Search

Your search keyword '"[INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG]"' showing total 9,679 results
Start Over Descriptor "[INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG]"
9,679 results on '"[INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG]"'

1. Historical Documents and Automatic Text Recognition: Introduction

Author

Ariane Pinche and Peter Stokes

Subjects
atr, escriptorium, kraken, htr-united, segmonto, [shs.hist]humanities and social sciences/history, [info]computer science [cs], [info.info-ai]computer science [cs]/artificial intelligence [cs.ai], [info.info-lg]computer science [cs]/machine learning [cs.lg], [info.info-mo]computer science [cs]/modeling and simulation, [shs.litt]humanities and social sciences/literature, History of scholarship and learning. The humanities, AZ20-999, Bibliography. Library science. Information resources
Abstract

With this special issue of the Journal of Data Mining and Digital Humanities (JDMDH), we bringtogether in one single volume several experiments, projects and reflections related to automatic textrecognition applied to historical documents. More and more research projects now include automatic text acquisition in their data processing chain, and this is true not only for projects focussed on Digital or Computational Humanities but increasingly also for those that are simply using existing digital tools as the means to an end. The increasing use of this technology has led to an automation of tasks that affects the role of the researcher in the textual production process. This new data-intensive practice makes it urgent to collect and harmonise the corpora necessary for the constitution of training sets, but also to make them available for exploitation. This special issue is therefore an opportunity to present articles combining philological and technical questions to make a scientific assessment of the use of automatic text recognition for ancient documents, its results, its contributions and the new practices induced by its use in the process of editing and exploring texts. We hope that practical aspects will be questioned on this occasion, while raising methodological challenges and its impact on research data.The special issue on Automatic Text Recognition (ATR) is therefore dedicated to providing a comprehensive overview of the use of ATR in the humanities field, particularly concerning historical documents in the early 2020s. This issue presents a fusion of engineering and philological aspects, catering to both beginners and experienced users interested in launching projects with ATR. The collection encompasses a diverse array of approaches, covering topics such as data creation or collection for training generic models, reaching specific objectives, technical and HTR machine architecture, segmentation methods, and image processing.

Published
2024
Full Text
View/download PDF

2. Generic HTR Models for Medieval Manuscripts. The CREMMALab Project

Author

Ariane Pinche

Subjects
htr, model, dataset, medieval, text, transcription, [shs.hist]humanities and social sciences/history, [info.info-ai]computer science [cs]/artificial intelligence [cs.ai], [info.info-lg]computer science [cs]/machine learning [cs.lg], [info.info-mo]computer science [cs]/modeling and simulation, History of scholarship and learning. The humanities, AZ20-999, Bibliography. Library science. Information resources
Abstract

In the Humanities, the emergence of digital methods has opened up research questions to quantitative analysis. This is why HTR technology is increasingly involved in humanities research projects following precursors such as the Himanis project. However, many research teams have limited resources, either financially or in terms of their expertise in artificial intelligence. It may therefore be difficult to integrate handwritten text recognition into their project pipeline if they need to train a model or to create data from scratch. The goal here is not to explain how to build or improve a new HTR engine, nor to find a way to automatically align a preexisting corpus with an image to quickly create ground truths for training. This paper aims to help humanists easily develop an HTR model for medieval manuscripts, create and gather training data by knowing the issues underlying their choices. The objective is also to show the importance of the constitution of consistent data as a prerequisite to allow their gathering and to train efficient HTR models. We will present an overview of our work and experiment in the CREMMALab project (2021-2022), showing first how we ensure the consistency of the data and then how we have developed a generic model for medieval French manuscripts from the 13 th to the 15 th century, ready to be shared (more than 94% accuracy) and/or fine-tuned by other projects.

Published
2023
Full Text
View/download PDF

3. Automatic medieval charters structure detection : A Bi-LSTM linear segmentation approach

Author

Sergio Torres Aguilar, Pierre Chastang, and Xavier Tannier

Subjects
linear text segmentation, latin nlp, medieval digital studies, digital diplomatics, medieval charters, automatic structure detection, [info.info-tt]computer science [cs]/document and text processing, [shs.hist]humanities and social sciences/history, [info.info-lg]computer science [cs]/machine learning [cs.lg], History of scholarship and learning. The humanities, AZ20-999, Bibliography. Library science. Information resources
Abstract

This paper presents a model aiming to automatically detect sections in medieval Latin charters. These legal sources are some of the most important sources for medieval studies as they reflect economic and social dynamics as well as legal and institutional writing practices. An automatic linear segmentation can greatly facilitate charter indexation and speed up the recovering of evidence to support historical hypothesis by the means of granular inquiries on these raw, rarely structured sources. Our model is based on a Bi-LSTM approach using a final CRF-layer and was trained using a large, annotated collection of medieval charters (4,700 documents) coming from Lombard monasteries: the CDLM corpus (11th-12th centuries). The evaluation shows a high performance in most sections on the test-set and on an external evaluation corpus consisting of the Montecassino abbey charters (10th-12th centuries). We describe the architecture of the model, the main problems related to the treatment of medieval Latin and formulaic discourse, and we discuss some implications of the results in terms of record-keeping practices in High Middle Ages.

Published
2022
Full Text
View/download PDF

4. Private Sampling with Identifiable Cheaters

Author

Sabater, César, Hahn, Florian, Peter, Andreas, Ramon, Jan, Machine Learning in Information Networks (MAGNET), 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 Twente, University of Oldenburg, Chair TIP project, ANR-20-CE23-0013,PMR,Recherche en Medicine respectant la vie Privée(2020), European Project: 101070038,TRUMPET, and Machine Learning in Information Networks [MAGNET]

Subjects
sampling, Cybersecurity, differential privacy, zero knowledge proofs, multiparty computation, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Horizon Europe, General Earth and Planetary Sciences, Private Sampling, General Environmental Science
Abstract

International audience; In this paper we study verifiable sampling from probability distributions in the context of multi-party computation. This has various applications in randomized algorithms performed collaboratively by parties not trusting each other. One example is differentially private machine learning where noise should be drawn, typically from a Laplace or Gaussian distribution, and it is desirable that no party can bias this process. In particular, we propose algorithms to draw random numbers from uniform, Laplace, Gaussian and arbitrary probability distributions, and to verify honest execution of the protocols through zero-knowledge proofs. We propose protocols that result in one party knowing the drawn number and protocols that deliver the drawn random number as a shared secret.

Published
2023

5. Machine learning identifies a profile of inadequate responder to methotrexate in rheumatoid arthritis

Author

Julien Duquesne, Vincent Bouget, Paul Henry Cournède, Bruno Fautrel, Francis Guillemin, Pascal H P de Jong, Judith W Heutz, Marloes Verstappen, Annette H M van der Helm-van Mil, Xavier Mariette, Samuel Bitoun, Scienta Lab [Gif-sur-Yvette, France], Mathématiques et Informatique pour la Complexité et les Systèmes (MICS), CentraleSupélec-Université Paris-Saclay, CHU Pitié-Salpêtrière [AP-HP], Assistance publique - Hôpitaux de Paris (AP-HP) (AP-HP)-Sorbonne Université (SU), Pharmacoépidémiologie et évaluation des soins [iPLesp] (PEPITES), Institut Pierre Louis d'Epidémiologie et de Santé Publique (iPLESP), Institut National de la Santé et de la Recherche Médicale (INSERM)-Sorbonne Université (SU)-Institut National de la Santé et de la Recherche Médicale (INSERM)-Sorbonne Université (SU), Adaptation, mesure et évaluation en santé. Approches interdisciplinaires (APEMAC), Université de Lorraine (UL), Erasmus University Medical Center [Rotterdam] (Erasmus MC), Service de Rhumatologie [CHU Bicêtre], Assistance publique - Hôpitaux de Paris (AP-HP) (AP-HP)-Hôpital Bicêtre, Immunologie des maladies virales, auto-immunes, hématologiques et bactériennes (IMVA-HB), Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Institut National de la Santé et de la Recherche Médicale (INSERM)-Université Paris-Saclay, Leiden University Medical Center (LUMC), and Universiteit Leiden

Subjects
machine learning, Rheumatology, [SDV.MHEP.RSOA]Life Sciences [q-bio]/Human health and pathology/Rhumatology and musculoskeletal system, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [SDV.SP.PHARMA]Life Sciences [q-bio]/Pharmaceutical sciences/Pharmacology, treatment response, biomarker, Pharmacology (medical), [SDV.SPEE]Life Sciences [q-bio]/Santé publique et épidémiologie, MTX, RA
Abstract

Objectives Around 30% of patients with RA have an inadequate response to MTX. We aimed to use routine clinical and biological data to build machine learning models predicting EULAR inadequate response to MTX and to identify simple predictive biomarkers. Methods Models were trained on RA patients fulfilling the 2010 ACR/EULAR criteria from the ESPOIR and Leiden EAC cohorts to predict the EULAR response at 9 months (± 6 months). Several models were compared on the training set using the AUROC. The best model was evaluated on an external validation cohort (tREACH). The model's predictions were explained using Shapley values to extract a biomarker of inadequate response. Results We included 493 therapeutic sequences from ESPOIR, 239 from EAC and 138 from tREACH. The model selected DAS28, Lymphocytes, Creatininemia, Leucocytes, AST, ALT, swollen joint count and corticosteroid co-treatment as predictors. The model reached an AUROC of 0.72 [95% CI (0.63, 0.80)] on the external validation set, where 70% of patients were responders to MTX. Patients predicted as inadequate responders had only 38% [95% CI (20%, 58%)] chance to respond and using the algorithm to decide to initiate MTX would decrease inadequate-response rate from 30% to 23% [95% CI: (17%, 29%)]. A biomarker was identified in patients with moderate or high activity (DAS28 > 3.2): patients with a lymphocyte count superior to 2000 cells/mm3 are significantly less likely to respond. Conclusion Our study highlights the usefulness of machine learning in unveiling subgroups of inadequate responders to MTX to guide new therapeutic strategies. Further work is needed to validate this approach.

Published
2023

6. A Survey on Deep Learning Resilience Assessment Methodologies

Author

Annachiara Ruospo, Ernesto Sanchez, Lucas Matana Luza, Luigi Dilillo, Marcello Traiola, Alberto Bosio, Politecnico di Torino = Polytechnic of Turin (Polito), Test and dEpendability of microelectronic integrated SysTems (TEST), 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), Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT), Architectures matérielles spécialisées pour l’ère post loi-de-Moore (TARAN), Inria Rennes – Bretagne Atlantique, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-ARCHITECTURE (IRISA-D3), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut des Nanotechnologies de Lyon (INL), École Centrale de Lyon (ECL), Université de Lyon-Université de Lyon-Université Claude Bernard Lyon 1 (UCBL), Université de Lyon-École Supérieure de Chimie Physique Électronique de Lyon (CPE)-Institut National des Sciences Appliquées de Lyon (INSA Lyon), and Université de Lyon-Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Centre National de la Recherche Scientifique (CNRS)

Subjects
learning, Resilience, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], General Computer Science, [INFO.INFO-ES]Computer Science [cs]/Embedded Systems, Safety, [SPI.NANO]Engineering Sciences [physics]/Micro and nanotechnologies/Microelectronics, Reliability, Focusing
Abstract

International audience; Deep Learning (DL) applications are gaining increasing interest in the industry and academia for their outstanding computational capabilities. Indeed, they have found successful applications in various areas and domains such as avionics, robotics, automotive, medical wearable devices, gaming; some have been labeled as safety-critical, as system failures can compromise human life. Consequently, DL reliability is becoming a growing concern, and efficient reliability assessment approaches are required to meet safety constraints. This paper presents a survey of the main DL reliability assessment methodologies, focusing mainly on Fault Injection (FI) techniques used to evaluate the DL resilience. The article describes some of the most representative state-of-the-art academic and industrial works describing FI methodologies at different levels of abstraction. Finally, a discussion of the advantages and disadvantages of each methodology is proposed to provide valuable guidelines for carrying out safety analyses.

Published
2023

7. Secure protocols for cumulative reward maximization in stochastic multi-armed bandits

Author

Ciucanu, Radu, Lafourcade, Pascal, Lombard-Platet, Marius, Soare, Marta, Laboratoire d'Informatique de Grenoble (LIG), Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes (UGA)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP ), Université Grenoble Alpes (UGA), Laboratoire d'Informatique, de Modélisation et d'Optimisation des Systèmes (LIMOS), Ecole Nationale Supérieure des Mines de St Etienne (ENSM ST-ETIENNE)-Centre National de la Recherche Scientifique (CNRS)-Université Clermont Auvergne (UCA)-Institut national polytechnique Clermont Auvergne (INP Clermont Auvergne), Université Clermont Auvergne (UCA)-Université Clermont Auvergne (UCA), Département d'informatique - ENS Paris (DI-ENS), École normale supérieure - Paris (ENS-PSL), Université Paris sciences et lettres (PSL)-Université Paris sciences et lettres (PSL)-Institut National de Recherche en Informatique et en Automatique (Inria)-Centre National de la Recherche Scientifique (CNRS), ANR-19-P3IA-0003,MIAI,MIAI @ Grenoble Alpes(2019), European Project: H2020,INODE, and European Project: 952215,TAILOR(2020)

Subjects
[INFO.INFO-CR]Computer Science [cs]/Cryptography and Security [cs.CR], [INFO.INFO-DB]Computer Science [cs]/Databases [cs.DB], [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Computer Networks and Communications, Hardware and Architecture, Safety, Risk, Reliability and Quality, Software, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]
Abstract

International audience; We consider the problem of cumulative reward maximization in multi-armed bandits. We address the security concerns that occur when data and computations are outsourced to an honest-but-curious cloud i.e., that executes tasks dutifully, but tries to gain as much information as possible. We consider situations where data used in bandit algorithms is sensitive and has to be protected e.g., commercial or personal data. We rely on cryptographic schemes and propose UCB - MS, a secure multi-party protocol based on the UCB algorithm. We prove that UCB - MS computes the same cumulative reward as UCB while satisfying desirable security properties. In particular, cloud nodes cannot learn the cumulative reward or the sum of rewards for more than one arm. Moreover, by analyzing messages exchanged among cloud nodes, an external observer cannot learn the cumulative reward or the sum of rewards produced by some arm. We show that the overhead due to cryptographic primitives is linear in the size of the input. Our implementation confirms the linear-time behavior and the practical feasibility of our protocol, on both synthetic and real-world data.

Published
2023

8. Indigenous frameworks for data-intensive humanities: recalibrating the past through knowledge engineering and generative modelling.

Author

Sydney Shep, Marcus Frean, Rhys Owen, Rere-No-A-Rangi Pope, Pikihuia Reihana, and Valerie Chan

Subjects
indigenous knowledge, semantic web, generative modelling, bayesian record linkage, network analysis, [info.info-lg]computer science [cs]/machine learning [cs.lg], [shs.hist]humanities and social sciences/history, [shs.museo]humanities and social sciences/cultural heritage and museology, History of scholarship and learning. The humanities, AZ20-999, Bibliography. Library science. Information resources
Abstract

Identifying, contacting and engaging missing shareholders constitutes an enormous challenge for Māori incorporations, iwi and hapū across Aotearoa New Zealand. Without accurate data or tools to har-monise existing fragmented or conflicting data sources, issues around land succession, opportunities for economic development, and maintenance of whānau relationships are all negatively impacted. This unique three-way research collaboration between Victoria University of Wellington (VUW), Parininihi ki Waitotara Incorporation (PKW), and University of Auckland funded by the National Science Challenge | Science for Technological Innovation catalyses innovation through new digital humanities-inflected data science modelling and analytics with the kaupapa of reconnecting missing Māori shareholders for a prosperous economic, cultural, and socially revitalised future. This paper provides an overview of VUW's culturally-embedded social network approach to the project, discusses the challenges of working within an indigenous worldview, and emphasises the importance of decolonising digital humanities.

Published
2021
Full Text
View/download PDF

9. Digital interfaces of historical newspapers: opportunities, restrictions and recommendations

Author

Eva Pfanzelter, Sarah Oberbichler, Jani Marjanen, Pierre-Carl Langlais, and Stefan Hechl

Subjects
digital humanities, frequencies, topic modelling, historical newspapers, interfaces, digital newspapers, [shs.hist]humanities and social sciences/history, [info.info-dl]computer science [cs]/digital libraries [cs.dl], [info.info-lg]computer science [cs]/machine learning [cs.lg], [shs]humanities and social sciences, [shs.stat]humanities and social sciences/methods and statistics, History of scholarship and learning. The humanities, AZ20-999, Bibliography. Library science. Information resources
Abstract

Many libraries offer free access to digitised historical newspapers via user interfaces. After an initial period of search and filter options as the only features, the availability of more advanced tools and the desire for more options among users has ushered in a period of interface development. However, this raises a number of open questions and challenges. For example, how can we provide interfaces for different user groups? What tools should be available on interfaces and how can we avoid too much complexity? What tools are helpful and how can we improve usability? This paper will not provide definite answers to these questions, but it gives an insight into the difficulties, challenges and risks of using interfaces to investigate historical newspapers. More importantly, it provides ideas and recommendations for the improvement of user interfaces and digital tools.

Published
2021
Full Text
View/download PDF

10. Computed Tomography Image Restoration Using a Quantum-Based Deep Unrolled Denoiser and a Plug-and-Play Framework

Author

Dutta, Sayantan, Nwigbo, Kenule Tuador, MICHETTI, Jerome, Georgeot, Bertrand, Pham, Duong-Hung, Kouamé, Denis, Basarab, Adrian, Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT), CoMputational imagINg anD viSion (IRIT-MINDS), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université de Toulouse (UT)-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université Toulouse Capitole (UT Capitole), Laboratoire de Physique Théorique (LPT), Université de Toulouse (UT)-Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Fédération de recherche « Matière et interactions » (FeRMI), Institut National des Sciences Appliquées - Toulouse (INSA Toulouse), Institut National des Sciences Appliquées (INSA)-Université de Toulouse (UT)-Institut National des Sciences Appliquées (INSA)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National des Sciences Appliquées - Toulouse (INSA Toulouse), Institut National des Sciences Appliquées (INSA)-Université de Toulouse (UT)-Institut National des Sciences Appliquées (INSA)-Centre National de la Recherche Scientifique (CNRS), Imagerie Ultrasonore, Centre de Recherche en Acquisition et Traitement de l'Image pour la Santé (CREATIS), Université Claude Bernard Lyon 1 (UCBL), Université de Lyon-Université de Lyon-Institut National des Sciences Appliquées de Lyon (INSA Lyon), Université de Lyon-Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université Jean Monnet - Saint-Étienne (UJM)-Institut National de la Santé et de la Recherche Médicale (INSERM)-Centre National de la Recherche Scientifique (CNRS)-Université Claude Bernard Lyon 1 (UCBL), Université de Lyon-Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université Jean Monnet - Saint-Étienne (UJM)-Institut National de la Santé et de la Recherche Médicale (INSERM)-Centre National de la Recherche Scientifique (CNRS), and Aalto University, Finland

Subjects
Plug-and-Play, Unrolling, Quantum image processing, Super-Resolution, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, Quantum denoising, [INFO.INFO-IM]Computer Science [cs]/Medical Imaging, Deep learning, DIVA, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]
Abstract

International audience; In this work, we address the problem of cone beam computed tomography (CBCT) image resolution enhancement by exploiting a newly introduced deep unrolled quantum denoiser, based on quantum interaction theory adapted to computational imaging. Following recent advances in image restoration using the Plug-and-Play (PnP) framework, we impose this external deep learning denoiser as a regularizer within the super-resolution (SR) problem. The quantum-based deep unrolled denoiser combined with a computationally efficient way to deal with the degradation operators, and the PnP scheme, result in an original way of approaching the image resolution enhancement problem. Experiments conducted on dental CBCT images are presented to illustrate the efficiency of the proposed model for image resolution enhancement tasks. The numerical results suggest that the proposed method allows better restoration performances compared to existing state-of-the-art approaches.

Published
2023

11. Learning a Riemannian manifold for the analysis-synthesis of nonstationary sounds

Author

Han, Han, Lostanlen, Vincent, Lagrange, Mathieu, Laboratoire des Sciences du Numérique de Nantes (LS2N), Institut National de Recherche en Informatique et en Automatique (Inria)-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-École Centrale de Nantes (Nantes Univ - ECN), Nantes Université (Nantes Univ)-Nantes Université (Nantes Univ)-Nantes université - UFR des Sciences et des Techniques (Nantes univ - UFR ST), Nantes Université - pôle Sciences et technologie, Nantes Université (Nantes Univ)-Nantes Université (Nantes Univ)-Nantes Université - pôle Sciences et technologie, Nantes Université (Nantes Univ), Signal, IMage et Son (LS2N - équipe SIMS ), and Nantes Université (Nantes Univ)-Institut National de Recherche en Informatique et en Automatique (Inria)-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique)

Subjects
[SPI.ACOU]Engineering Sciences [physics]/Acoustics [physics.class-ph], [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [INFO.INFO-SD]Computer Science [cs]/Sound [cs.SD]
Abstract

International audience; Computer sound matching poses an inverse problem, namely, the identification of resynthesis parameters. Borrowing from the differentiable digital signal processing (DDSP) framework, we propose to automate its resolution by training a deep neural network. In this context, we aim to reach a compromise between the computational efficiency of parametric loss (P-loss) versus the psychoacoustical fidelity of spectral loss. Our approach, named ``perceptual--neural--physical'' (PNP), estimates the Riemannian metric which is associated to the composition between parametric synthesis and time--frequency scattering. By doing so, we locally linearize spectral loss and accelerate convergence. Furthermore, resorting to Tikhonov regularization improves the conditioning of the inverse problem. On an analysis--synthesis task for musical arpeggios, PNP training outperforms state-of-the-art methods P-loss (wav2shape) and STFT-based DDSP, as measured in terms of JTFS-based similarity between reference signal and reconstructed signal.; La transformation de sons par ordinateur pose un problème inverse d'identification des paramètres de resynthèse adéquats. Empruntant au formalisme du traitement du signal différentiable (DDSP), nous proposons d'automatiser sa résolution par entrainement d'un réseau de neurones profond. Dans ce contexte, nous visons un compromis entre l'efficacité computationnelle de la perte paramétrique et la fidélité psychoacoustique de la perte spectrale. Notre approche, baptisée perceptuelle-neuronalephysique (PNP), consiste à estimer la métrique riemannienne associée à la composition entre synthèse paramétrique et diffusion temps-fréquence (JTFS). Ce faisant, nous linéarisons localement la perte spectrale et accélérons la convergence. De plus, le recours à une régularisation de Tikhonov améliore le conditionnement du problème inverse. Par rapport à l'état de l'art (wav2shape et DDSP), et pour une tâche difficile d'analyse-synthèse d'arpège musical, l'entrainement via PNP rapproche le signal reconstruit du signal de référence, d'après une mesure de similarité de timbre fondée sur la JTFS.

Published
2023

12. How to (Virtually) Train Your Speaker Localizer

Author

Srivastava, Prerak, Deleforge, Antoine, Politis, Archontis, Vincent, Emmanuel, 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), Institut National de Recherche en Informatique et en Automatique (Inria), and University of Tampere [Finland]

Subjects
FOS: Computer and information sciences, Sound (cs.SD), Computer Science - Sound, localization, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], image source, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Audio and Speech Processing (eess.AS), [INFO.INFO-SD]Computer Science [cs]/Sound [cs.SD], FOS: Electrical engineering, electronic engineering, information engineering, room acoustic simulation, directivity, direction-of-arrival, Electrical Engineering and Systems Science - Audio and Speech Processing
Abstract

Learning-based methods have become ubiquitous in speaker localization. Existing systems rely on simulated training sets for the lack of sufficiently large, diverse and annotated real datasets. Most room acoustics simulators used for this purpose rely on the image source method (ISM) because of its computational efficiency. This paper argues that carefully extending the ISM to incorporate more realistic surface, source and microphone responses into training sets can significantly boost the real-world performance of speaker localization systems. It is shown that increasing the training-set realism of a state-of-the-art direction-of-arrival estimator yields consistent improvements across three different real test sets featuring human speakers in a variety of rooms and various microphone arrays. An ablation study further reveals that every added layer of realism contributes positively to these improvements., Comment: Published in INTERSPEECH 2023

Published
2023

13. Learning meshless parameterization with graph convolutional neural networks

Author

Giannelli, Carlotta, Imperatore, Sofia, Mantzaflaris, Angelos, Scholz, Felix, Dipartimento di Matematica e Informatica 'Ulisse Dini' [Firenze] (DIMAI UniFI), Università degli Studi di Firenze = University of Florence (UniFI), COMUE Université Côte d'Azur (2015-2019) (COMUE UCA), AlgebRe, geOmetrie, Modelisation et AlgoriTHmes (AROMATH), 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)-National and Kapodistrian University of Athens (NKUA), Institute of Applied Geometry [Linz], Johannes Kepler Universität Linz (JKU), PHC GALILEE project 47786N, and European Project: 860843,H2020-EU.1.3. - EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions,GRAPES(2019)

Subjects
meshless parameterization, geometric deep learning, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], surface reconstruction, least squares fitting, graph convolutional neural networks, [INFO.INFO-IA]Computer Science [cs]/Computer Aided Engineering
Abstract

International audience; This paper proposes a deep learning approach for parameterizing an unorganized or scattered point cloud in R 3 with graph convolutional neural networks. It builds upon a graph convolutional neural network that predicts the weights (called parameterization weights) of certain convex combinations that lead to a mapping of the 3D points into a planar parameter domain. First, we compute a radius neighbours graph that yields proximity information to each 3D point in the cloud. This radius graph is then converted to its line graph, which encodes edge adjacencies, and is equipped with appropriate weights. The line graph is used as input to a graph convolutional neural network trained to predict optimal parameterizations. The proposed model outperforms closed-form choices of the parameterization weights and produces high quality parameterizations for surface reconstruction schemes.

Published
2023

14. MEG Encoding using Word Context Semantics in Listening Stories

Author

Reddy Oota, Subba, Trouvain, Nathan, Alexandre, Frederic, Hinaut, Xavier, Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université de Bordeaux (UB)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Centre National de la Recherche Scientifique (CNRS), Institut des Maladies Neurodégénératives [Bordeaux] (IMN), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS), Mnemonic Synergy (Mnemosyne), 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, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Institut des Maladies Neurodégénératives [Bordeaux] (IMN), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-Centre National de la Recherche Scientifique (CNRS), and INTERSPEECH

Subjects
MEG, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], human-computer interaction, [SCCO.NEUR]Cognitive science/Neuroscience, context length, [SDV.NEU]Life Sciences [q-bio]/Neurons and Cognition [q-bio.NC], [SCCO.LING]Cognitive science/Linguistics, brain encoding, syntax, semantics, [INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL], [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]
Abstract

International audience; Brain encoding is the process of mapping stimuli to brain activity. There is a vast literature on linguistic brain encoding for functional MRI (fMRI) related to syntactic and semantic representations. Magnetoencephalography (MEG), with higher temporal resolution than fMRI, enables us to look more precisely at the timing of linguistic feature processing. Unlike MEG decoding, few studies on MEG encoding using natural stimuli exist. Existing ones on story listening focus on phoneme and simple word-based features, ignoring more abstract features such as context, syntactic and semantic aspects. Inspired by previous fMRI studies, we study MEG brain encoding using basic syntactic and semantic features, with various context lengths and directions (past vs. future), for a dataset of 8 subjects listening to stories. We find that BERT representations predict MEG significantly but not other syntactic features or word embeddings (e.g. GloVe), allowing us to encode MEG in a distributed way across auditory and language regions in time. In particular, past context is crucial in obtaining significant results.

Published
2023

15. Unsupervised speech enhancement with deep dynamical generative speech and noise models

Author

Lin, Xiaoyu, Leglaive, Simon, Girin, Laurent, Alameda-Pineda, Xavier, Vers des robots à l’intelligence sociale au travers de l’apprentissage, de la perception et de la commande (ROBOTLEARN), Inria Grenoble - Rhône-Alpes, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Université Grenoble Alpes (UGA), Institut d'Électronique et des Technologies du numéRique (IETR), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-Nantes Université - pôle Sciences et technologie, Nantes Université (Nantes Univ)-Nantes Université (Nantes Univ), GIPSA - Cognitive Robotics, Interactive Systems, & Speech Processing (GIPSA-CRISSP), GIPSA Pôle Parole et Cognition (GIPSA-PPC), Grenoble Images Parole Signal Automatique (GIPSA-lab), Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes (UGA)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP ), Université Grenoble Alpes (UGA)-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes (UGA)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP ), Université Grenoble Alpes (UGA)-Grenoble Images Parole Signal Automatique (GIPSA-lab), Université Grenoble Alpes (UGA), ANR-19-P3IA-0003,MIAI,MIAI @ Grenoble Alpes(2019), ANR-19-CE33-0008,ML3RI,Apprentissage de bas-niveau d'ineractions robotiques multi-modales avec plusieurs personnes(2019), and European Project: 871245,H2020-EU.2.1.1. - INDUSTRIAL LEADERSHIP - Leadership in enabling and industrial technologies - Information and Communication Technologies (ICT),SPRING(2020)

Subjects
FOS: Computer and information sciences, Sound (cs.SD), Computer Science - Machine Learning, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Audio and Speech Processing (eess.AS), FOS: Electrical engineering, electronic engineering, information engineering, Computer Science - Sound, Machine Learning (cs.LG), Electrical Engineering and Systems Science - Audio and Speech Processing, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]
Abstract

International audience; This work builds on a previous work on unsupervised speech enhancement using a dynamical variational autoencoder (DVAE) as the clean speech model and non-negative matrix factorization (NMF) as the noise model. We propose to replace the NMF noise model with a deep dynamical generative model (DDGM) depending either on the DVAE latent variables, or on the noisy observations, or on both. This DDGM can be trained in three configurations: noise-agnostic, noise-dependent and noise adaptation after noise-dependent training. Experimental results show that the proposed method achieves competitive performance compared to state-of-the-art unsupervised speech enhancement methods, while the noise-dependent training configuration yields a much more time-efficient inference process.

Published
2023

16. Speech Taskonomy: Which Speech Tasks are the most Predictive of fMRI Brain Activity?

Author

Reddy Oota, Subba, Veeral, Agarwal, Mounika, Marreddy, Manish, Gupta, Bapi, Raju Surampudi, Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université de Bordeaux (UB)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Centre National de la Recherche Scientifique (CNRS), Mnemonic Synergy (Mnemosyne), 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, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Institut des Maladies Neurodégénératives [Bordeaux] (IMN), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-Centre National de la Recherche Scientifique (CNRS), International Institute of Information Technology, Hyderabad [Hyderabad] (IIIT-H), and Microsoft Research (MSR)

Subjects
computational paralinguistics, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], human-computer interaction, [SCCO.NEUR]Cognitive science/Neuroscience, FMRI = functional MRI, speech recognition, Speech Downstream Tasks, Brain Encoding, [INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL], Transfomers, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]
Abstract

International audience; Self-supervised speech based models have been found to be successful in predicting brain recordings of subjects experiencing naturalistic story listening. Inspired by the recent progress on deep learning models for various speech-processing tasks, existing literature has leveraged pretrained speech Transformer models for brain encoding. However, there is no work on exploring the efficacy of task-specific finetuned Transformer representations for this task. Hence, in this paper, we explore transfer learning from representations finetuned for eight different tasks from Speech processing Universal PERformance Benchmark (SUPERB) for predicting brain responses. Encoding models based on task features are used to predict activity in different regions across the whole brain, and also in language and auditory brain regions. Our experiments on finetuning the Wav2Vec2.0 model for these eight tasks show that the model finetuned on automatic speech recognition (ASR) yields the best encoding performance for the whole brain, language and auditory regions.

Published
2023

17. GAP: Differentially Private Graph Neural Networks with Aggregation Perturbation

Author

Sajadmanesh, Sina, Shahin Shamsabadi, Ali, Bellet, Aurélien, Gatica-Perez, Daniel, IDIAP Research Institute, Ecole Polytechnique Fédérale de Lausanne (EPFL), The Alan Turing Institute, Machine Learning in Information Networks (MAGNET), 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), ANR-20-CE23-0015,PRIDE,Apprentissage automatique décentralisé et préservant la vie privée(2020), Ecole Polytechnique Fédérale de Lausanne [EPFL], and Machine Learning in Information Networks [MAGNET]

Subjects
FOS: Computer and information sciences, Computer Science - Machine Learning, Computer Science - Cryptography and Security, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [STAT.ML]Statistics [stat]/Machine Learning [stat.ML], Cryptography and Security (cs.CR), Machine Learning (cs.LG)
Abstract

In this paper, we study the problem of learning Graph Neural Networks (GNNs) with Differential Privacy (DP). We propose a novel differentially private GNN based on Aggregation Perturbation (GAP), which adds stochastic noise to the GNN's aggregation function to statistically obfuscate the presence of a single edge (edge-level privacy) or a single node and all its adjacent edges (node-level privacy). Tailored to the specifics of private learning, GAP's new architecture is composed of three separate modules: (i) the encoder module, where we learn private node embeddings without relying on the edge information; (ii) the aggregation module, where we compute noisy aggregated node embeddings based on the graph structure; and (iii) the classification module, where we train a neural network on the private aggregations for node classification without further querying the graph edges. GAP's major advantage over previous approaches is that it can benefit from multi-hop neighborhood aggregations, and guarantees both edge-level and node-level DP not only for training, but also at inference with no additional costs beyond the training's privacy budget. We analyze GAP's formal privacy guarantees using R\'enyi DP and conduct empirical experiments over three real-world graph datasets. We demonstrate that GAP offers significantly better accuracy-privacy trade-offs than state-of-the-art DP-GNN approaches and naive MLP-based baselines. Our code is publicly available at https://github.com/sisaman/GAP., Comment: Accepted at USENIX Security '23

Published
2023
Full Text
View/download PDF

18. Multiobjective optimization with a quadratic surrogate-assisted CMA-ES

Author

Gharafi, Mohamed, Hansen, Nikolaus, Brockhoff, Dimo, Le Riche, Rodolphe, Randomized Optimisation (RANDOPT ), Centre de Mathématiques Appliquées - Ecole Polytechnique (CMAP), École polytechnique (X)-Centre National de la Recherche Scientifique (CNRS)-École polytechnique (X)-Centre National de la Recherche Scientifique (CNRS)-Inria Saclay - Ile de France, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria), Ecole Nationale Supérieure des Mines de St Etienne (ENSM ST-ETIENNE), Centre National de la Recherche Scientifique (CNRS), GECCO, Laboratoire d'Informatique, de Modélisation et d'Optimisation des Systèmes (LIMOS), Ecole Nationale Supérieure des Mines de St Etienne (ENSM ST-ETIENNE)-Université Clermont Auvergne [2017-2020] (UCA [2017-2020])-Centre National de la Recherche Scientifique (CNRS), and Université Clermont Auvergne [2017-2020] (UCA [2017-2020])

Subjects
multiobjective optimization CMA-ES evolution strategies surrogateassisted optimization quadratic metamodel, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], evolution strategies, multiobjective optimization, surrogateassisted optimization, CMA-ES, quadratic metamodel, [MATH.MATH-OC]Mathematics [math]/Optimization and Control [math.OC], [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE]
Abstract

International audience; We present a surrogate-assisted multiobjective optimization algorithm. The aggregation of the objectives relies on the Uncrowded Hypervolume Improvement (UHVI) which is partly replaced by a linear-quadratic surrogate that is integrated into the CMA-ES algorithm. Surrogating the UHVI poses two challenges. First, the UHVI is a dynamic function, changing with the empirical Pareto set. Second, it is a composite function, defined differently for dominated and nondominated points. The presented algorithm is thought to be used with expensive functions of moderate dimension (up to about 50) with a quadratic surrogate which is updated based on its ranking ability. We report numerical experiments which include tests on the COCO benchmark. The algorithm shows in particular linear convergence on the double sphere function with a convergence rate that is 6-20 times faster than without surrogate assistance.

Published
2023

19. On the Tacit Linearity Assumption in Common Cascaded Models of RIS-Parametrized Wireless Channels

Author

Rabault, Antonin, Magoarou, Luc Le, Sol, Jérôme, Alexandropoulos, George C., Shlezinger, Nir, Poor, H. Vincent, del Hougne, Philipp, Institut d'Électronique et des Technologies du numéRique (IETR), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-Nantes Université - pôle Sciences et technologie, Nantes Université (Nantes Univ)-Nantes Université (Nantes Univ), Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA), National and Kapodistrian University of Athens (NKUA), Ben-Gurion University of the Negev (BGU), Department of Electrical and Computer Engineering [Princeton] (ECE), and Princeton University

Subjects
Signal Processing (eess.SP), FOS: Computer and information sciences, linearity, Computer Science - Information Theory, Information Theory (cs.IT), end-to-end channel modeling, FOS: Physical sciences, Physics - Applied Physics, Applied Physics (physics.app-ph), Born series, discrete dipole approximation, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], FOS: Electrical engineering, electronic engineering, information engineering, Reconfigurable intelligent surfaces, fading channels, Electrical Engineering and Systems Science - Signal Processing, Reconfigurable intelligent surfaces end-to-end channel modeling fading channels discrete dipole approximation Born series linearity, [SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processing
Abstract

We analytically derive from first physical principles the functional dependence of wireless channels on the RIS configuration for generic (i.e., potentially complex-scattering) RIS-parametrized radio environments. The wireless channel is a linear input-output relation that depends non-linearly on the RIS configuration because of two independent mechanisms: i) proximity-induced mutual coupling between close-by RIS elements; ii) reverberation-induced long-range coupling between all RIS elements. Mathematically, this "structural" non-linearity originates from the inversion of an "interaction" matrix that can be cast as the sum of an infinite Born series [for i)] or Born-like series [for ii)] whose $K$th term physically represents paths involving $K$ bounces between the RIS elements [for i)] or wireless entities [for ii)]. We identify the key physical parameters that determine whether these series can be truncated after the first and second term, respectively, as tacitly done in common cascaded models of RIS-parametrized wireless channels. Numerical results obtained with the physics-compliant PhysFad model and experimental results obtained with a RIS prototype in an anechoic (echo-free) chamber and rich-scattering reverberation chambers corroborate our analysis. Our findings raise doubts about the reliability of existing performance analysis and channel-estimation protocols for cases in which cascaded models poorly describe the physical reality., Comment: 30 pages, 5 figures, submitted to an IEEE Journal

Published
2023

20. How does the brain process syntactic structure while listening?

Author

Reddy Oota, Subba, Marreddy, Mounika, Gupta, Manish, Bapi, Raju Surampudi, Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université de Bordeaux (UB)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Centre National de la Recherche Scientifique (CNRS), Mnemonic Synergy (Mnemosyne), 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, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Institut des Maladies Neurodégénératives [Bordeaux] (IMN), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-Centre National de la Recherche Scientifique (CNRS), International Institute of Information Technology, Hyderabad [Hyderabad] (IIIT-H), Microsoft Research (MSR), and Association for Computational Linguistics

Subjects
[SCCO.NEUR]Cognitive science/Neuroscience, Incremental Top-Down Parsing, Constituency parsers, Language Regions, [SCCO.LING]Cognitive science/Linguistics, Brain Encoding, [INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL], Semantics, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Transformers, Dependency Parsing, FMRI = functional MRI, Syntax
Abstract

International audience; Syntactic parsing is the task of assigning a syntactic structure to a sentence. There are two popular syntactic parsing methods: constituency and dependency parsing. Recent works have used syntactic embeddings based on constituency trees, incremental top-down parsing, and other word syntactic features for brain activity prediction given the text stimuli to study how the syntax structure is represented in the brain's language network. However, the effectiveness of dependency parse trees or the relative predictive power of the various syntax parsers across brain areas, especially for the listening task, is yet unexplored. In this study, we investigate the predictive power of the brain encoding models in three settings: (i) individual performance of the constituency and dependency syntactic parsing based embedding methods, (ii) efficacy of these syntactic parsing based embedding methods when controlling for basic syntactic signals, (iii) relative effectiveness of each of the syntactic embedding methods when controlling for the other. Further, we explore the relative importance of syntactic information (from these syntactic embedding methods) versus semantic information using BERT embeddings. We find that constituency parsers help explain activations in the temporal lobe and middle-frontal gyrus, while dependency parsers better encode syntactic structure in the angular gyrus and posterior cingulate cortex. Although semantic signals from BERT are more effective compared to any of the syntactic features or embedding methods, syntactic embedding methods explain additional variance for a few brain regions.

Published
2023

21. Continuous–Discrete Time Neural Network Observer for Nonlinear Dynamic Systems Application to Vehicle Systems

Author

Ghani, Hasan Abdl, Dam, Quang Truc, Laghmara, Hind, Ahmed-Ali, Sofiane, Ainouz, Samia, Gao, Xing, Institut de Recherche en Systèmes Electroniques Embarqués (IRSEEM), Université de Rouen Normandie (UNIROUEN), Normandie Université (NU)-Normandie Université (NU)-École Supérieure d’Ingénieurs en Génie Électrique (ESIGELEC), Laboratoire d'Informatique, du Traitement de l'Information et des Systèmes (LITIS), Université Le Havre Normandie (ULH), Normandie Université (NU)-Normandie Université (NU)-Université de Rouen Normandie (UNIROUEN), Normandie Université (NU)-Institut national des sciences appliquées Rouen Normandie (INSA Rouen Normandie), Institut National des Sciences Appliquées (INSA)-Normandie Université (NU)-Institut National des Sciences Appliquées (INSA), Informatique, BioInformatique, Systèmes Complexes (IBISC), and Université d'Évry-Val-d'Essonne (UEVE)-Université Paris-Saclay

Subjects
Nonlinear observer, Backpropagation algorithm, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Vehicle systems, Sampled data neural network, [SPI.AUTO]Engineering Sciences [physics]/Automatic
Abstract

International audience; This paper proposes a novel continuous-discrete (sampled data) time neural network (NSNN) observer for nonlinear systems. It can therefore be applied to systems with a high degree of non-linearity with no prior knowledge of the system dynamics. The proposed observer is a three-layer feedforward neural network that has been intensively trained using the error backpropagation learning algorithm, which includes an e-modification term to ensure robustness of the observer. A structure of the output predictor with a corrective term is added in the structure of the NN observer to overcome the problem of discrete time measurement. Simulations using MATLAB and CarSim are illustrated to demonstrate the performance of the proposed state observer strategy to reconstruct the state variables and parameters of a vehicle system.

Published
2023

22. Feasibility study on using house-tree-person drawings for automatic analysis of depression

Author

Zhang, Jie, Yu, Yaoxiang, Barra, Vincent, Ruan, Xiaoming, Chen, Yu, Cai, Bo, School of Computer Engineering, Nanyang Technological University [Singapour], Laboratoire d'Informatique, de Modélisation et d'Optimisation des Systèmes (LIMOS), Ecole Nationale Supérieure des Mines de St Etienne (ENSM ST-ETIENNE)-Université Clermont Auvergne [2017-2020] (UCA [2017-2020])-Centre National de la Recherche Scientifique (CNRS), and University of South Florida [Tampa] (USF)

Subjects
[INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing
Published
2023

23. A New Remote Hyperspectral Imaging System Embedded on an Unmanned Aquatic Drone for the Detection and Identification of Floating Plastic Litter Using Machine Learning

Author

Alboody, Ahed, Vandenbroucke, Nicolas, Porebski, Alice, Sawan, Rosa, Viudes, Florence, Doyen, Perine, Amara, Rachid, Laboratoire d'Informatique Signal et Image de la Côte d'Opale (LISIC), Université du Littoral Côte d'Opale (ULCO), Laboratoire d’Océanologie et de Géosciences (LOG) - UMR 8187 (LOG), Institut national des sciences de l'Univers (INSU - CNRS)-Université du Littoral Côte d'Opale (ULCO)-Université de Lille-Centre National de la Recherche Scientifique (CNRS)-Institut de Recherche pour le Développement (IRD [France-Nord]), Biochimie des Produits Aquatiques (BPA), Université du Littoral Côte d'Opale (ULCO)-Transfrontalière BioEcoAgro - UMR 1158 (BioEcoAgro), Université d'Artois (UA)-Université de Liège-Université de Picardie Jules Verne (UPJV)-Université du Littoral Côte d'Opale (ULCO)-Université de Lille-Institut National de Recherche pour l’Agriculture, l’Alimentation et l’Environnement (INRAE)-JUNIA (JUNIA), Université catholique de Lille (UCL)-Université catholique de Lille (UCL)-Université d'Artois (UA)-Université de Liège-Université de Picardie Jules Verne (UPJV)-Université du Littoral Côte d'Opale (ULCO)-Université de Lille-Institut National de Recherche pour l’Agriculture, l’Alimentation et l’Environnement (INRAE)-JUNIA (JUNIA), Université catholique de Lille (UCL)-Université catholique de Lille (UCL), This work was financially supported by the European Union, the European Regional Development Fund (ERDF), the French State, and the French Region Hauts-de-France and Ifremer, in the framework of the project CPER IDEAL 2021–2027., and ANR-21-EXES-0011,IFSEA,Transdisciplinary graduate school for marIne, Fisheries and SEAfood sciences(2021)

Subjects
unmanned aquatic drone, remote sensing, plastic pollution, machine learning, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], hyperspectral imaging, [SDE.IE]Environmental Sciences/Environmental Engineering, [INFO.INFO-TI]Computer Science [cs]/Image Processing [eess.IV], [INFO.INFO-ES]Computer Science [cs]/Embedded Systems, spectral reflectance, plastic litter identification, [SDU.STU.OC]Sciences of the Universe [physics]/Earth Sciences/Oceanography
Abstract

International audience; This paper presents a new Remote Hyperspectral Imaging System (RHIS) embedded on an Unmanned Aquatic Drone (UAD) for plastic detection and identification in coastal and freshwater environments. This original system, namely the Remotely Operated Vehicle of the University of Littoral Côte d’Opale (ROV-ULCO), works in a near-field of view, where the distance between the hyperspectral camera and the water surface is about 45 cm. In this paper, the new ROV-ULCO system with all its components is firstly presented. Then, a hyperspectral image database of plastic litter acquired with this system is described. This database contains hyperspectral data cubes of different plastic types and polymers corresponding to the most-common plastic litter items found in aquatic environments. An in situ spectral analysis was conducted from this benchmark database to characterize the hyperspectral reflectance of these items in order to identify the absorption feature wavelengths for each type of plastic. Finally, the ability of our original system RHIS to automatically recognize different types of plastic litter was assessed by applying different supervised machine learning methods on a set of representative image patches of marine litter. The obtained results highlighted the plastic litter classification capability with an overall accuracy close to 90%. This paper showed that the newly presented RHIS coupled with the UAD is a promising approach to identify plastic waste in aquatic environments.

Published
2023

24. DITAN: A deep-learning domain agnostic framework for Detection and Interpretation of Temporally-based multivariate ANomalies

Author

Giannoulis, Michail, Harris, Andrew, Barra, Vincent, Laboratoire d'Informatique, de Modélisation et d'Optimisation des Systèmes (LIMOS), Ecole Nationale Supérieure des Mines de St Etienne (ENSM ST-ETIENNE)-Centre National de la Recherche Scientifique (CNRS)-Université Clermont Auvergne (UCA)-Institut national polytechnique Clermont Auvergne (INP Clermont Auvergne), Université Clermont Auvergne (UCA)-Université Clermont Auvergne (UCA), Laboratoire Magmas et Volcans (LMV), Institut national des sciences de l'Univers (INSU - CNRS)-Institut de Recherche pour le Développement et la société-Centre National de la Recherche Scientifique (CNRS)-Université Clermont Auvergne (UCA)-Observatoire de Physique du Globe de Clermont-Ferrand (OPGC), Institut national des sciences de l'Univers (INSU - CNRS)-Centre National de la Recherche Scientifique (CNRS)-Université Clermont Auvergne (UCA)-Institut national des sciences de l'Univers (INSU - CNRS)-Centre National de la Recherche Scientifique (CNRS)-Université Clermont Auvergne (UCA), and Ecole Nationale Supérieure des Mines de St Etienne (ENSM ST-ETIENNE)-Université Clermont Auvergne [2017-2020] (UCA [2017-2020])-Centre National de la Recherche Scientifique (CNRS)

Subjects
[INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing
Abstract

International audience

Published
2023

25. Anytime Performance Assessment in Blackbox Optimization Benchmarking

Author

Nikolaus Hansen, Anne Auger, Dimo Brockhoff, Tea Tusar, Randomized Optimisation (RANDOPT ), Centre de Mathématiques Appliquées - Ecole Polytechnique (CMAP), École polytechnique (X)-Centre National de la Recherche Scientifique (CNRS)-École polytechnique (X)-Centre National de la Recherche Scientifique (CNRS)-Inria Saclay - Ile de France, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria), École polytechnique (X)-Centre National de la Recherche Scientifique (CNRS), Jozef Stefan Institute [Ljubljana] (IJS), Slovenian Research Agency Grant P2-0209, Slovenian Research Agency Grant N2-0254, and ANR-12-MONU-0009,NumBBO,Analyse, Amelioration, Evaluation d'algorithmes numériques pour l'optimisation boîte-noire(2012)

Subjects
blackbox optimization, [INFO.INFO-CE]Computer Science [cs]/Computational Engineering, Finance, and Science [cs.CE], [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], [INFO.INFO-NA]Computer Science [cs]/Numerical Analysis [cs.NA], Theoretical Computer Science, [INFO.INFO-PF]Computer Science [cs]/Performance [cs.PF], [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Computational Theory and Mathematics, anytime optimization, quality indicator, performance assessment, benchmarking, Software, [INFO.INFO-MS]Computer Science [cs]/Mathematical Software [cs.MS]
Abstract

open access; International audience; We present concepts and recipes for the anytime performance assessment when benchmarking optimization algorithms in a blackbox scenario. We consider runtime-oftentimes measured in number of blackbox evaluations needed to reach a target quality-to be a universally measurable cost for solving a problem. Starting from the graph that depicts the solution quality versus runtime, we argue that runtime is the only performance measure with a generic, meaningful, and quantitative interpretation. Hence, our assessment is solely based on runtime measurements. We discuss proper choices for solution quality indicators in single-and multiobjective optimization, as well as in the presence of noise and constraints. We also discuss the choice of the target values, budget-based targets, and the aggregation of runtimes by using simulated restarts, averages, and empirical cumulative distributions which generalize convergence graphs of single runs. The presented performance assessment is to a large extent implemented in the comparing continuous optimizers (COCO) platform freely available at https://github.com/numbbo/coco.

Published
2022

26. Innovation Paths for Machine Learning in Robotics [Industry Activities]

Author

Freek Stulp, Michael Spranger, Kim Listmann, Stéphane Doncieux, Moritz Tenorth, George Konidaris, Pieter Abbeel, German Aerospace Center (DLR), Sony AI, Tokyo, Bender GmbH, Institut des Systèmes Intelligents et de Robotique (ISIR), Sorbonne Université (SU)-Centre National de la Recherche Scientifique (CNRS), Magazino GmbH, Brown University, University of California (UC), and European Project: 951992

Subjects
[INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Control and Systems Engineering, [INFO.INFO-RB]Computer Science [cs]/Robotics [cs.RO], Electrical and Electronic Engineering, Computer Science Applications
Abstract

International audience; Presents interviews conducted with robotics engineers discussing advances in artificial intelligence, with particular use in machine learning. Advances in artificial intelligence (AI), especially in machine learning (ML), are changing the business models of many companies, and creating entirely new ones. Recent research estimates that AI could boost profitability rates by 38% worldwide, leading to an economic boost of €12 trillion across a variety of industries by 2035. This immense number is an accumulation of many smaller numbers, related to the successful deployment of ML at individual companies, including smalland medium-sized enterprises (SMEs) and start-ups.

Published
2022

27. A New Belief-Based Incomplete Pattern Unsupervised Classification Method

Author

Zhe Liu, Zuo-Wei Zhang, Ma Zongfang, Hao Wang, Yiru Zhang, Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université de Bretagne Sud (UBS)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National de Recherche en Informatique et en Automatique (Inria)-École normale supérieure - Rennes (ENS Rennes)-Centre National de la Recherche Scientifique (CNRS)-Université de Rennes 1 (UR1), Université de Rennes (UNIV-RENNES)-CentraleSupélec-IMT Atlantique Bretagne-Pays de la Loire (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT), Xi'an University of Science and Technology, Norwegian University of Science and Technology (NTNU), Declarative & Reliable management of Uncertain, user-generated Interlinked Data (DRUID), GESTION DES DONNÉES ET DE LA CONNAISSANCE (IRISA-D7), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), and Université de Rennes (UR)

Subjects
0209 industrial biotechnology, unsupervised classification, Computer science, 02 engineering and technology, Fuzzy logic, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], Task (project management), 020901 industrial engineering & automation, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], belief functions theory, Classifier (linguistics), 0202 electrical engineering, electronic engineering, information engineering, incomplete pattern, Cluster analysis, business.industry, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], Pattern recognition, Missing data, fuzzy c-means, Computer Science Applications, Computational Theory and Mathematics, Classification methods, 020201 artificial intelligence & image processing, Artificial intelligence, business, clustering, Information Systems
Abstract

The clustering of incomplete patterns is a very challenging task because the estimations may negatively affect the distribution of real centers and thus cause uncertainty and imprecision in the results. To address this problem, a new belief-based incomplete pattern unsupervised classification method (BPC) is proposed in this paper. Firstly, the complete patterns are grouped into a few clusters by a classical soft method like fuzzy c-means to obtain the corresponding reliable centers and thereby are partitioned into reliable patterns and unreliable ones by an optimization method. Secondly, a basic classifier trained by reliable patterns is employed to classifies unreliable patterns and the incomplete patterns edited by the neighbors. In this way, most of the edited incomplete patterns can be submitted to specific clusters. Finally, some ambiguous patterns will be carefully repartitioned again by a new distance-based rule depending on the obtained reliable centers and belief functions theory. By doing this, a few patterns that are very difficult to classify between different specific clusters will be reasonably submitted to meta-cluster which can characterize the uncertainty and imprecision of the clusters due to missing values. The simulation results show that the BPC has the potential to deal with real datasets. © 2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

Published
2022

28. Wavelets in the Deep Learning Era

Author

Zaccharie Ramzi, Kevin Michalewicz, Jean-Luc Starck, Thomas Moreau, Philippe Ciuciu, Modelling brain structure, function and variability based on high-field MRI data (PARIETAL), Inria Saclay - Ile de France, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Service NEUROSPIN (NEUROSPIN), Direction de Recherche Fondamentale (CEA) (DRF (CEA)), Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Université Paris-Saclay-Direction de Recherche Fondamentale (CEA) (DRF (CEA)), Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Université Paris-Saclay, Astrophysique Interprétation Modélisation (AIM (UMR_7158 / UMR_E_9005 / UM_112)), Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Institut national des sciences de l'Univers (INSU - CNRS)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université de Paris (UP), Service NEUROSPIN (NEUROSPIN), Université Paris-Saclay-Direction de Recherche Fondamentale (CEA) (DRF (CEA)), Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Inria Saclay - Ile de France, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria), Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Institut national des sciences de l'Univers (INSU - CNRS)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université Paris Cité (UPCité), Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Commissariat à l'énergie atomique et aux énergies alternatives (CEA), Modèles et inférence pour les données de Neuroimagerie (MIND), and IFR49 - Neurospin - CEA

Subjects
Statistics and Probability, Denoising, Neural Networks, Applied Mathematics, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], Wavelets, Condensed Matter Physics, Machine Learning, Image restoration, Deep Learning, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [STAT.ML]Statistics [stat]/Machine Learning [stat.ML], [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, 0102 Applied Mathematics, Modeling and Simulation, 0801 Artificial Intelligence and Image Processing, Artificial Intelligence & Image Processing, Geometry and Topology, Computer Vision and Pattern Recognition, 0802 Computation Theory and Mathematics
Abstract

International audience; Sparsity-based methods, such as wavelets, have been the state of the art for more than 20 years for inverse problems before being overtaken by neural networks. In particular, U-nets have proven to be extremely effective. Their main ingredients are a highly nonlinear processing, a massive learning made possible by the flourishing of optimization algorithms with the power of computers (GPU) and the use of large available datasets for training. It is far from obvious to say which of these three ingredients has the biggest impact on the performance. While the many stages of nonlinearity are intrinsic to deep learning, the usage of learning with training data could also be exploited by sparsity-based approaches. The aim of our study is to push the limits of sparsity to use, similarly to U-nets, massive learning and large datasets, and then to compare the results with U-nets. We present a new network architecture, called learnlets, which conserves the properties of sparsity-based methods such as exact reconstruction and good generalization properties, while fostering the power of neural networks for learning and fast calculation. We evaluate the model on image denoising tasks. Our conclusion is that U-nets perform better than learnlets on image quality metrics in distribution, while learnlets have better generalization properties.

Published
2022

29. Order selection with confidence for finite mixture models

Author

Hien D. Nguyen, Daniel Fryer, Geoffrey J. McLachlan, La Trobe University [Melbourne], University of Queensland [Brisbane], and Nguyen, Hien

Subjects
FOS: Computer and information sciences, [INFO.INFO-AI] Computer Science [cs]/Artificial Intelligence [cs.AI], Statistics and Probability, [STAT.AP]Statistics [stat]/Applications [stat.AP], [INFO.INFO-IM] Computer Science [cs]/Medical Imaging, [INFO.INFO-NE] Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], [SCCO.COMP]Cognitive science/Computer science, [INFO.INFO-LG] Computer Science [cs]/Machine Learning [cs.LG], [STAT.OT]Statistics [stat]/Other Statistics [stat.ML], [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], [INFO.INFO-NA]Computer Science [cs]/Numerical Analysis [cs.NA], [STAT.OT] Statistics [stat]/Other Statistics [stat.ML], [INFO.INFO-MO]Computer Science [cs]/Modeling and Simulation, Physics::Geophysics, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], Methodology (stat.ME), [STAT.AP] Statistics [stat]/Applications [stat.AP], [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [SCCO.COMP] Cognitive science/Computer science, [INFO.INFO-IT]Computer Science [cs]/Information Theory [cs.IT], [INFO.INFO-NA] Computer Science [cs]/Numerical Analysis [cs.NA], [INFO.INFO-IM]Computer Science [cs]/Medical Imaging, [INFO.INFO-IT] Computer Science [cs]/Information Theory [cs.IT], [INFO.INFO-MO] Computer Science [cs]/Modeling and Simulation, Statistics - Methodology
Abstract

The determination of the number of mixture components (the order) of a finite mixture model has been an enduring problem in statistical inference. We prove that the closed testing principle leads to a sequential testing procedure (STP) that allows for confidence statements to be made regarding the order of a finite mixture model. We construct finite sample tests, via data splitting and data swapping, for use in the STP, and we prove that such tests are consistent against fixed alternatives. Simulation studies and real data examples are used to demonstrate the performance of the finite sample tests-based STP, yielding practical recommendations of their use as confidence estimators in combination with point estimates such as the Akaike information or Bayesian information criteria. In addition, we demonstrate that a modification of the STP yields a method that consistently selects the order of a finite mixture model, in the asymptotic sense. Our STP is not only applicable for order selection of finite mixture models, but is also useful for making confidence statements regarding any sequence of nested models.

Published
2022

30. Echocardiography Segmentation With Enforced Temporal Consistency

Author

Nathan Painchaud, Nicolas Duchateau, Olivier Bernard, Pierre-Marc Jodoin, Painchaud, Nathan, Département d'informatique [Sherbrooke] (UdeS), Faculté des sciences [Sherbrooke] (UdeS), Université de Sherbrooke (UdeS)-Université de Sherbrooke (UdeS), Modeling & analysis for medical imaging and Diagnosis (MYRIAD), Centre de Recherche en Acquisition et Traitement de l'Image pour la Santé (CREATIS), Université Claude Bernard Lyon 1 (UCBL), Université de Lyon-Université de Lyon-Institut National des Sciences Appliquées de Lyon (INSA Lyon), Université de Lyon-Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Hospices Civils de Lyon (HCL)-Université Jean Monnet - Saint-Étienne (UJM)-Institut National de la Santé et de la Recherche Médicale (INSERM)-Centre National de la Recherche Scientifique (CNRS)-Université Claude Bernard Lyon 1 (UCBL), Université de Lyon-Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Hospices Civils de Lyon (HCL)-Université Jean Monnet - Saint-Étienne (UJM)-Institut National de la Santé et de la Recherche Médicale (INSERM)-Centre National de la Recherche Scientifique (CNRS), Institut Universitaire de France (IUF), and Ministère de l'Education nationale, de l’Enseignement supérieur et de la Recherche (M.E.N.E.S.R.)

Subjects
FOS: Computer and information sciences, [INFO.INFO-AI] Computer Science [cs]/Artificial Intelligence [cs.AI], Computer Science - Machine Learning, left ventricle, Computer Vision and Pattern Recognition (cs.CV), [INFO.INFO-IM] Computer Science [cs]/Medical Imaging, Computer Science - Computer Vision and Pattern Recognition, Machine Learning (cs.LG), [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], [INFO.INFO-CV] Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Image Processing, Computer-Assisted, FOS: Electrical engineering, electronic engineering, information engineering, myocardium, [INFO.INFO-IM]Computer Science [cs]/Medical Imaging, Humans, variational autoencoder, cardiac segmentation, Electrical and Electronic Engineering, Observer Variation, Radiological and Ultrasound Technology, ultrasound, Image and Video Processing (eess.IV), [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], Heart, Deep learning, [INFO.INFO-LG] Computer Science [cs]/Machine Learning [cs.LG], Electrical Engineering and Systems Science - Image and Video Processing, Computer Science Applications, [INFO.INFO-TI] Computer Science [cs]/Image Processing [eess.IV], Echocardiography, [INFO.INFO-TI]Computer Science [cs]/Image Processing [eess.IV], Neural Networks, Computer, CNN, Software
Abstract

Convolutional neural networks (CNN) have demonstrated their ability to segment 2D cardiac ultrasound images. However, despite recent successes according to which the intra-observer variability on end-diastole and end-systole images has been reached, CNNs still struggle to leverage temporal information to provide accurate and temporally consistent segmentation maps across the whole cycle. Such consistency is required to accurately describe the cardiac function, a necessary step in diagnosing many cardiovascular diseases. In this paper, we propose a framework to learn the 2D+time apical long-axis cardiac shape such that the segmented sequences can benefit from temporal and anatomical consistency constraints. Our method is a post-processing that takes as input segmented echocardiographic sequences produced by any state-of-the-art method and processes it in two steps to (i) identify spatio-temporal inconsistencies according to the overall dynamics of the cardiac sequence and (ii) correct the inconsistencies. The identification and correction of cardiac inconsistencies relies on a constrained autoencoder trained to learn a physiologically interpretable embedding of cardiac shapes, where we can both detect and fix anomalies. We tested our framework on 98 full-cycle sequences from the CAMUS dataset, which are available alongside this paper. Our temporal regularization method not only improves the accuracy of the segmentation across the whole sequences, but also enforces temporal and anatomical consistency., 12 pages, accepted for publication in IEEE TMI

Published
2022

31. Ethical Aspects of Artificial Intelligence in Radiation Oncology

Author

Lucien, Lahmi, Marie-France, Mamzer, Anita, Burgun, Catherine, Durdux, Jean-Emmanuel, Bibault, Institut Curie [Paris], Centre de Recherche des Cordeliers (CRC (UMR_S_1138 / U1138)), École Pratique des Hautes Études (EPHE), Université Paris sciences et lettres (PSL)-Université Paris sciences et lettres (PSL)-Institut National de la Santé et de la Recherche Médicale (INSERM)-Sorbonne Université (SU)-Université Paris Cité (UPCité), Health data- and model- driven Knowledge Acquisition (HeKA), Inria de Paris, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Centre de Recherche des Cordeliers (CRC (UMR_S_1138 / U1138)), Université Paris sciences et lettres (PSL)-Université Paris sciences et lettres (PSL)-Institut National de la Santé et de la Recherche Médicale (INSERM)-Sorbonne Université (SU)-Université Paris Cité (UPCité)-École Pratique des Hautes Études (EPHE), Service d'informatique médicale, Hôpital Européen Georges Pompidou [APHP] (HEGP), Assistance publique - Hôpitaux de Paris (AP-HP) (AP-HP)-Hôpitaux Universitaires Paris Ouest - Hôpitaux Universitaires Île de France Ouest (HUPO)-Assistance publique - Hôpitaux de Paris (AP-HP) (AP-HP)-Hôpitaux Universitaires Paris Ouest - Hôpitaux Universitaires Île de France Ouest (HUPO), Service d'informatique médicale et biostatistiques [CHU Necker], Assistance publique - Hôpitaux de Paris (AP-HP) (AP-HP)-CHU Necker - Enfants Malades [AP-HP], Assistance publique - Hôpitaux de Paris (AP-HP) (AP-HP), and Assistance publique - Hôpitaux de Paris (AP-HP) (AP-HP)-Hôpitaux Universitaires Paris Ouest - Hôpitaux Universitaires Île de France Ouest (HUPO)

Subjects
Cancer Research, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Oncology, Artificial Intelligence, Radiation Oncology, [SDV.SP.PHARMA]Life Sciences [q-bio]/Pharmaceutical sciences/Pharmacology, Humans, Radiology, Nuclear Medicine and imaging, Delivery of Health Care, Workflow, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]
Abstract

International audience; Radiation oncology is a field that heavily relies on new technology. Data science and artificial intelligence will have an important role in the entire radiotherapy workflow. A new paradigm of routine healthcare data reuse to automate treatments and provide decision support is emerging. This review will discuss the ethical aspects of the use of artificial intelligence (AI) in radiation oncology. More specifically, the review will discuss the evolution of work through the ages, as well as the impact AI will have on it. We will then explain why AI opens a new technical era for the field and we will conclude on the challenges in the years to come.

Published
2022

32. Interpreting convolutional neural network decision for earthquake detection with feature map visualization, backward optimization and layer-wise relevance propagation methods

Author

Josipa Majstorović, Sophie Giffard-Roisin, Piero Poli, Institut des Sciences de la Terre (ISTerre), Institut national des sciences de l'Univers (INSU - CNRS)-Institut de recherche pour le développement [IRD] : UR219-Université Savoie Mont Blanc (USMB [Université de Savoie] [Université de Chambéry])-Centre National de la Recherche Scientifique (CNRS)-Université Gustave Eiffel-Université Grenoble Alpes (UGA), and Centre National de la Recherche Scientifique (CNRS)

Subjects
[SDU.STU.TE]Sciences of the Universe [physics]/Earth Sciences/Tectonics, Geophysics, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, [SDU.STU.GP]Sciences of the Universe [physics]/Earth Sciences/Geophysics [physics.geo-ph], Geochemistry and Petrology
Abstract

SUMMARY In the recent years, the seismological community has adopted deep learning (DL) models for many diverse tasks such as discrimination and classification of seismic events, identification of P- and S-phase wave arrivals or earthquake early warning systems. Numerous models recently developed are showing high accuracy values, and it has been attested for several tasks that DL models perform better than the classical seismological state-of-art models. However, their performances strongly depend on the DL architecture, the training hyperparameters, and the training data sets. Moreover, due to their complex nature, we are unable to understand how the model is learning and therefore how it is making a prediction. Thus, DL models are usually referred to as a ‘black-box’. In this study, we propose to apply three complementary techniques to address the interpretability of a convolutional neural network (CNN) model for the earthquake detection. The implemented techniques are: feature map visualization, backward optimization and layer-wise relevance propagation. Since our model reaches a good accuracy performance (97%), we can suppose that the CNN detector model extracts relevant characteristics from the data, however a question remains: can we identify these characteristics? The proposed techniques help to answer the following questions: How is an earthquake processed by a CNN model? What is the optimal earthquake signal according to a CNN? Which parts of the earthquake signal are more relevant for the model to correctly classify an earthquake sample? The answer to these questions help understand why the model works and where it might fail, and whether the model is designed well for the predefined task. The CNN used in this study had been trained for single-station detection, where an input sample is a 25 s three-component waveform. The model outputs a binary target: earthquake (positive) or noise (negative) class. The training database contains a balanced number of samples from both classes. Our results shows that the CNN model correctly learned to recognize where is the earthquake within the sample window, even though the position of the earthquake in the window is not explicitly given during the training. Moreover, we give insights on how a neural network builds its decision process: while some aspects can be linked to clear physical characteristics, such as the frequency content and the P and S waves, we also see how different a DL detection is compared to a visual expertise or an STA/LTA detection. On top of improving our model designs, we also think that understanding how such models work, how they perceive an earthquake, can be useful for the comprehension of events that are not fully understood yet such as tremors or low frequency earthquakes.

Published
2022

33. A distributed and incremental algorithm for large-scale graph clustering

Author

Wissem Inoubli, Sabeur Aridhi, Haithem Mezni, Mondher Maddouri, Engelbert Mephu Nguifo, Université de Tunis El Manar (UTM), Computational Algorithms for Protein Structures and Interactions (CAPSID), 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 Complex Systems, Artificial Intelligence & Robotics (LORIA - AIS), 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), Strategies for Modelling and ARtificial inTelligence Laboratory (SMART-LAB), Université de Tunis, University of Jeddah, Laboratoire d'Informatique, de Modélisation et d'Optimisation des Systèmes (LIMOS), Ecole Nationale Supérieure des Mines de St Etienne (ENSM ST-ETIENNE)-Université Clermont Auvergne [2017-2020] (UCA [2017-2020])-Centre National de la Recherche Scientifique (CNRS), This work was partially supported by the CNRS-INRIA/FAPs project 'TempoGraphs' (PRC2243)., Tallinn University, Institut National de Recherche en Informatique et en Automatique (Inria)-Université de Lorraine (UL)-Centre National de la Recherche Scientifique (CNRS), Université de Jendouba (UJ), Taibah University, Université Clermont Auvergne (UCA), Ecole Nationale Supérieure des Mines de St Etienne (ENSM ST-ETIENNE)-Centre National de la Recherche Scientifique (CNRS)-Université Clermont Auvergne (UCA)-Institut national polytechnique Clermont Auvergne (INP Clermont Auvergne), Université Clermont Auvergne (UCA)-Université Clermont Auvergne (UCA), and Ecole Nationale Supérieure des Mines de St Etienne-Université Clermont Auvergne [2017-2020] (UCA [2017-2020])-Centre National de la Recherche Scientifique (CNRS)

Subjects
Big Data, [INFO.INFO-DB]Computer Science [cs]/Databases [cs.DB], Community detection, Computer Networks and Communications, Graph processing, Outliers detection, Graph processing Structural graph clustering Big Graph Analysis Community detection Outliers detection hubs detection, hubs detection, Distributed computing, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Structural graph clustering, Graph clustering, Big graph, Hardware and Architecture, Distributed graph clustering, [INFO.INFO-BI]Computer Science [cs]/Bioinformatics [q-bio.QM], [INFO.INFO-DC]Computer Science [cs]/Distributed, Parallel, and Cluster Computing [cs.DC], SCAN, Software, Big Graph Analysis
Abstract

International audience; Graph clustering is one of the key techniques to understand structures that are presented in networks. In addition to clusters, bridges and outliers detection is also a critical task as it plays an important role in the analysis of networks. Recently, several graph clustering methods are developed and used in multiple application domains such as biological network analysis, recommendation systems and community detection. Most of these algorithms are based on the structural clustering algorithm. Yet, this kind of algorithm is based on the structural similarity. This latter requires to parse all graph’ edges in order to compute the structural similarity. However, the height needs of similarity computing make this algorithm more adequate for small graphs, without significant support to deal with large-scale networks. In this paper, we propose a novel distributed graph clustering algorithm based on structural graph clustering. The experimental results show the efficiency in terms of running time of the proposed algorithm in large networks compared to existing structural graph clustering methods.

Published
2022

34. Dynamic Social Learning Under Graph Constraints

Author

Konstantin Avrachenkov, Suhail M. Shah, Sharayu Moharir, Vivek S. Borkar, Network Engineering and Operations (NEO ), 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), Department of Electrical Engineering [IIT-Bombay] (EE-IIT), Indian Institute of Technology Kanpur (IIT Kanpur), Hong Kong University of Science and Technology (HKUST), This work was supported by the grant `Machine Learning for NetworkAnalytics' from the Indo-French Centre for Promotion of AdvancedScientific Research (Cefipra) and by the grant `Distributed Learning and Controlfor Network Analysis' from Nokia Bell Labs., and The Hong Kong University of Sciences & Technology (HKUST)

Subjects
FOS: Computer and information sciences, Vertex (graph theory), Computer Science - Machine Learning, Control and Optimization, Computer Networks and Communications, annealed dynamics, Stochastic approximation, Machine Learning (cs.LG), [INFO.INFO-NI]Computer Science [cs]/Networking and Internet Architecture [cs.NI], [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], graphical constraints, FOS: Mathematics, Mathematics - Optimization and Control, Equivalence (measure theory), Empirical process, Mathematics, Discrete mathematics, Markov chain, dynamic choice with reinforcement, Probability (math.PR), vertex reinforced random walk, Recursion (computer science), Random walk, [MATH.MATH-PR]Mathematics [math]/Probability [math.PR], Optimization and Control (math.OC), Control and Systems Engineering, Signal Processing, Graph (abstract data type), optimal choice, [MATH.MATH-OC]Mathematics [math]/Optimization and Control [math.OC], Mathematics - Probability
Abstract

International audience; We introduce a model of graph-constrained dynamic choice with reinforcement modeled by positively $\alpha$-homogeneous rewards. We show that its empirical process, which can be written as a stochastic approximation recursion with Markov noise, has the same probability law as a certain vertex reinforced random walk. We use this equivalence to show that for $\alpha > 0$, the asymptotic outcome concentrates around the optimum in a certain limiting sense when 'annealed' by letting $\alpha \to \infty$ slowly.

Published
2022

35. Bayesian uncertainty quantification for machine-learned models in physics

Author

Yarin Gal, Petros Koumoutsakos, Francois Lanusse, Gilles Louppe, Costas Papadimitriou, Department of Computer Science, University of Oxford, Oxford, UK, Computational Science and Engineering Laboratory, School of Engineering and Applied Sciences, Harvard University, Cambridge, MA, USA, Astrophysique Interprétation Modélisation (AIM (UMR_7158 / UMR_E_9005 / UM_112)), Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Institut national des sciences de l'Univers (INSU - CNRS)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université Paris Cité (UPCité), University of Liege, Liege, Belgium, and Department of Mechanical Engineering, University of Thessaly, Volos, Greece

Subjects
[INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], General Physics and Astronomy
Abstract

International audience

Published
2022

36. DiscoPG

Author

Angela Bonifati, Stefania Dumbrava, Emile Martinez, Fatemeh Ghasemi, Malo Jaffré, Pacôme Luton, Thomas Pickles, Université Claude Bernard Lyon 1 - Faculté des sciences (UCBL FS), Université Claude Bernard Lyon 1 (UCBL), Université de Lyon-Université de Lyon, Laboratoire d'InfoRmatique en Image et Systèmes d'information (LIRIS), Université Lumière - Lyon 2 (UL2)-École Centrale de Lyon (ECL), Université de Lyon-Université de Lyon-Université Claude Bernard Lyon 1 (UCBL), Université de Lyon-Institut National des Sciences Appliquées de Lyon (INSA Lyon), Université de Lyon-Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Centre National de la Recherche Scientifique (CNRS), Méthodes et modèles pour les réseaux (METHODES-SAMOVAR), Services répartis, Architectures, MOdélisation, Validation, Administration des Réseaux (SAMOVAR), Institut Mines-Télécom [Paris] (IMT)-Télécom SudParis (TSP)-Institut Mines-Télécom [Paris] (IMT)-Télécom SudParis (TSP), Institut Polytechnique de Paris (IP Paris), Ecole Nationale Supérieure d'Informatique pour l'Industrie et l'Entreprise (ENSIIE), and École normale supérieure de Lyon (ENS de Lyon)

Subjects
Graph databases, [INFO.INFO-DB]Computer Science [cs]/Databases [cs.DB], [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Noeuds, General Engineering, ACM: H.: Information Systems/H.2: DATABASE MANAGEMENT, Property graphs, Graph applications
Abstract

Property graphs are becoming pervasive in a variety of graph processing applications using interconnected data. They allow to encode multi-labeled nodes and edges, as well as their properties, represented as key/value pairs. Although property graphs are widely used in several open-source and commercial graph databases, they lack a schema definition, unlike their relational counterparts. The property graph schema discovery problem consists of extracting the underlying schema concepts and types from such graph datasets. We showcase DiscoPG, a system for efficiently and accurately discovering and exploring property graph schemas. To this end, it leverages hierarchical clustering using a Gaussian Mixture Model, which accounts for both node labels and properties. DiscoPG allows users to perform schema discovery for both static and dynamic graph datasets. Suitable visualization layouts and dedicated dashboards enable the user perception of the static and dynamic inferred schema on the node clusters, as well as the differences in runtimes and clustering quality. To the best of our knowledge, DiscoPG is the first system to tackle the property graph schema discovery problem. As such, it supports the insightful exploration of the graph schema components and their evolving behavior, while revealing the underpinnings of the clustering-based discovery process.

Published
2022

37. Mesh Denoising With Facet Graph Convolutions

Author

Edmond Boyer, Matthieu Armando, Jean-Sébastien Franco, Capture and Analysis of Shapes in Motion (MORPHEO), Inria Grenoble - Rhône-Alpes, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Laboratoire Jean Kuntzmann (LJK), Institut National de Recherche en Informatique et en Automatique (Inria)-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes (UGA)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP ), Université Grenoble Alpes (UGA)-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes (UGA)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP ), Université Grenoble Alpes (UGA), Institut National de Recherche en Informatique et en Automatique (Inria), MSR-Inria, and ANR-19-P3IA-0003,MIAI,MIAI @ Grenoble Alpes(2019)

Subjects
Computer science, Noise reduction, Mesh denoising, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], 020207 software engineering, 02 engineering and technology, [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], Geometric deep learning, [INFO.INFO-CG]Computer Science [cs]/Computational Geometry [cs.CG], Computer Graphics and Computer-Aided Design, Graph, Vertex (geometry), Convolution, Feature preserving, Normal filtering, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Signal Processing, 0202 electrical engineering, electronic engineering, information engineering, Polygon mesh, Computer Vision and Pattern Recognition, Algorithm, Graph convolutions, Software, ComputingMethodologies_COMPUTERGRAPHICS
Abstract

code available at the following address:https://gitlab.inria.fr/marmando/deep-mesh-denoizing; International audience; We examine the problem of mesh denoising, which consists of removing noise from corrupted 3D meshes while preserving existing geometric features. Most mesh denoising methods require a lot of mesh-specific parameter fine-tuning, to account for specific features and noise types. In recent years, data-driven methods have demonstrated their robustness and effectiveness with respect to noise and feature properties on a wide variety of geometry and image problems. Most existing mesh denoising methods still use hand-crafted features, and locally denoise facets rather than examine the mesh globally. In this work, we propose the use of a fully end-to-end learning strategy based on graph convolutions, where meaningful features are learned directly by our network. It operates on a graph of facets, directly on the existing topology of the mesh, without resampling, and follows a multi-scale design to extract geometric features at different resolution levels. Similar to most recent pipelines, given a noisy mesh, we first denoise face normals with our novel approach, then update vertex positions accordingly. Our method performs significantly better than the current state-of-the-art learning-based methods. Additionally, we show that it can be trained on noisy data, without explicit correspondence between noisy and ground-truth facets. We also propose a multi-scale denoising strategy, better suited to correct noise with a low spatial frequency.

Published
2022

38. Kernel-based convolution expansion for facial expression recognition

Author

M. Amine Mahmoudi, Aladine Chetouani, Fatma Boufera, Hedi Tabia, University of Mustapha Stambouli Mascara, Laboratoire pluridisciplinaire de recherche en ingénierie des systèmes, mécanique et énergétique (PRISME), Université d'Orléans (UO)-Institut National des Sciences Appliquées - Centre Val de Loire (INSA CVL), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA), Informatique, BioInformatique, Systèmes Complexes (IBISC), and Université d'Évry-Val-d'Essonne (UEVE)-Université Paris-Saclay

Subjects
[INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, Artificial Intelligence, Signal Processing, Kernel methods, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], Deep learning, Emotion recognition, Computer Vision and Pattern Recognition, Facial expression recognition, Software
Abstract

International audience; The ever-growing depth and width of Convolutional Neural Networks (CNNs) drastically increases the number of their parameters and requires more powerful devices to train and deploy. In this paper, we propose a new architecture that outperforms the classical linear convolution function by expanding the latter to a higher degree kernel function without additional weights. We opt for Taylor Series Kernel which maps input data to a higher-dimensional Reproducing Kernel Hilbert Space (RKHS). Mapping features to a higher-order RKHS is performed in both implicit and explicit ways. For the former way, we compute several polynomial kernels of different degrees leveraging the kernel trick. Whereas, the latter way is achieved by concatenating the result of these polynomial kernels. The proposed Taylor Series Kernelized Convolution (TSKC) is able to learn more complex patterns than the linear convolution kernel and thus be more discriminative. The experiments conducted on Facial Expression Recognition (FER) datasets demonstrate that TSKC outperforms the ordinary convolution layer without additional parameters.

Published
2022

39. mpNet: Variable Depth Unfolded Neural Network for Massive MIMO Channel Estimation

Author

Taha Yassine, Luc Le Magoarou, Artificial Intelligence, Domaine Hypermedia (IRT b<>com) (Hypermedia), Institut de Recherche Technologique b-com (IRT b-com)-Institut de Recherche Technologique b-com (IRT b-com), Institut de Recherche Technologique b-com (IRT b-com), B<>COM [Cesson Sévigné], Institut d'Électronique et des Technologies du numéRique (IETR), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-Nantes Université - pôle Sciences et technologie, and Nantes Université (Nantes Univ)-Nantes Université (Nantes Univ)

Subjects
sparse recovery, Signal Processing (eess.SP), FOS: Computer and information sciences, Computer Science - Machine Learning, neural network, Computer Science - Artificial Intelligence, online learning, Channel estimation, unsupervised learning, Machine Learning (cs.LG), [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], Computer Science - Networking and Internet Architecture, [SPI]Engineering Sciences [physics], [INFO.INFO-NI]Computer Science [cs]/Networking and Internet Architecture [cs.NI], [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], FOS: Electrical engineering, electronic engineering, information engineering, Analytical models, Electrical Engineering and Systems Science - Signal Processing, Electrical and Electronic Engineering, Manifolds, matching pursuit (MP), Computer Science::Information Theory, Networking and Internet Architecture (cs.NI), Signal to noise ratio, Applied Mathematics, MIMO channel estimation, Matching pursuit algorithms, Computer Science Applications, Artificial Intelligence (cs.AI), Deep unfolding, Estimation, [SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processing, Neural networks
Abstract

International audience; Massive multiple-input multiple-output (MIMO) communication systems have a huge potential both in terms of data rate and energy efficiency, although channel estimation becomes challenging for a large number of antennas. Using a physical model allows to ease the problem by injecting a priori information based on the physics of propagation. However, such a model rests on simplifying assumptions and requires to know precisely the configuration of the system, which is unrealistic in practice. In this paper we present mpNet, an unfolded neural network specifically designed for massive MIMO channel estimation. It is trained online in an unsupervised way. Moreover, mpNet is computationally efficient and automatically adapts its depth to the signal-to-noise ratio (SNR). The method we propose adds flexibility to physical channel models by allowing a base station (BS) to automatically correct its channel estimation algorithm based on incoming data, without the need for a separate offline training phase. It is applied to realistic millimeter wave channels and shows great performance, achieving a channel estimation error almost as low as one would get with a perfectly calibrated system. It also allows incident detection and automatic correction, making the BS resilient and able to automatically adapt to changes in its environment.

Published
2022

40. Convolutional neural networks contain structured strong lottery tickets

Author

Carvalho Walraven da Cunha, Arthur, d'Amore, Francesco, Natale, Emanuele, Combinatorics, Optimization and Algorithms for Telecommunications (COATI), 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)-COMmunications, Réseaux, systèmes Embarqués et Distribués (Laboratoire I3S - COMRED), Laboratoire d'Informatique, Signaux, et Systèmes de Sophia Antipolis (I3S), Université Nice Sophia Antipolis (1965 - 2019) (UNS), COMUE Université Côte d'Azur (2015-2019) (COMUE UCA)-COMUE Université Côte d'Azur (2015-2019) (COMUE UCA)-Centre National de la Recherche Scientifique (CNRS)-Université Côte d'Azur (UCA)-Université Nice Sophia Antipolis (1965 - 2019) (UNS), COMUE Université Côte d'Azur (2015-2019) (COMUE UCA)-COMUE Université Côte d'Azur (2015-2019) (COMUE UCA)-Centre National de la Recherche Scientifique (CNRS)-Université Côte d'Azur (UCA)-Laboratoire d'Informatique, Signaux, et Systèmes de Sophia Antipolis (I3S), COMUE Université Côte d'Azur (2015-2019) (COMUE UCA)-COMUE Université Côte d'Azur (2015-2019) (COMUE UCA)-Centre National de la Recherche Scientifique (CNRS)-Université Côte d'Azur (UCA), Aalto University, Centre National de la Recherche Scientifique (CNRS), and COMUE Université Côte d'Azur (2015-2019) (COMUE UCA)

Subjects
[INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG]
Abstract

The Strong Lottery Ticket Hypothesis (SLTH) states that randomly-initialised neural networks contain subnetworks that can perform well without any training. Although unstructured pruning has been extensively studied in this context, its structured counterpart, which can deliver significant computational and memory efficiency gains, has been largely unexplored. One of the main reasons for this gap is the limitations of the underlying mathematical tools used in formal analyses of the SLTH. In this paper, we overcome these limitations: we leverage recent advances in the multidimensional generalisation of the Random Subset-Sum Problem and obtain a variant that admits the stochastic dependencies that arise when addressing structured pruning in the SLTH. We apply this result to prove, for a wide class of random Convolutional Neural Networks, the existence of structured subnetworks that can approximate any sufficiently smaller network. This is the first work to address the SLTH for structured pruning, opening up new avenues for further research on the hypothesis and contributing to the understanding of the role of overparameterization in deep learning.

Published
2023

41. Event-based neural learning for quadrotor control

Author

Carvalho, Estéban, Susbielle, Pierre, Marchand, Nicolas, Hably, Ahmad, Dibangoye, Jilles, GIPSA - COntrol, PErception, Robots, navigation and Intelligent Computing (GIPSA-COPERNIC), GIPSA Pôle Sciences des Données (GIPSA-PSD), Grenoble Images Parole Signal Automatique (GIPSA-lab), Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes (UGA)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP ), Université Grenoble Alpes (UGA)-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes (UGA)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP ), Université Grenoble Alpes (UGA)-Grenoble Images Parole Signal Automatique (GIPSA-lab), Université Grenoble Alpes (UGA), Institut National des Sciences Appliquées de Lyon (INSA Lyon), Université de Lyon-Institut National des Sciences Appliquées (INSA), CITI Centre of Innovation in Telecommunications and Integration of services (CITI), Université de Lyon-Institut National des Sciences Appliquées (INSA)-Université de Lyon-Institut National des Sciences Appliquées (INSA)-Institut National de Recherche en Informatique et en Automatique (Inria), GIPSA-Services (GIPSA-Services), Université de Lyon, Robots coopératifs et adaptés à la présence humaine en environnements (CHROMA), Inria Grenoble - Rhône-Alpes, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-CITI Centre of Innovation in Telecommunications and Integration of services (CITI), Université de Lyon-Institut National des Sciences Appliquées (INSA)-Université de Lyon-Institut National des Sciences Appliquées (INSA)-Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National des Sciences Appliquées de Lyon (INSA Lyon), Université de Lyon-Institut National des Sciences Appliquées (INSA)-Université de Lyon-Institut National des Sciences Appliquées (INSA)-Inria Lyon, Institut National de Recherche en Informatique et en Automatique (Inria), ANR-11-LABX-0025,PERSYVAL-lab,Systemes et Algorithmes Pervasifs au confluent des mondes physique et numérique(2011), ANR-10-EQPX-0044,ROBOTEX,Réseau national de plateformes robotiques d'excellence(2010), and ANR-21-ESRE-0015,TIRREX,Infrastructure technologique pour la recherche d'excellence en robotique(2021)

Subjects
event-based control, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Quadrotor, [INFO.INFO-AU]Computer Science [cs]/Automatic Control Engineering, online learning, deep neural network, trajectory tracking, [SPI.AUTO]Engineering Sciences [physics]/Automatic
Abstract

International audience; The design of a simple and adaptive flight controller is a real challenge in aerial robotics. A simple flight controller often generates a poor flight tracking performance. Furthermore, adaptive algorithms might be costly in time and resources or deep learning based methods may cause instability problems, for instance in presence of disturbances. In this paper, we propose an event-based neural learning control strategy that combines the use of a standard cascaded flight controller enhanced by a deep neural network that learns the disturbances in order to improve the tracking performance. The strategy relies on two events: one allowing the improvement of tracking errors and the second to ensure closed-loop system stability. After a validation of the proposed strategy in a ROS/Gazebo simulation environment, its effectiveness is confirmed in real experiments in the presence of wind disturbance.

Published
2023

42. Expert Variability and Deep Learning Performance in Spinal Cord Lesion Segmentation for Multiple Sclerosis Patients

Author

Walsh, Ricky, Meurée, Cédric, Kerbrat, Anne, Masson, Arthur, Hussein, Burhan Rashid, Gaubert, Malo, Galassi, Francesca, Combès, Benoit, Neuroimagerie: méthodes et applications (EMPENN), Institut National de la Santé et de la Recherche Médicale (INSERM)-Inria Rennes – Bretagne Atlantique, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-SIGNAL, IMAGE ET LANGAGE (IRISA-D6), Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT), Service de Neurologie [Rennes] = Neurology [Rennes], CHU Pontchaillou [Rennes], Centre d'Investigation Clinique [Rennes] (CIC), Université de Rennes (UR)-Hôpital Pontchaillou-Institut National de la Santé et de la Recherche Médicale (INSERM), Service de Neuroradiologie [Rennes], ANR-20-THIA-0018,AI4SDA,IA pour l'analyse de données sémantiques(2020), ANR-21-RHUS-0014,Primus,Transforming the care of patients with Multiple Sclerosis using a multidimensional data-driven clinical decision support system(2021), and ANR-10-COHO-0002,OFSEP,Observatoire Français de la Sclérose en Plaques(2010)

Subjects
Multiple sclerosis spinal cord magnetic resonance imaging lesion segmentation inter-rater variability intrarater variability deep learning automated segmentation, spinal cord, deep learning, intrarater variability, inter-rater variability, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], Multiple sclerosis, intra-rater variability, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [INFO.INFO-TI]Computer Science [cs]/Image Processing [eess.IV], lesion segmentation, [INFO.INFO-IM]Computer Science [cs]/Medical Imaging, magnetic resonance imaging, automated segmentation
Abstract

Accepted at 2023 IEEE 36th International Symposium on Computer-Based Medical Systems (CBMS).© 2023 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other work.; International audience; Multiple sclerosis (MS) patients often present with lesions in spinal cord magnetic resonance (MR) volumes. However, accurately detecting these lesions is challenging and prone to inter-and intra-rater variability. Deep learning-based methods have the potential to aid clinicians in detecting and segmenting MS lesions, but can also be affected by rater variability. This study assesses the inter-and intra-rater variability in manual segmentation of spinal cord lesions, and evaluates raters and a state-of-the-art nnU-Net model against a ground truth (GT) segmentation of a senior expert. Four experts segmented twelve spinal cord MR volumes from six patients twice, at a time distance of two weeks. Considerable inter-and intra-rater variability were observed, with the total number of detected lesions ranging from 28 to 60, depending on the rater. Moreover, the segmented volumes of individual lesions varied substantially between raters. All raters and the model achieved high precision when evaluated against the senior expert GT, but sensitivity was notably lower. These results motivate the need for more sensitive automated methods to aid clinicians in lesion detection, and suggest that consideration should be given to inter-rater variability when training and evaluating automated methods.

Published
2023

43. A Framework for Fast and Stable Representations of Multiparameter Persistent Homology Decompositions

Author

Loiseaux, David, Carrière, Mathieu, Blumberg, Andrew J., Understanding the Shape of Data (DATASHAPE), 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)-Inria Saclay - Ile de France, Institut National de Recherche en Informatique et en Automatique (Inria), Columbia University [New York], and ANR-19-P3IA-0002,3IA@cote d'azur,3IA Côte d'Azur(2019)

Subjects
Computational Geometry (cs.CG), FOS: Computer and information sciences, Computer Science - Machine Learning, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [MATH.MATH-AT]Mathematics [math]/Algebraic Topology [math.AT], FOS: Mathematics, Algebraic Topology (math.AT), Computer Science - Computational Geometry, Mathematics - Algebraic Topology, [INFO.INFO-CG]Computer Science [cs]/Computational Geometry [cs.CG], Machine Learning (cs.LG)
Abstract

Topological data analysis (TDA) is an area of data science that focuses on using invariants from algebraic topology to provide multiscale shape descriptors for geometric data sets such as point clouds. One of the most important such descriptors is {\em persistent homology}, which encodes the change in shape as a filtration parameter changes; a typical parameter is the feature scale. For many data sets, it is useful to simultaneously vary multiple filtration parameters, for example feature scale and density. While the theoretical properties of single parameter persistent homology are well understood, less is known about the multiparameter case. In particular, a central question is the problem of representing multiparameter persistent homology by elements of a vector space for integration with standard machine learning algorithms. Existing approaches to this problem either ignore most of the multiparameter information to reduce to the one-parameter case or are heuristic and potentially unstable in the face of noise. In this article, we introduce a new general representation framework that leverages recent results on {\em decompositions} of multiparameter persistent homology. This framework is rich in information, fast to compute, and encompasses previous approaches. Moreover, we establish theoretical stability guarantees under this framework as well as efficient algorithms for practical computation, making this framework an applicable and versatile tool for analyzing geometric and point cloud data. We validate our stability results and algorithms with numerical experiments that demonstrate statistical convergence, prediction accuracy, and fast running times on several real data sets.

Published
2023

44. A Multimodal Dynamical Variational Autoencoder for Audiovisual Speech Representation Learning

Author

Sadok, Samir, Leglaive, Simon, Girin, Laurent, Alameda-Pineda, Xavier, Séguier, Renaud, Institut d'Électronique et des Technologies du numéRique (IETR), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-Nantes Université - pôle Sciences et technologie, Nantes Université (Nantes Univ)-Nantes Université (Nantes Univ), GIPSA - Cognitive Robotics, Interactive Systems, & Speech Processing (GIPSA-CRISSP), GIPSA Pôle Parole et Cognition (GIPSA-PPC), Grenoble Images Parole Signal Automatique (GIPSA-lab), Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes (UGA)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP ), Université Grenoble Alpes (UGA)-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes (UGA)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP ), Université Grenoble Alpes (UGA)-Grenoble Images Parole Signal Automatique (GIPSA-lab), Université Grenoble Alpes (UGA), Vers des robots à l’intelligence sociale au travers de l’apprentissage, de la perception et de la commande (ROBOTLEARN), Inria Grenoble - Rhône-Alpes, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Université Grenoble Alpes (UGA), ANR-19-P3IA-0003,MIAI,MIAI @ Grenoble Alpes(2019), ANR-19-CE33-0008,ML3RI,Apprentissage de bas-niveau d'ineractions robotiques multi-modales avec plusieurs personnes(2019), and European Project: 871245,H2020-EU.2.1.1. - INDUSTRIAL LEADERSHIP - Leadership in enabling and industrial technologies - Information and Communication Technologies (ICT),SPRING(2020)

Subjects
FOS: Computer and information sciences, Computer Science - Machine Learning, Sound (cs.SD), [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Audio and Speech Processing (eess.AS), FOS: Electrical engineering, electronic engineering, information engineering, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], Computer Science - Sound, Computer Science - Multimedia, Electrical Engineering and Systems Science - Audio and Speech Processing, Machine Learning (cs.LG), Multimedia (cs.MM)
Abstract

In this paper, we present a multimodal \textit{and} dynamical VAE (MDVAE) applied to unsupervised audio-visual speech representation learning. The latent space is structured to dissociate the latent dynamical factors that are shared between the modalities from those that are specific to each modality. A static latent variable is also introduced to encode the information that is constant over time within an audiovisual speech sequence. The model is trained in an unsupervised manner on an audiovisual emotional speech dataset, in two stages. In the first stage, a vector quantized VAE (VQ-VAE) is learned independently for each modality, without temporal modeling. The second stage consists in learning the MDVAE model on the intermediate representation of the VQ-VAEs before quantization. The disentanglement between static versus dynamical and modality-specific versus modality-common information occurs during this second training stage. Extensive experiments are conducted to investigate how audiovisual speech latent factors are encoded in the latent space of MDVAE. These experiments include manipulating audiovisual speech, audiovisual facial image denoising, and audiovisual speech emotion recognition. The results show that MDVAE effectively combines the audio and visual information in its latent space. They also show that the learned static representation of audiovisual speech can be used for emotion recognition with few labeled data, and with better accuracy compared with unimodal baselines and a state-of-the-art supervised model based on an audiovisual transformer architecture., Comment: 25 pages, 14 figures, https://samsad35.github.io/site-mdvae/

Published
2023

45. Clustering Multivariate Longitudinal Data Application on Disease Progression Modeling ⋆

Author

Liekah, Loujain, Elghazel, Haytham, de Marchi, Fabien, Hacid, Mohand-Saïd, Laboratoire d'InfoRmatique en Image et Systèmes d'information (LIRIS), Université Lumière - Lyon 2 (UL2)-École Centrale de Lyon (ECL), Université de Lyon-Université de Lyon-Université Claude Bernard Lyon 1 (UCBL), Université de Lyon-Institut National des Sciences Appliquées de Lyon (INSA Lyon), Université de Lyon-Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Centre National de la Recherche Scientifique (CNRS), Base de Données (BD), Université de Lyon-Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Centre National de la Recherche Scientifique (CNRS)-Université Lumière - Lyon 2 (UL2)-École Centrale de Lyon (ECL), Data Mining and Machine Learning (DM2L), and European Project: 875171,QUALITOP

Subjects
Subtyping, Disease progression modeling, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Alzheimer's disease, Multivariate longitudinal clustering, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]
Abstract

International audience; Methods for identifying homogeneous groups with varying characteristics in longitudinal data have been receiving increasing attention in recent years, especially in the medical domain. Exploiting electronic health records (EHRs) to infer patient subtypes can support practitioners in improving the decision-making process. In this paper, we propose a dynamic method for clustering multivariate longitudinal data, which constitutes a transparent solution for patient subtyping and modeling disease progression. Based on the assumption that subjects with similar disease trajectories share the same patterns, we subtype patients based on their medical history then learn the disease progression model. We cluster data periodically, and maintain the results and update the deduced subtypes by applying a borrowed approach from the data integration domain, namely entity matching. We test our method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) real-world dataset. We demonstrate how our results can facilitate the early detection of dementia two years on average before the actual diagnosis, and therefore assist in the development of support and prevention procedures for patients. We compare our clustering with the state-of-the-art data stream clustering algorithms, and show that our method exhibits higher effectiveness in terms of both internal and external validation metrics.

Published
2023

46. Motion-DVAE: Unsupervised learning for fast human motion denoising

Author

Fiche, Guénolé, Leglaive, Simon, Alameda-Pineda, Xavier, Séguier, Renaud, Institut d'Électronique et des Technologies du numéRique (IETR), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-Nantes Université - pôle Sciences et technologie, Nantes Université (Nantes Univ)-Nantes Université (Nantes Univ), Vers des robots à l’intelligence sociale au travers de l’apprentissage, de la perception et de la commande (ROBOTLEARN), Inria Grenoble - Rhône-Alpes, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Université Grenoble Alpes (UGA), and ANR-18-EURE-0022,DIGISPORT,Digital Sports Sciences(2018)

Subjects
FOS: Computer and information sciences, Artificial Intelligence (cs.AI), [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Computer Science - Artificial Intelligence, Computer Vision and Pattern Recognition (cs.CV), Computer Science - Computer Vision and Pattern Recognition, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV]
Abstract

Pose and motion priors are crucial for recovering realistic and accurate human motion from noisy observations. Substantial progress has been made on pose and shape estimation from images, and recent works showed impressive results using priors to refine frame-wise predictions. However, a lot of motion priors only model transitions between consecutive poses and are used in time-consuming optimization procedures, which is problematic for many applications requiring real-time motion capture. We introduce Motion-DVAE, a motion prior to capture the short-term dependencies of human motion. As part of the dynamical variational autoencoder (DVAE) models family, Motion-DVAE combines the generative capability of VAE models and the temporal modeling of recurrent architectures. Together with Motion-DVAE, we introduce an unsupervised learned denoising method unifying regression- and optimization-based approaches in a single framework for real-time 3D human pose estimation. Experiments show that the proposed approach reaches competitive performance with state-of-the-art methods while being much faster.

Published
2023

47. pyBumpHunter: A model independent bump hunting tool in Python for High Energy Physics analyses

Author

Louis Vaslin, Samuel Calvet, Vincent Barra, Julien Donini, Laboratoire de Physique de Clermont (LPC), Institut National de Physique Nucléaire et de Physique des Particules du CNRS (IN2P3)-Centre National de la Recherche Scientifique (CNRS)-Université Clermont Auvergne (UCA), Laboratoire d'Informatique, de Modélisation et d'Optimisation des Systèmes (LIMOS), Ecole Nationale Supérieure des Mines de St Etienne (ENSM ST-ETIENNE)-Centre National de la Recherche Scientifique (CNRS)-Université Clermont Auvergne (UCA)-Institut national polytechnique Clermont Auvergne (INP Clermont Auvergne), Université Clermont Auvergne (UCA)-Université Clermont Auvergne (UCA), BARRA, Vincent, and HEP, INSPIRE

Subjects
data analysis method, [PHYS.HEXP] Physics [physics]/High Energy Physics - Experiment [hep-ex], [INFO.INFO-TS] Computer Science [cs]/Signal and Image Processing, FOS: Physical sciences, [INFO.INFO-LG] Computer Science [cs]/Machine Learning [cs.LG], programming, High Energy Physics - Experiment, [PHYS.HPHE] Physics [physics]/High Energy Physics - Phenomenology [hep-ph], High Energy Physics - Experiment (hep-ex), High Energy Physics - Phenomenology, High Energy Physics - Phenomenology (hep-ph), statistical analysis, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [PHYS.HPHE]Physics [physics]/High Energy Physics - Phenomenology [hep-ph], [PHYS.HEXP]Physics [physics]/High Energy Physics - Experiment [hep-ex], performance, new particle
Abstract

The BumpHunter algorithm is widely used in the search for new particles in High Energy Physics analysis. This algorithm offers the advantage of evaluating the local and global p-values of a localized deviation in the observed data without making any hypothesis on the supposed signal. The increasing popularity of the Python programming language motivated the development of a new public implementation of this algorithm in Python, called pyBumpHunter, together with several improvements and additional features. It is the first public implementation of the BumpHunter algorithm to be added to Scikit-HEP. This paper presents in detail the BumpHunter algorithm as well as all the features proposed in this implementation. All these features have been tested in order to demonstrate their behaviour and performance., 14 pages, 9 figures

Published
2023

48. Neural State-Dependent Delay Differential Equations

Author

Monsel, Thibault, Semeraro, Onofrio, Mathelin, Lionel, Charpiat, Guillaume, DAtascience, trAnsition, Fluid instability, contrOl, Turbulence (DATAFLOT), Laboratoire Interdisciplinaire des Sciences du Numérique (LISN), Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Mécanique-Energétique (M.-E.), Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS), TAckling the Underspecified (TAU), Inria Saclay - Ile de France, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Laboratoire Interdisciplinaire des Sciences du Numérique (LISN), Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-CentraleSupélec-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS), and ANR-20-CE23-0025,SPEED,Simulation de systèmes d'EDP de la Physique par apprentissage profond(2020)

Subjects
FOS: Computer and information sciences, Delay, Neural Networks, Continuous-depth models, Computer Science - Artificial Intelligence, Neural Ordinary Differential Equations, [MATH.MATH-DS]Mathematics [math]/Dynamical Systems [math.DS], Physical Modelling, Dynamical Systems (math.DS), Delay Differential Equations, Dynamical Systems, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], Artificial Intelligence (cs.AI), [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Discontinuities, Delay Differential Equations Delay Differential Equations Neural Networks Discontinuities Neural Ordinary Differential Equations NODE Physical Modelling Dynamical Systems Numerical Integration Continuous-depth models Software DDE solver, Differential Equations, [NLIN.NLIN-CD]Nonlinear Sciences [physics]/Chaotic Dynamics [nlin.CD], FOS: Mathematics, DDE solver, Mathematics - Dynamical Systems, Numerical Integration, NODE, Software
Abstract

Discontinuities and delayed terms are encountered in the governing equations of a large class of problems ranging from physics, engineering, medicine to economics. These systems are impossible to be properly modelled and simulated with standard Ordinary Differential Equations (ODE), or any data-driven approximation including Neural Ordinary Differential Equations (NODE). To circumvent this issue, latent variables are typically introduced to solve the dynamics of the system in a higher dimensional space and obtain the solution as a projection to the original space. However, this solution lacks physical interpretability. In contrast, Delay Differential Equations (DDEs) and their data-driven, approximated counterparts naturally appear as good candidates to characterize such complicated systems. In this work we revisit the recently proposed Neural DDE by introducing Neural State-Dependent DDE (SDDDE), a general and flexible framework featuring multiple and state-dependent delays. The developed framework is auto-differentiable and runs efficiently on multiple backends. We show that our method is competitive and outperforms other continuous-class models on a wide variety of delayed dynamical systems.

Published
2023

49. Energy-Aware Spreading Factor Selection in LoRaWAN Using Delayed-Feedback Bandits

Author

Navas, Renzo Efrain, Dandachi, Ghina, Hadjadj-Aoul, Yassine, Maillé, Patrick, Département Systèmes Réseaux, Cybersécurité et Droit du numérique (IMT Atlantique - SRCD), IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT), Self-prOtecting The futurE inteRNet (SOTERN), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-RÉSEAUX, TÉLÉCOMMUNICATION ET SERVICES (IRISA-D2), Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS), mEasuRing and ManagIng Network operation and Economic (ERMINE), Inria Rennes – Bretagne Atlantique, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-RÉSEAUX, TÉLÉCOMMUNICATION ET SERVICES (IRISA-D2), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut National de Recherche en Informatique et en Automatique (Inria), Université de Rennes (UR), IFIP, and ANR-20-CE25-0011,INTELLIGENTSIA,Intelligence de la périphérie du réseau par des boucles d'apprentissage et la génération de données pour une automatisation du réseau basée sur l'inférence de ses états.(2020)

Subjects
reinforcement learning, IoT, ADR, Spreading Factor selection, Thompson sampling, LoRa, LoRaWAN, [INFO.INFO-NI]Computer Science [cs]/Networking and Internet Architecture [cs.NI], [INFO.INFO-PF]Computer Science [cs]/Performance [cs.PF], bandits, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], energy consumption, LPWAN, Adaptive Data Rate, [INFO.INFO-ET]Computer Science [cs]/Emerging Technologies [cs.ET]
Abstract

International audience; LoRaWAN networks can involve large numbers of wireless devices relying on batteries to sense the environment and send data to gateways. A critical trade-off for transmission performance (packet delivery ratio) versus energy conservation (and hence, the device lifespan) appears when deciding the transmission parameters, in particular, the Spreading Factor (SF) to be used by each node. In this paper, we use lightweight reinforcement learning techniques, namely multi-armed bandits, for each node to select an appropriate SF, based on preferences regarding that trade-off. Unlike previous works on that topic, we relax some assumptions to aim at a realistic implementation: our solution does not assume immediate rewards, or that each device communicates with only one gateway. Additionally, we build explicit MAC commands for the method to work in practice and implement it in the ns-3 simulator using a state-of-the-art LoRaWAN module. We share the source code of our implementation and our simulation results. Those simulations show that when energy conservation is critical for IoT nodes, such lightweight learning algorithms outperform LoRaWAN's legacy Adaptive Data Rate algorithm, both in single- and multi-gateway scenarios.

Published
2023

50. Domain adaptation for pen trajectory reconstruction from kinematic sensors

Author

Imbert, Florent, Soullard, Yann, Tavenard, Romain, Anquetil, Eric, Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT), Université de Rennes (UR), Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA), Intuitive user-interaction for document (INTUIDOC), SIGNAL, IMAGE ET LANGAGE (IRISA-D6), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université de Rennes 2 (UR2), Littoral, Environnement, Télédétection, Géomatique UMR 6554 (LETG), Université de Brest (UBO)-Université de Rennes 2 (UR2)-Centre National de la Recherche Scientifique (CNRS)-Institut de Géographie et d'Aménagement Régional de l'Université de Nantes (Nantes Univ - IGARUN), Nantes Université - pôle Humanités, Nantes Université (Nantes Univ)-Nantes Université (Nantes Univ)-Nantes Université - pôle Humanités, Nantes Université (Nantes Univ)-Nantes Université (Nantes Univ), and Observation de l’environnement par imagerie complexe (OBELIX)

Subjects
[INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], [INFO]Computer Science [cs], Inertial Measurement Units, Digital Pen, [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], Online Handwriting, Domain Adaptation, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]
Abstract

International audience; Handwriting with connected pen becomes one of the major human-computer easy interaction methods. In comparison with traditional touch screen handwriting systems, the pen-based interaction method has the advantage of producing the online handwriting signal without surface constraints. Indeed, people who write on a paper obtain the corresponding pen trajectory coordinates that represent the online handwriting signal. Furthermore, the feeling of writing on paper is important particularly for the children during the learning of writing. In this work which is part of ANR Franco-German KIHT project with Stabilo, we introduce a domain adaptation-based approach that reconstructs the paper handwriting traces of the digital stylus Digipen of STABILO which is equipped with a wireless trajectory tracking system based on kinematic sensors. We use unsupervised domain adaptation method, to pass from the tablet domain where the ground truth (the online trace of the writing on the tablet) is known, to the paper domain where only the input sensors data are known.

Published
2023

Catalog

Books, media, physical & digital resources
See catalog results