Search

Showing total 9 results
Start Over Topic artificial neural network Language english
9 results

1. Binary level toolchain provenance identification with graph neural networks

Author

Tristan Benoit, Jean-Yves Marion, Sébastien Bardin, Carbone (CARBONE), Department of Formal Methods (LORIA - FM), Laboratoire Lorrain de Recherche en Informatique et ses Applications (LORIA), Institut National de Recherche en Informatique et en Automatique (Inria)-Université de Lorraine (UL)-Centre National de la Recherche Scientifique (CNRS)-Institut National de Recherche en Informatique et en Automatique (Inria)-Université de Lorraine (UL)-Centre National de la Recherche Scientifique (CNRS)-Laboratoire Lorrain de Recherche en Informatique et ses Applications (LORIA), Institut National de Recherche en Informatique et en Automatique (Inria)-Université de Lorraine (UL)-Centre National de la Recherche Scientifique (CNRS)-Institut National de Recherche en Informatique et en Automatique (Inria)-Université de Lorraine (UL)-Centre National de la Recherche Scientifique (CNRS), CEA- Saclay (CEA), Commissariat à l'énergie atomique et aux énergies alternatives (CEA), This work is supported by (i) a public grant overseen by the French National Research Agency (ANR) as part of the 'Investissements d'Avenir' French PIA project 'Lorraine Université d'Excellence', reference ANR-15-IDEX-04-LUE, and (ii) has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 830927 (Concordia). Experiments presented in this paper were carried out using the Grid’5000 testbed, supported by a scientific interest group hosted by Inria and including CNRS, RENATER and several Universities as well as other organizations (see https://www.grid5000.fr)., GRID5000, IMPACT-DIGITRUST, ANR-15-IDEX-0004,LUE,Isite LUE(2015), European Project: 830927,CONCORDIA(2019), Centre National de la Recherche Scientifique (CNRS)-Université de Lorraine (UL)-Institut National de Recherche en Informatique et en Automatique (Inria)-Centre National de la Recherche Scientifique (CNRS)-Université de Lorraine (UL)-Institut National de Recherche en Informatique et en Automatique (Inria)-Laboratoire Lorrain de Recherche en Informatique et ses Applications (LORIA), and Centre National de la Recherche Scientifique (CNRS)-Université de Lorraine (UL)-Institut National de Recherche en Informatique et en Automatique (Inria)-Centre National de la Recherche Scientifique (CNRS)-Université de Lorraine (UL)-Institut National de Recherche en Informatique et en Automatique (Inria)

Subjects
Theoretical computer science, Artificial neural network, Computer science, Binary number, graph neural networks, 020207 software engineering, 02 engineering and technology, computer.software_genre, toolchain provenance, Toolchain, [INFO.INFO-CR]Computer Science [cs]/Cryptography and Security [cs.CR], [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Scalability, 0202 electrical engineering, electronic engineering, information engineering, Graph reduction, ACM: D.: Software/D.2: SOFTWARE ENGINEERING/D.2.7: Distribution, Maintenance, and Enhancement/D.2.7.5: Restructuring, reverse engineering, and reengineering, ACM: D.: Software/D.2: SOFTWARE ENGINEERING/D.2.5: Testing and Debugging/D.2.5.2: Diagnostics, Control flow graph, binary code analysis, 020201 artificial intelligence & image processing, Binary code, Compiler, computer, ACM: I.: Computing Methodologies/I.2: ARTIFICIAL INTELLIGENCE/I.2.6: Learning
Abstract

International audience; We consider the problem of recovering the compiling chain used to generate a given stripped binary code. We present a Graph Neural Network framework at the binary level to solve this problem, with the idea to take into account the shallow semantics provided by the binary code's structured control flow graph (CFG).We introduce a Graph Neural Network, called Site Neural Network (SNN), dedicated to this problem. To attain scalability at the binary level, feature extraction is simplified by forgetting almost everything in a CFG except transfer control instructions and performing a parametric graph reduction. Our experiments show that our method recovers the compiler family with a very high F1-Score of 0.9950 while the optimization level is recovered with a moderately high F1-Score of 0.7517. On the compiler version prediction task, the F1-Score is about 0.8167 excluding the clang family. A comparison with a previous work demonstrates the accuracy and performance of this framework.

Published
2021

2. Evolution Control for parallel ANN-assisted simulation-based optimization application to Tuberculosis Transmission Control

Author

Mohand-Said Mezmaz, Nouredine Melab, Romain Ragonnet, Guillaume Briffoteaux, Daniel Tuyttens, Optimisation de grande taille et calcul large échelle (BONUS), Inria Lille - Nord Europe, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 (CRIStAL), Centrale Lille-Université de Lille-Centre National de la Recherche Scientifique (CNRS)-Centrale Lille-Université de Lille-Centre National de la Recherche Scientifique (CNRS), University of Mons [Belgium] (UMONS), Monash University [Melbourne], Laboratoire d'Informatique Fondamentale de Lille (LIFL), Université de Lille, Sciences et Technologies-Institut National de Recherche en Informatique et en Automatique (Inria)-Université de Lille, Sciences Humaines et Sociales-Centre National de la Recherche Scientifique (CNRS), and Experiments presented in this paper were carried out using the Grid’5000 testbed, supported by a scientific interest group hosted by Inria and including CNRS, RENATER and several Universities as well as other organizations (see https://www.grid5000.fr).

Subjects
Artificial Neural Network, Computer Networks and Communications, Computer science, Monte Carlo method, Context (language use), 02 engineering and technology, [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], Simulation-based optimization, Surrogate-assisted Optimization, 0202 electrical engineering, electronic engineering, information engineering, Massively parallel, Dropout (neural networks), Artificial neural network, Evolution Control, business.industry, Deep learning, 020206 networking & telecommunications, [INFO.INFO-RO]Computer Science [cs]/Operations Research [cs.RO], Supercomputer, Computer engineering, Hardware and Architecture, 020201 artificial intelligence & image processing, Artificial intelligence, Massively Parallel Computing, [INFO.INFO-DC]Computer Science [cs]/Distributed, Parallel, and Cluster Computing [cs.DC], business, Software, Simulation
Abstract

International audience; In many optimal design searches, the function to optimise is a simulator that is computationally expensive. While current High Performance Computing (HPC) methods are not able to solve such problems efficiently, parallelism can be coupled with approximate models (surrogates or meta-models) that imitate the simulator in timely fashion to achieve better results. This combined approach reduces the number of simulations thanks to surrogate use whereas the remaining evaluations are handled by supercomputers. While the surrogates' ability to limit computational times is very attractive, integrating them into the over-arching optimization process can be challenging. Indeed, it is critical to address the major trade-off between the quality (precision) and the efficiency (execution time) of the resolution. In this article, we investigate Evolution Controls (ECs) which are strategies that define the alternation between the simulator and the surrogate within the optimization process. We propose a new EC based on the prediction uncertainty obtained from Monte Carlo Dropout (MCDropout), a technique originally dedicated to quantifying uncertainty in deep learning. Investigations of such uncertainty-aware ECs remain uncommon in surrogate-assisted evolutionary optimization. In addition, we use parallel computing in a complementary way to address the high computational burden. Our new strategy is implemented in the context of a pioneering application to Tuberculosis Transmission Control. The reported results show that the MCDropout-based EC coupled with massively parallel computing outperforms strategies previously proposed in the field of surrogate-assisted optimization.

Published
2020

3. DNN-based mask estimation for distributed speech enhancement in spatially unconstrained microphone arrays

Author

Romain Serizel, Irina Illina, Nicolas Furnon, Slim Essid, Speech Modeling for Facilitating Oral-Based Communication (MULTISPEECH), Inria Nancy - Grand Est, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Department of Natural Language Processing & Knowledge Discovery (LORIA - NLPKD), Laboratoire Lorrain de Recherche en Informatique et ses Applications (LORIA), Institut National de Recherche en Informatique et en Automatique (Inria)-Université de Lorraine (UL)-Centre National de la Recherche Scientifique (CNRS)-Institut National de Recherche en Informatique et en Automatique (Inria)-Université de Lorraine (UL)-Centre National de la Recherche Scientifique (CNRS)-Laboratoire Lorrain de Recherche en Informatique et ses Applications (LORIA), Institut National de Recherche en Informatique et en Automatique (Inria)-Université de Lorraine (UL)-Centre National de la Recherche Scientifique (CNRS)-Université de Lorraine (UL)-Centre National de la Recherche Scientifique (CNRS), Laboratoire Traitement et Communication de l'Information (LTCI), Institut Mines-Télécom [Paris] (IMT)-Télécom Paris, This work was made with the support of the French National Research Agency, in the framework of the project DiSCogs (ANR-17-CE23-0026-01). Experiments presented in this paper were partially out using the Grid5000 testbed, supported by a scientific interest group hosted by Inria and including CNRS, RENATER and several Universities as well as other organizations (see https://www.grid5000)., Grid'5000, ANR-17-CE23-0026,DiSCogs,Antennes acoustiques hétérogènes et non contraintes pour la communication parlée(2017), Institut Polytechnique de Paris (IP Paris), Département Images, Données, Signal (IDS), Télécom ParisTech, Signal, Statistique et Apprentissage (S2A), and Institut Mines-Télécom [Paris] (IMT)-Télécom Paris-Institut Mines-Télécom [Paris] (IMT)-Télécom Paris

Subjects
Signal Processing (eess.SP), Microphone array, Acoustics and Ultrasonics, Noise measurement, Artificial neural network, Computer science, Microphone, Noise reduction, Speech recognition, Context (language use), Speech processing, Speech enhancement, Computational Mathematics, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, Computer Science::Sound, FOS: Electrical engineering, electronic engineering, information engineering, Computer Science (miscellaneous), Electrical and Electronic Engineering, Electrical Engineering and Systems Science - Signal Processing
Abstract

Deep neural network (DNN)-based speech enhancement algorithms in microphone arrays have now proven to be efficient solutions to speech understanding and speech recognition in noisy environments. However, in the context of ad-hoc microphone arrays, many challenges remain and raise the need for distributed processing. In this paper, we propose to extend a previously introduced distributed DNN-based time-frequency mask estimation scheme that can efficiently use spatial information in form of so-called compressed signals which are pre-filtered target estimations. We study the performance of this algorithm under realistic acoustic conditions and investigate practical aspects of its optimal application. We show that the nodes in the microphone array cooperate by taking profit of their spatial coverage in the room. We also propose to use the compressed signals not only to convey the target estimation but also the noise estimation in order to exploit the acoustic diversity recorded throughout the microphone array., Submitted to TASLP

Published
2020

4. Machines and Masterpieces: Predicting Prices in the Art Auction Market

Author

Roman Kräussl, Gustavo Manso, Mathieu Aubry, Christophe Spaenjers, Laboratoire d'Informatique Gaspard-Monge (LIGM), Centre National de la Recherche Scientifique (CNRS)-Fédération de Recherche Bézout-ESIEE Paris-École des Ponts ParisTech (ENPC)-Université Paris-Est Marne-la-Vallée (UPEM), Department of Economics, Tilburg University [Netherlands], HEC Research Paper Series, and Université Paris-Est Marne-la-Vallée (UPEM)-École des Ponts ParisTech (ENPC)-ESIEE Paris-Fédération de Recherche Bézout-Centre National de la Recherche Scientifique (CNRS)

Subjects
asset valuation, History, Polymers and Plastics, Computer science, Big data, Machine learning, computer.software_genre, Industrial and Manufacturing Engineering, big data, 0502 economics and business, Common value auction, Price level, auctions, 050207 economics, Business and International Management, JEL: C - Mathematical and Quantitative Methods/C.C5 - Econometric Modeling/C.C5.C50 - General, Valuation (finance), JEL: D - Microeconomics/D.D4 - Market Structure, Pricing, and Design/D.D4.D44 - Auctions, 050208 finance, Artificial neural network, Ex-ante, business.industry, 05 social sciences, experts, Hedonic pricing, TheoryofComputation_GENERAL, JEL: Z - Other Special Topics/Z.Z1 - Cultural Economics • Economic Sociology • Economic Anthropology/Z.Z1.Z11 - Economics of the Arts and Literature, machine learning, [SHS.GESTION]Humanities and Social Sciences/Business administration, Artificial intelligence, business, computer, JEL: G - Financial Economics/G.G1 - General Financial Markets/G.G1.G12 - Asset Pricing • Trading Volume • Bond Interest Rates
Abstract

We study the accuracy and usefulness of automated (i.e., machine-generated) valuations for illiquid and heterogeneous real assets. We assemble a database of 1.1 million paintings auctioned between 2008 and 2015. We use a popular machine-learning technique - neural networks - to develop a pricing algorithm based on both non-visual and visual artwork characteristics. Our out-of-sample valuations predict auction prices dramatically better than valuations based on a standard hedonic pricing model. Moreover, they help explaining price levels and sale probabilities even after conditioning on auctioneers' pre-sale estimates. Machine learning is particularly helpful for assets that are associated with high price uncertainty. It can also correct human experts' systematic biases in expectations formation - and identify ex ante situations in which such biases are likely to arise.

Published
2020

5. Parallel Surrogate-assisted Optimization: Batched Bayesian Neural Network-assisted GA versus q-EGO

Author

Guillaume Briffoteaux, Romain Ragonnet, Jan Gmys, Nouredine Melab, Mohand Mezmaz, Maxime Gobert, Daniel Tuyttens, Optimisation de grande taille et calcul large échelle (BONUS), Inria Lille - Nord Europe, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 (CRIStAL), Centrale Lille-Université de Lille-Centre National de la Recherche Scientifique (CNRS)-Centrale Lille-Université de Lille-Centre National de la Recherche Scientifique (CNRS), University of Mons [Belgium] (UMONS), Monash University [Melbourne], Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 (CRIStAL), Centrale Lille-Université de Lille-Centre National de la Recherche Scientifique (CNRS), Institut de Mathématiques [Mons], Université de Mons (UMons), Laboratoire d'Informatique Fondamentale de Lille (LIFL), Université de Lille, Sciences et Technologies-Institut National de Recherche en Informatique et en Automatique (Inria)-Université de Lille, Sciences Humaines et Sociales-Centre National de la Recherche Scientifique (CNRS), and Experiments presented in this paper were carried out using the Grid’5000 testbed, supported by a scientific interest group hosted by Inria and including CNRS, RENATER and several Universities as well as other organizations (see https://www.grid5000.fr).

Subjects
General Computer Science, Computer science, General Mathematics, Evolutionary algorithm, 02 engineering and technology, [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], Machine learning, computer.software_genre, Surrogate model, Genetic algorithm, Surrogate-assisted Optimization, 0202 electrical engineering, electronic engineering, information engineering, Massively parallel, Global optimization, Artificial neural network, Bayesian Optimization, business.industry, 05 social sciences, Bayesian optimization, Evolutionary Algorithm, 050301 education, [INFO.INFO-RO]Computer Science [cs]/Operations Research [cs.RO], Efficient Global Optimization, Benchmark (computing), 020201 artificial intelligence & image processing, Artificial intelligence, Massively Parallel Computing, [INFO.INFO-DC]Computer Science [cs]/Distributed, Parallel, and Cluster Computing [cs.DC], business, 0503 education, computer, Simulation
Abstract

International audience; Surrogate-based optimization is widely used to deal with long-running black-box simulation-based objective functions. Actually, the use of a surrogate model such as Kriging or Artificial Neural Network allows to reduce the number of calls to the CPU time-intensive simulator. Bayesian optimization uses the ability of surrogates to provide useful information to help guiding effectively the optimization process. In this paper, the Efficient Global Optimization (EGO) reference framework is challenged by a Bayesian Neural Network-assisted Genetic Algorithm, namely BNN-GA. The Bayesian Neural Network (BNN) surrogate is chosen for its ability to provide an uncertainty measure of the prediction that allows to compute the Expected Improvement of a candidate solution in order to improve the exploration of the objective space. BNN is also more reliable than Kriging models for high-dimensional problems and faster to set up thanks to its incremental training. In addition, we propose a batch-based approach for the parallelization of BNN-GA that is challenged by a parallel version of EGO, called q-EGO. Parallel computing is a highly important complementary way (to surrogates) to deal with the computational burden of simulation-based optimization. The comparison of the two parallel approaches is experimentally performed through several benchmark functions and two real-world problems within the scope of Tuberculosis Transmission Control (TBTC). The study presented in this paper proves that parallel batched BNN-GA is a viable alternative to q-EGO approaches being more suitable for high-dimensional problems, parallelization impact, bigger databases and moderate search budgets. Moreover, a significant improvement of the solutions is obtained for the two TBTC problems tackled.

Published
2020

6. Flood forecasting based on an artificial neural network scheme

Author

Chafiq Titouna, Abdelhak Mourad Gueroui, Ousmane Thiare, Ado Adamou Abba Ari, Francis Yongwa Dtissibe, University of Maroua (UMa), Laboratoire d'Informatique Parallélisme Réseaux Algorithmes Distribués (LI-PaRAD), Université de Versailles Saint-Quentin-en-Yvelines (UVSQ), Laboratoire d'Informatique Paris Descartes (LIPADE (URP_2517)), Université de Paris (UP), Laboratoire d'Analyse Numérique et Informatique [Sénégal] (LANI), Université Gaston Berger de Saint-Louis Sénégal (UGB), and We like to thank the editor and the anonymous reviewers for their valuable remarks that helped us in better improving the content and presentation of the paper.

Subjects
021110 strategic, defence & security studies, Atmospheric Science, Flood forecasting, 010504 meteorology & atmospheric sciences, Artificial neural network, Flood myth, Artificial neural networks, Computer science, 0211 other engineering and technologies, 02 engineering and technology, 01 natural sciences, Industrial engineering, Hazard, 13. Climate action, Order (exchange), Multilayer perceptron, Natural hazard, Machine learning, [SDE]Environmental Sciences, Earth and Planetary Sciences (miscellaneous), 0105 earth and related environmental sciences, Water Science and Technology, Communication channel
Abstract

International audience; Nowadays, floods have become the widest global environmental and economic hazard in many countries, causing huge loss of lives and materials damages. It is, therefore, necessary to build an efficient flood forecasting system. The physical-based flood forecasting methods have indeed proven to be limited and ineffective. In most cases, they are only applicable under certain conditions. Indeed, some methods do not take into account all the parameters involved in the flood modeling, and these parameters can vary along a channel, which results in obtaining forecasted discharges very different from observed discharges. While using machine learning tools, especially artificial neural networks schemes appears to be an alternative. However, the performance of forecasting models, as well as a minimum error of prediction, is very interesting and challenging issues. In this paper, we used the multilayer perceptron in order to design a flood forecasting model and used discharge as input–output variables. The designed model has been tested upon intensive experiments and the results showed the effectiveness of our proposal with a good forecasting capacity.

Published
2020

7. Towards Portable Online Prediction of Network Utilization using MPI-level Monitoring

Author

Emmanuel Jeannot, Franck Cappello, Shu Mei Tseng, Bogdan Nicolae, Aparna Chandramowlishwaran, George Bosilca, University of California [Irvine] (UCI), University of California, Argonne National Laboratory [Lemont] (ANL), The University of Tennessee [Knoxville], Topology-Aware System-Scale Data Management for High-Performance Computing (TADAAM), Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Inria Bordeaux - Sud-Ouest, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria), This research was supported by the Exascale Computing Project (17-SC-20-SC), a collaborative effort of the U.S. Department of Energy Office of Science and the National Nuclear Security Administration. This material was based upon work supported by the U.S. Department of Energy, Office of Science, under contract DE-AC02-06CH11357,and by the National Science Foundation under Grant No. #1664142. The experimentspresented in this paper were carried out using the Grid’5000/ALADDIN-G5K experimental testbed, an initiative of the French Ministry of Research through the ACI GRID incentive action, INRIA, CNRS and RENATER and other contributing partners (see http://www.grid5000.fr/)., University of California [Irvine] (UC Irvine), University of California (UC), and Université de Bordeaux (UB)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Centre National de la Recherche Scientifique (CNRS)-Université de Bordeaux (UB)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Centre National de la Recherche Scientifique (CNRS)-Inria Bordeaux - Sud-Ouest

Subjects
Artificial neural network, business.industry, Computer science, Network monitoring, Deep learning, Distributed computing, 010103 numerical & computational mathematics, 010501 environmental sciences, Prediction of resource utilization, 01 natural sciences, Timeseries forecasting, Work stealing, Online learning, Leverage (statistics), Artificial intelligence, 0101 mathematics, [INFO.INFO-DC]Computer Science [cs]/Distributed, Parallel, and Cluster Computing [cs.DC], business, 0105 earth and related environmental sciences
Abstract

International audience; Stealing network bandwidth helps a variety of HPC runtimes and services to run additional operations in the background without negatively affecting the applications. A key ingredient to make this possible is an accurate prediction of the future network utilization, enabling the runtime to plan the background operations in advance, such as to avoid competing with the application for network bandwidth. In this paper, we propose a portable deep learning predictor that only uses the information available through MPI introspection to construct a recurrent sequence-to-sequence neural network capable of forecasting network utilization. We leverage the fact that most HPC applications exhibit periodic behaviors to enable predictions far into the future (at least the length of a period). Our on-line approach does not have an initial training phase, it continuously improves itself during application execution without incurring significant computational overhead. Experimental results show better accuracy and lower computational overhead compared with the state-of-the-art on two representative applications.

Published
2019

8. CountNet: Estimating the Number of Concurrent Speakers Using Supervised Learning

Author

Bernd Edler, Emanuel A. P. Habets, Fabian-Robert Stöter, Soumitro Chakrabarty, Scientific Data Management (ZENITH), Laboratoire d'Informatique de Robotique et de Microélectronique de Montpellier (LIRMM), Centre National de la Recherche Scientifique (CNRS)-Université de Montpellier (UM)-Centre National de la Recherche Scientifique (CNRS)-Université de Montpellier (UM)-Inria Sophia Antipolis - Méditerranée (CRISAM), Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria), International Audio Laboratories Erlangen (AUDIO LABS), Friedrich-Alexander Universität Erlangen-Nürnberg (FAU)-Fraunhofer Institute for Integrated Circuits (Fraunhofer IIS), Fraunhofer (Fraunhofer-Gesellschaft)-Fraunhofer (Fraunhofer-Gesellschaft), The authors gratefully acknowledge the compute resources and support provided by the Erlangen Regional Computing Center (RRZE). They would like to thank A. Liutkus for his constructive criticism of the paper., and Université de Montpellier (UM)-Centre National de la Recherche Scientifique (CNRS)-Université de Montpellier (UM)-Centre National de la Recherche Scientifique (CNRS)-Inria Sophia Antipolis - Méditerranée (CRISAM)

Subjects
Speaker count estimation, Reverberation, cocktail-party, overlap detection, Acoustics and Ultrasonics, Artificial neural network, Computer science, Speech recognition, Supervised learning, Probabilistic logic, Blind signal separation, Speaker diarisation, 030507 speech-language pathology & audiology, 03 medical and health sciences, Computational Mathematics, [MATH.MATH-LO]Mathematics [math]/Logic [math.LO], Recurrent neural network, Computer Science (miscellaneous), number of concurrent speakers, Point estimation, Electrical and Electronic Engineering, 0305 other medical science
Abstract

International audience; Estimating the maximum number of concurrent speakers from single-channel mixtures is a challenging problem and an essential first step to address various audio-based tasks such as blind source separation, speaker diarization, and audio surveillance. We propose a unifying probabilistic paradigm, where deep neural network architectures are used to infer output posterior distributions. These probabilities are in turn processed to yield discrete point estimates. Designing such architectures often involves two important and complementary aspects that we investigate and discuss. First, we study how recent advances in deep architectures may be exploited for the task of speaker count estimation. In particular, we show that convolutional recurrent neural networks outperform recurrent networks used in a previous study when adequate input features are used. Even for short segments of speech mixtures, we can estimate up to five speakers, with a significantly lower error than other methods. Second, through comprehensive evaluation, we compare the best-performing method to several baselines, as well as the influence of gain variations, different data sets, and reverberation. The output of our proposed method is compared to human performance. Finally, we give insights into the strategy used by our proposed method.

Published
2019

9. Disentangling the attention network test: behavioral, event related potentials, and neural source analyses

Author

Guillermo Izquierdo, Manuel Vázquez-Marrufo, Alejandro Galvao-Carmona, Dolores Páramo, María Luisa Benítez, Javier J. Gonzalez-Rosa, Antonio R. Hidalgo-Muñoz, [Galvao-Carmona,A, Hidalgo-Muñoz,AR, Vázquez-Marrufo,M] Psychophysiology Unit (Lab B508), Department of Experimental Psychology, Faculty of Psychology, University of Seville, Seville, Spain. [González-Rosa,JJ] Laboratory for Clinical Neuroscience, Centre of Biomedical Technology (CTB), Technical University of Madrid (UPM), Madrid, Spain. [Páramo,D, Izquierdo,G] Multiple Sclerosis Unit, Virgen Macarena University Hospital, Seville, Spain. [Benítez,ML] Department of Physiotherapy, Faculty of Nursing, University of Seville, Seville, Spain., This study was supported by Plan Nacional I+D+i (Ministerio de Economía y Competitividad, Government of Spain, PSI2010- 16825), Asociación Neuroinvest and Asociación Sanitaria Virgen Macarena. which has greatly improved this paper, Universidad de Sevilla. Departamento de Psicología Experimental, Universidad de Sevilla. Departamento de Fisioterapia, and Ministerio de Economía y Competitividad (MINECO). España

Subjects
CNV, Neurociencias, Sensory system, source analysis, lcsh:RC321-571, Organisms::Eukaryota::Animals::Chordata::Vertebrates::Mammals::Primates::Haplorhini::Catarrhini::Hominidae::Humans [Medical Subject Headings], Behavioral Neuroscience, Psychiatry and Psychology::Psychological Phenomena and Processes::Mental Processes::Learning::Cues [Medical Subject Headings], Psychiatry and Psychology::Psychological Phenomena and Processes::Mental Processes [Medical Subject Headings], Disciplines and Occupations::Natural Science Disciplines::Biological Science Disciplines::Neurosciences [Medical Subject Headings], Event-related potential, Attention network, Ssource analysis, Psychiatry and Psychology::Psychological Phenomena and Processes::Psychophysiology::Arousal::Attention [Medical Subject Headings], Orienting, Phenomena and Processes::Physiological Phenomena::Electrophysiological Phenomena::Evoked Potentials [Medical Subject Headings], Attention, Original Research Article, P300, Source analysis, lcsh:Neurosciences. Biological psychiatry. Neuropsychiatry, Biological Psychiatry, Expectancy theory, Artificial neural network, orienting, Information processing, Psychiatry and Psychology::Behavior and Behavior Mechanisms::Adaptation, Psychological::Orientation [Medical Subject Headings], neural networks, attention, Psychiatry and Mental health, Neuropsychology and Physiological Psychology, Neurology, Potenciales evocados, alerting, Procesos mentales, Basic level, ANT, Alerting, Psychology, Neuroscience, Neural networks
Abstract

Background: The study of the attentional system remains a challenge for current neuroscience. The “Attention Network Test” (ANT) was designed to study simultaneously three different attentional networks (alerting, orienting, and executive) based in subtraction of different experimental conditions. However, some studies recommend caution with these calculations due to the interactions between the attentional networks. In particular, it is highly relevant that several interpretations about attentional impairment have arisen from these calculations in diverse pathologies. Event related potentials (ERPs) and neural source analysis can be applied to disentangle the relationships between these attentional networks not specifically shown by behavioral measures. Results: This study shows that there is a basic level of alerting (tonic alerting) in the no cue (NC) condition, represented by a slow negative trend in the ERP trace prior to the onset of the target stimuli. A progressive increase in the CNV amplitude related to the amount of information provided by the cue conditions is also shown. Neural source analysis reveals specific modulations of the CNV related to a task-related expectancy presented in the NC condition; a late modulation triggered by the central cue (CC) condition and probably representing a generic motor preparation; and an early and late modulation for spatial cue (SC) condition suggesting specific motor and sensory preactivation. Finally, the first component in the information processing of the target stimuli modulated by the interaction between orienting network and the executive system can be represented by N1. Conclusions: The ANT is useful as a paradigm to study specific attentional mechanisms and their interactions. However, calculation of network effects is based in subtractions with non-comparable experimental conditions, as evidenced by the present data, which can induce misinterpretations in the study of the attentional capacity in human subjects. Ministerio de Economía y Competitividad PSI2010- 16825

Published
2014