an SR database containing over 70,000 articles, we annotated articles from 9 topics that had received an update during the study period. The algorithm was then evaluated in terms of the overall correct and incorrect alert rate for publications meeting the topic inclusion criteria, as well as in terms of its ability to identify important, update-motivating publications in a topic area. Our initial approach, based on our previous work in topic-specific SR publication classification, identifies over 70% of the most important new publications, while maintaining a low overall alert rate. We performed an initial analysis of the opportunities and challenges in aiding the SR update planning process with an informatics-based machine learning approach. Alerts could be a useful tool in the planning, scheduling, and allocation of resources for SR updates, providing an improvement in timeliness and coverage for the large number of medical topics needing SRs. While the performance of this initial method is not perfect, it could be a useful supplement to current approaches to scheduling an SR update. Approaches specifically targeting the types of important publications identified by this work are likely to improve results.

Studying the potential impact of automated document classification on scheduling a systematic review update

PubMed Central

2012-01-01

Background Systematic Reviews (SRs) are an essential part of evidence-based medicine, providing support for clinical practice and policy on a wide range of medical topics. However, producing SRs is resource-intensive, and progress in the research they review leads to SRs becoming outdated, requiring updates. Although the question of how and when to update SRs has been studied, the best method for determining when to update is still unclear, necessitating further research. Methods In this work we study the potential impact of a machine learning-based automated system for providing alerts when new publications become available within an SR topic. Some of these new publications are especially important, as they report findings that are more likely to initiate a review update. To this end, we have designed a classification algorithm to identify articles that are likely to be included in an SR update, along with an annotation scheme designed to identify the most important publications in a topic area. Using an SR database containing over 70,000 articles, we annotated articles from 9 topics that had received an update during the study period. The algorithm was then evaluated in terms of the overall correct and incorrect alert rate for publications meeting the topic inclusion criteria, as well as in terms of its ability to identify important, update-motivating publications in a topic area. Results Our initial approach, based on our previous work in topic-specific SR publication classification, identifies over 70% of the most important new publications, while maintaining a low overall alert rate. Conclusions We performed an initial analysis of the opportunities and challenges in aiding the SR update planning process with an informatics-based machine learning approach. Alerts could be a useful tool in the planning, scheduling, and allocation of resources for SR updates, providing an improvement in timeliness and coverage for the large number of medical topics needing SRs. While the performance of this initial method is not perfect, it could be a useful supplement to current approaches to scheduling an SR update. Approaches specifically targeting the types of important publications identified by this work are likely to improve results. PMID:22515596

Rb-Sr, Sm-Nd, K-Ca, O, and H isotopic study of Cretaceous-Tertiary boundary sediments, Caravaca, Spain: evidence for an oceanic impact site

USGS Publications Warehouse

DePaolo, D.J.; Kyte, F.T.; Marshall, B.D.; O'Neil, J.R.; Smit, J.

1983-01-01

Isotopic ratios and trace element abundances were measured on samples of Ir-enriched clay at the Cretaceous-Tertiary boundary, and in carbonate and marl from 5 cm below and 3 cm above the boundary. Samples were leached with acetic acid to remove carbonate, and with hydrochloric acid. Leachates and residues were measured. The Sr, Nd, O and H isotopic compositions of the boundary clay residues are distinct from those of the stratigraphically neighboring materials. The data indicate that most of the clay material was derived from a terrestrial source with relatively low 87Sr/86Sr and high 143Nd/144Nd ratios. The ??18O data suggest that the detritus has been modified by submarine weathering. K-Ca and Rb-Sr systematics, as well as O isotope ratios of K-feldspar spherules within the boundary clay, suggest that they are predominantly authigenic and may have formed after the time of deposition. However, Sm-Nd and Rb-Sr isotopic data indicate that the spherules contain relict material that provides information on the nature of the original detritus. The isotopic evidence for foreign terrestrial detritus in the boundary clay, the low rare earth element concentrations and high Ni concentration, support the hypothesis of a terminal Cretaceous asteroidal impact that produced a global layer of fallout. The data are most easily explained if the impact site was on oceanic crust rather than continental crust, and if a substantial fraction of the fallout was derived from relatively deep within the lithosphere (>3 km). This would probably require a single large impactor. ?? 1983.

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).

Concussion classification via deep learning using whole-brain white matter fiber strains

PubMed Central

Cai, Yunliang; Wu, Shaoju; Zhao, Wei; Li, Zhigang; Wu, Zheyang

2018-01-01

Developing an accurate and reliable injury predictor is central to the biomechanical studies of traumatic brain injury. State-of-the-art efforts continue to rely on empirical, scalar metrics based on kinematics or model-estimated tissue responses explicitly pre-defined in a specific brain region of interest. They could suffer from loss of information. A single training dataset has also been used to evaluate performance but without cross-validation. In this study, we developed a deep learning approach for concussion classification using implicit features of the entire voxel-wise white matter fiber strains. Using reconstructed American National Football League (NFL) injury cases, leave-one-out cross-validation was employed to objectively compare injury prediction performances against two baseline machine learning classifiers (support vector machine (SVM) and random forest (RF)) and four scalar metrics via univariate logistic regression (Brain Injury Criterion (BrIC), cumulative strain damage measure of the whole brain (CSDM-WB) and the corpus callosum (CSDM-CC), and peak fiber strain in the CC). Feature-based machine learning classifiers including deep learning, SVM, and RF consistently outperformed all scalar injury metrics across all performance categories (e.g., leave-one-out accuracy of 0.828–0.862 vs. 0.690–0.776, and .632+ error of 0.148–0.176 vs. 0.207–0.292). Further, deep learning achieved the best cross-validation accuracy, sensitivity, AUC, and .632+ error. These findings demonstrate the superior performances of deep learning in concussion prediction and suggest its promise for future applications in biomechanical investigations of traumatic brain injury. PMID:29795640

Concussion classification via deep learning using whole-brain white matter fiber strains.

PubMed

Cai, Yunliang; Wu, Shaoju; Zhao, Wei; Li, Zhigang; Wu, Zheyang; Ji, Songbai

2018-01-01

Developing an accurate and reliable injury predictor is central to the biomechanical studies of traumatic brain injury. State-of-the-art efforts continue to rely on empirical, scalar metrics based on kinematics or model-estimated tissue responses explicitly pre-defined in a specific brain region of interest. They could suffer from loss of information. A single training dataset has also been used to evaluate performance but without cross-validation. In this study, we developed a deep learning approach for concussion classification using implicit features of the entire voxel-wise white matter fiber strains. Using reconstructed American National Football League (NFL) injury cases, leave-one-out cross-validation was employed to objectively compare injury prediction performances against two baseline machine learning classifiers (support vector machine (SVM) and random forest (RF)) and four scalar metrics via univariate logistic regression (Brain Injury Criterion (BrIC), cumulative strain damage measure of the whole brain (CSDM-WB) and the corpus callosum (CSDM-CC), and peak fiber strain in the CC). Feature-based machine learning classifiers including deep learning, SVM, and RF consistently outperformed all scalar injury metrics across all performance categories (e.g., leave-one-out accuracy of 0.828-0.862 vs. 0.690-0.776, and .632+ error of 0.148-0.176 vs. 0.207-0.292). Further, deep learning achieved the best cross-validation accuracy, sensitivity, AUC, and .632+ error. These findings demonstrate the superior performances of deep learning in concussion prediction and suggest its promise for future applications in biomechanical investigations of traumatic brain injury.

Prediction of Bispectral Index during Target-controlled Infusion of Propofol and Remifentanil: A Deep Learning Approach.

PubMed

Lee, Hyung-Chul; Ryu, Ho-Geol; Chung, Eun-Jin; Jung, Chul-Woo

2018-03-01

The discrepancy between predicted effect-site concentration and measured bispectral index is problematic during intravenous anesthesia with target-controlled infusion of propofol and remifentanil. We hypothesized that bispectral index during total intravenous anesthesia would be more accurately predicted by a deep learning approach. Long short-term memory and the feed-forward neural network were sequenced to simulate the pharmacokinetic and pharmacodynamic parts of an empirical model, respectively, to predict intraoperative bispectral index during combined use of propofol and remifentanil. Inputs of long short-term memory were infusion histories of propofol and remifentanil, which were retrieved from target-controlled infusion pumps for 1,800 s at 10-s intervals. Inputs of the feed-forward network were the outputs of long short-term memory and demographic data such as age, sex, weight, and height. The final output of the feed-forward network was the bispectral index. The performance of bispectral index prediction was compared between the deep learning model and previously reported response surface model. The model hyperparameters comprised 8 memory cells in the long short-term memory layer and 16 nodes in the hidden layer of the feed-forward network. The model training and testing were performed with separate data sets of 131 and 100 cases. The concordance correlation coefficient (95% CI) were 0.561 (0.560 to 0.562) in the deep learning model, which was significantly larger than that in the response surface model (0.265 [0.263 to 0.266], P < 0.001). The deep learning model-predicted bispectral index during target-controlled infusion of propofol and remifentanil more accurately compared to the traditional model. The deep learning approach in anesthetic pharmacology seems promising because of its excellent performance and extensibility.

Improved Automated Detection of Diabetic Retinopathy on a Publicly Available Dataset Through Integration of Deep Learning.

PubMed

Abràmoff, Michael David; Lou, Yiyue; Erginay, Ali; Clarida, Warren; Amelon, Ryan; Folk, James C; Niemeijer, Meindert

2016-10-01

To compare performance of a deep-learning enhanced algorithm for automated detection of diabetic retinopathy (DR), to the previously published performance of that algorithm, the Iowa Detection Program (IDP)-without deep learning components-on the same publicly available set of fundus images and previously reported consensus reference standard set, by three US Board certified retinal specialists. We used the previously reported consensus reference standard of referable DR (rDR), defined as International Clinical Classification of Diabetic Retinopathy moderate, severe nonproliferative (NPDR), proliferative DR, and/or macular edema (ME). Neither Messidor-2 images, nor the three retinal specialists setting the Messidor-2 reference standard were used for training IDx-DR version X2.1. Sensitivity, specificity, negative predictive value, area under the curve (AUC), and their confidence intervals (CIs) were calculated. Sensitivity was 96.8% (95% CI: 93.3%-98.8%), specificity was 87.0% (95% CI: 84.2%-89.4%), with 6/874 false negatives, resulting in a negative predictive value of 99.0% (95% CI: 97.8%-99.6%). No cases of severe NPDR, PDR, or ME were missed. The AUC was 0.980 (95% CI: 0.968-0.992). Sensitivity was not statistically different from published IDP sensitivity, which had a CI of 94.4% to 99.3%, but specificity was significantly better than the published IDP specificity CI of 55.7% to 63.0%. A deep-learning enhanced algorithm for the automated detection of DR, achieves significantly better performance than a previously reported, otherwise essentially identical, algorithm that does not employ deep learning. Deep learning enhanced algorithms have the potential to improve the efficiency of DR screening, and thereby to prevent visual loss and blindness from this devastating disease.

Automated detection of exudative age-related macular degeneration in spectral domain optical coherence tomography using deep learning.

PubMed

Treder, Maximilian; Lauermann, Jost Lennart; Eter, Nicole

2018-02-01

Our purpose was to use deep learning for the automated detection of age-related macular degeneration (AMD) in spectral domain optical coherence tomography (SD-OCT). A total of 1112 cross-section SD-OCT images of patients with exudative AMD and a healthy control group were used for this study. In the first step, an open-source multi-layer deep convolutional neural network (DCNN), which was pretrained with 1.2 million images from ImageNet, was trained and validated with 1012 cross-section SD-OCT scans (AMD: 701; healthy: 311). During this procedure training accuracy, validation accuracy and cross-entropy were computed. The open-source deep learning framework TensorFlow™ (Google Inc., Mountain View, CA, USA) was used to accelerate the deep learning process. In the last step, a created DCNN classifier, using the information of the above mentioned deep learning process, was tested in detecting 100 untrained cross-section SD-OCT images (AMD: 50; healthy: 50). Therefore, an AMD testing score was computed: 0.98 or higher was presumed for AMD. After an iteration of 500 training steps, the training accuracy and validation accuracies were 100%, and the cross-entropy was 0.005. The average AMD scores were 0.997 ± 0.003 in the AMD testing group and 0.9203 ± 0.085 in the healthy comparison group. The difference between the two groups was highly significant (p < 0.001). With a deep learning-based approach using TensorFlow™, it is possible to detect AMD in SD-OCT with high sensitivity and specificity. With more image data, an expansion of this classifier for other macular diseases or further details in AMD is possible, suggesting an application for this model as a support in clinical decisions. Another possible future application would involve the individual prediction of the progress and success of therapy for different diseases by automatically detecting hidden image information.

Deep SOMs for automated feature extraction and classification from big data streaming

NASA Astrophysics Data System (ADS)

Sakkari, Mohamed; Ejbali, Ridha; Zaied, Mourad

2017-03-01

In this paper, we proposed a deep self-organizing map model (Deep-SOMs) for automated features extracting and learning from big data streaming which we benefit from the framework Spark for real time streams and highly parallel data processing. The SOMs deep architecture is based on the notion of abstraction (patterns automatically extract from the raw data, from the less to more abstract). The proposed model consists of three hidden self-organizing layers, an input and an output layer. Each layer is made up of a multitude of SOMs, each map only focusing at local headmistress sub-region from the input image. Then, each layer trains the local information to generate more overall information in the higher layer. The proposed Deep-SOMs model is unique in terms of the layers architecture, the SOMs sampling method and learning. During the learning stage we use a set of unsupervised SOMs for feature extraction. We validate the effectiveness of our approach on large data sets such as Leukemia dataset and SRBCT. Results of comparison have shown that the Deep-SOMs model performs better than many existing algorithms for images classification.

DeepID-Net: Deformable Deep Convolutional Neural Networks for Object Detection.

PubMed

Ouyang, Wanli; Zeng, Xingyu; Wang, Xiaogang; Qiu, Shi; Luo, Ping; Tian, Yonglong; Li, Hongsheng; Yang, Shuo; Wang, Zhe; Li, Hongyang; Loy, Chen Change; Wang, Kun; Yan, Junjie; Tang, Xiaoou

2016-07-07

In this paper, we propose deformable deep convolutional neural networks for generic object detection. This new deep learning object detection framework has innovations in multiple aspects. In the proposed new deep architecture, a new deformation constrained pooling (def-pooling) layer models the deformation of object parts with geometric constraint and penalty. A new pre-training strategy is proposed to learn feature representations more suitable for the object detection task and with good generalization capability. By changing the net structures, training strategies, adding and removing some key components in the detection pipeline, a set of models with large diversity are obtained, which significantly improves the effectiveness of model averaging. The proposed approach improves the mean averaged precision obtained by RCNN [16], which was the state-of-the-art, from 31% to 50.3% on the ILSVRC2014 detection test set. It also outperforms the winner of ILSVRC2014, GoogLeNet, by 6.1%. Detailed component-wise analysis is also provided through extensive experimental evaluation, which provides a global view for people to understand the deep learning object detection pipeline.

First-Year Students' Approaches to Learning, and Factors Related to Change or Stability in Their Deep Approach during a Pharmacy Course

ERIC Educational Resources Information Center

Varunki, Maaret; Katajavuori, Nina; Postareff, Liisa

2017-01-01

Research shows that a surface approach to learning is more common among students in the natural sciences, while students representing the "soft" sciences are more likely to apply a deep approach. However, findings conflict concerning the stability of approaches to learning in general. This study explores the variation in students'…

Nonparametric Representations for Integrated Inference, Control, and Sensing

DTIC Science & Technology

2015-10-01

Learning (ICML), 2013. [20] Jeff Donahue, Yangqing Jia, Oriol Vinyals, Judy Hoffman, Ning Zhang, Eric Tzeng, and Trevor Darrell. DeCAF: A deep ...unlimited. Multi-layer feature learning “SuperVision” Convolutional Neural Network (CNN) ImageNet Classification with Deep Convolutional Neural Networks...to develop a new framework for autonomous operations that will extend the state of the art in distributed learning and modeling from data, and

Time-lagged autoencoders: Deep learning of slow collective variables for molecular kinetics

NASA Astrophysics Data System (ADS)

Wehmeyer, Christoph; Noé, Frank

2018-06-01

Inspired by the success of deep learning techniques in the physical and chemical sciences, we apply a modification of an autoencoder type deep neural network to the task of dimension reduction of molecular dynamics data. We can show that our time-lagged autoencoder reliably finds low-dimensional embeddings for high-dimensional feature spaces which capture the slow dynamics of the underlying stochastic processes—beyond the capabilities of linear dimension reduction techniques.

CGBVS-DNN: Prediction of Compound-protein Interactions Based on Deep Learning.

PubMed

Hamanaka, Masatoshi; Taneishi, Kei; Iwata, Hiroaki; Ye, Jun; Pei, Jianguo; Hou, Jinlong; Okuno, Yasushi

2017-01-01

Computational prediction of compound-protein interactions (CPIs) is of great importance for drug design as the first step in in-silico screening. We previously proposed chemical genomics-based virtual screening (CGBVS), which predicts CPIs by using a support vector machine (SVM). However, the CGBVS has problems when training using more than a million datasets of CPIs since SVMs require an exponential increase in the calculation time and computer memory. To solve this problem, we propose the CGBVS-DNN, in which we use deep neural networks, a kind of deep learning technique, instead of the SVM. Deep learning does not require learning all input data at once because the network can be trained with small mini-batches. Experimental results show that the CGBVS-DNN outperformed the original CGBVS with a quarter million CPIs. Results of cross-validation show that the accuracy of the CGBVS-DNN reaches up to 98.2 % (σ<0.01) with 4 million CPIs. © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

Relevance of deep learning to facilitate the diagnosis of HER2 status in breast cancer

NASA Astrophysics Data System (ADS)

Vandenberghe, Michel E.; Scott, Marietta L. J.; Scorer, Paul W.; Söderberg, Magnus; Balcerzak, Denis; Barker, Craig

2017-04-01

Tissue biomarker scoring by pathologists is central to defining the appropriate therapy for patients with cancer. Yet, inter-pathologist variability in the interpretation of ambiguous cases can affect diagnostic accuracy. Modern artificial intelligence methods such as deep learning have the potential to supplement pathologist expertise to ensure constant diagnostic accuracy. We developed a computational approach based on deep learning that automatically scores HER2, a biomarker that defines patient eligibility for anti-HER2 targeted therapies in breast cancer. In a cohort of 71 breast tumour resection samples, automated scoring showed a concordance of 83% with a pathologist. The twelve discordant cases were then independently reviewed, leading to a modification of diagnosis from initial pathologist assessment for eight cases. Diagnostic discordance was found to be largely caused by perceptual differences in assessing HER2 expression due to high HER2 staining heterogeneity. This study provides evidence that deep learning aided diagnosis can facilitate clinical decision making in breast cancer by identifying cases at high risk of misdiagnosis.

Geochemical evidence for possible natural migration of Marcellus Formation brine to shallow aquifers in Pennsylvania

PubMed Central

Warner, Nathaniel R.; Jackson, Robert B.; Darrah, Thomas H.; Osborn, Stephen G.; Down, Adrian; Zhao, Kaiguang; White, Alissa; Vengosh, Avner

2012-01-01

The debate surrounding the safety of shale gas development in the Appalachian Basin has generated increased awareness of drinking water quality in rural communities. Concerns include the potential for migration of stray gas, metal-rich formation brines, and hydraulic fracturing and/or flowback fluids to drinking water aquifers. A critical question common to these environmental risks is the hydraulic connectivity between the shale gas formations and the overlying shallow drinking water aquifers. We present geochemical evidence from northeastern Pennsylvania showing that pathways, unrelated to recent drilling activities, exist in some locations between deep underlying formations and shallow drinking water aquifers. Integration of chemical data (Br, Cl, Na, Ba, Sr, and Li) and isotopic ratios (87Sr/86Sr, 2H/H, 18O/16O, and 228Ra/226Ra) from this and previous studies in 426 shallow groundwater samples and 83 northern Appalachian brine samples suggest that mixing relationships between shallow ground water and a deep formation brine causes groundwater salinization in some locations. The strong geochemical fingerprint in the salinized (Cl > 20 mg/L) groundwater sampled from the Alluvium, Catskill, and Lock Haven aquifers suggests possible migration of Marcellus brine through naturally occurring pathways. The occurrences of saline water do not correlate with the location of shale-gas wells and are consistent with reported data before rapid shale-gas development in the region; however, the presence of these fluids suggests conductive pathways and specific geostructural and/or hydrodynamic regimes in northeastern Pennsylvania that are at increased risk for contamination of shallow drinking water resources, particularly by fugitive gases, because of natural hydraulic connections to deeper formations. PMID:22778445

Geochemical evidence for possible natural migration of Marcellus Formation brine to shallow aquifers in Pennsylvania.

PubMed

Warner, Nathaniel R; Jackson, Robert B; Darrah, Thomas H; Osborn, Stephen G; Down, Adrian; Zhao, Kaiguang; White, Alissa; Vengosh, Avner

2012-07-24

The debate surrounding the safety of shale gas development in the Appalachian Basin has generated increased awareness of drinking water quality in rural communities. Concerns include the potential for migration of stray gas, metal-rich formation brines, and hydraulic fracturing and/or flowback fluids to drinking water aquifers. A critical question common to these environmental risks is the hydraulic connectivity between the shale gas formations and the overlying shallow drinking water aquifers. We present geochemical evidence from northeastern Pennsylvania showing that pathways, unrelated to recent drilling activities, exist in some locations between deep underlying formations and shallow drinking water aquifers. Integration of chemical data (Br, Cl, Na, Ba, Sr, and Li) and isotopic ratios ((87)Sr/(86)Sr, (2)H/H, (18)O/(16)O, and (228)Ra/(226)Ra) from this and previous studies in 426 shallow groundwater samples and 83 northern Appalachian brine samples suggest that mixing relationships between shallow ground water and a deep formation brine causes groundwater salinization in some locations. The strong geochemical fingerprint in the salinized (Cl > 20 mg/L) groundwater sampled from the Alluvium, Catskill, and Lock Haven aquifers suggests possible migration of Marcellus brine through naturally occurring pathways. The occurrences of saline water do not correlate with the location of shale-gas wells and are consistent with reported data before rapid shale-gas development in the region; however, the presence of these fluids suggests conductive pathways and specific geostructural and/or hydrodynamic regimes in northeastern Pennsylvania that are at increased risk for contamination of shallow drinking water resources, particularly by fugitive gases, because of natural hydraulic connections to deeper formations.

Evaluation of the sparse coding super-resolution method for improving image quality of up-sampled images in computed tomography

NASA Astrophysics Data System (ADS)

Ota, Junko; Umehara, Kensuke; Ishimaru, Naoki; Ohno, Shunsuke; Okamoto, Kentaro; Suzuki, Takanori; Shirai, Naoki; Ishida, Takayuki

2017-02-01

As the capability of high-resolution displays grows, high-resolution images are often required in Computed Tomography (CT). However, acquiring high-resolution images takes a higher radiation dose and a longer scanning time. In this study, we applied the Sparse-coding-based Super-Resolution (ScSR) method to generate high-resolution images without increasing the radiation dose. We prepared the over-complete dictionary learned the mapping between low- and highresolution patches and seek a sparse representation of each patch of the low-resolution input. These coefficients were used to generate the high-resolution output. For evaluation, 44 CT cases were used as the test dataset. We up-sampled images up to 2 or 4 times and compared the image quality of the ScSR scheme and bilinear and bicubic interpolations, which are the traditional interpolation schemes. We also compared the image quality of three learning datasets. A total of 45 CT images, 91 non-medical images, and 93 chest radiographs were used for dictionary preparation respectively. The image quality was evaluated by measuring peak signal-to-noise ratio (PSNR) and structure similarity (SSIM). The differences of PSNRs and SSIMs between the ScSR method and interpolation methods were statistically significant. Visual assessment confirmed that the ScSR method generated a high-resolution image with sharpness, whereas conventional interpolation methods generated over-smoothed images. To compare three different training datasets, there were no significance between the CT, the CXR and non-medical datasets. These results suggest that the ScSR provides a robust approach for application of up-sampling CT images and yields substantial high image quality of extended images in CT.