Publications

2019

• ORSUM 2019 2nd workshop on online recommender systems and user modeling
  
  • Vinagre João
  • Jorge Alípio Mário
  • Bifet Albert
  • Al-Ghossein Marie
  , 2019, pp.562-563. (10.1145/3298689.3347057)
  DOI : 10.1145/3298689.3347057
• Recommendation System-based Upper Confidence Bound for Online Advertising
  
  • Nguyen-Thanh Nhan
  • Marinca Dana
  • Khawam Kinda
  • Rohde David
  • Vasile Flavian
  • Lohan Elena Simona
  • Martin Steven
  • Quadri Dominique
  , 2019. In this paper, the method UCB-RS, which resorts to recommendation system (RS) for enhancing the upper-confidence bound algorithm UCB, is presented. The proposed method is used for dealing with non-stationary and large-state spaces multi-armed bandit problems. The proposed method has been targeted to the problem of the product recommendation in the online advertising. Through extensive testing with RecoGym, an OpenAI Gym-based reinforcement learning environment for the product recommendation in online advertising, the proposed method outperforms the widespread reinforcement learning schemes such as Epsilon-Greedy, Upper Confidence (UCB1) and Exponential Weights for Exploration and Exploitation (EXP3).
• Preface to the 1st Multi-Paradigm Modeling for Cyber-Physical Systems (MPM4CPS 2019)
  
  • van Mierlo Simon
  • Syriani Eugene
  • Blouin Dominique
  • Amrani Moussa
  • Deantoni Julien
  • Wimmer Manuel
  , 2019, pp.2. (10.1109/MODELS-C.2019.00066)
  DOI : 10.1109/MODELS-C.2019.00066
• A Fundamental Storage-Communication Tradeoff in Distributed Computing with Straggling Nodes
  
  • Yan Qifa
  • Wigger Michèle
  • Yang Sheng
  • Tang Xiaohu
  , 2019. The optimal storage-computation tradeoff is characterized for a MapReduce-like distributed computing system with straggling nodes, where only a part of the nodes can be utilized to compute the desired output functions. The result holds for arbitrary output functions and thus generalizes previous results that restricted to linear functions. Specifically, in this work, we propose a new information-theoretical converse and a new matching coded computing scheme, that we call coded computing for straggling systems (CCS).
• Storage-Computation-Communication Tradeoff in Distributed Computing: Fundamental Limits and Complexity
  
  • Yan Qifa
  • Yang Sheng
  • Wigger Michèle
  , 2019. Distributed computing has become one of the most important frameworks in dealing with large computation tasks. In this paper, we propose a systematic construction of coded computing schemes for MapReduce-type distributed systems. The construction builds upon placement delivery arrays (PDA), originally proposed by Yan et al. for coded caching schemes. The main contributions of our work are threefold. First, we identify a class of PDAs, called Comp-PDAs, and show how to obtain a coded computing scheme from any Comp-PDA. We also characterize the normalized number of stored files (storage load), computed intermediate values (computation load), and communicated bits (communication load), of the obtained schemes in terms of the Comp-PDA parameters. Then, we show that the performance achieved by Comp-PDAs describing Maddah-Ali and Niesen's coded caching schemes matches a new information-theoretic converse, thus establishing the fundamental region of all achievable performance triples. In particular, we characterize all the Comp-PDAs achieving the pareto-optimal storage, computation, and communication (SCC) loads of the fundamental region. Finally, we investigate the file complexity of the proposed schemes, i.e., the smallest number of files required for implementation. In particular, we describe Comp-PDAs that achieve pareto-optimal SCC triples with significantly lower file complexity than the originally proposed Comp-PDAs.
• Towards a Formal Specification of Multi-paradigm Modelling
  
  • Amrani Moussa
  • Blouin Dominique
  • Heinrich Robert
  • Rensink Arend
  • Vangheluwe Hans
  • Wortmann Andreas
  , 2019.
• MAD-EEG: an EEG dataset for decoding auditory attention to a target instrument in polyphonic music
  
  • Cantisani Giorgia
  • Trégoat Gabriel
  • Essid Slim
  • Richard Gael
  , 2019. We present MAD-EEG, a new, freely available dataset for studying EEG-based auditory attention decoding considering the challenging case of subjects attending to a target instrument in polyphonic music. The dataset represents the first music-related EEG dataset of its kind, enabling, in particular, studies on single-trial EEG-based attention decoding, while also opening the path for research on other EEG-based music analysis tasks. MAD-EEG has so far collected 20-channel EEG signals recorded from 8 subjects listening to solo, duo and trio music excerpts and attending to one pre-specified instrument. The proposed experimental setting differs from the ones previously considered as the stimuli are polyphonic and are played to the subject using speakers instead of headphones. The stimuli were designed considering variations in terms of number and type of instruments in the mixture, spatial rendering, music genre and melody that is played. Preliminary results obtained with a state-of-the-art stimulus reconstruction algorithm commonly used for speech stimuli show that the audio representation reconstructed from the EEG response is more correlated with that of the attended source than with the one of the unattended source, proving the dataset to be suitable for such kind of studies.
• Learning about random media from near-surface backscattering: using machine learning to measure particle size and concentration
  
  • Gower Artur L.
  • Gower Robert M.
  • Deakin Jonathan
  • J. Parnell William
  • Abrahams. I. David
  , 2018. We ask what can be measured from a random media by using backscattered waves, emitted from and received at one source. We show that in 2D both the particle radius and concentration can be accurately measured for particles with Dirichlet boundary conditions. This is challenging to do for a wide range of particle volume fractions, 1% to 21%, because for high volume fraction the effects of multiple scattering are not completely understood. Across this range we show that the concentration can be accurately measured just from the mean backscattered wave, but the particle radius requires the backscattered variance, or intensity. We also show that using incident wavenumbers 0 ≤ k ≤ 0.8 is ideal to measure particle radius between 0 and 2. To answer these questions we use supervised machine learning (kernel ridge regression) together with a large, precise, dataset of simulated backscattered waves. One long term aim is to develop a device, powered by data, that can characterise random media from backscattering with little prior knowledge. Here we take the first steps towards this goal.
• Smoothing technique for nonsmooth composite minimization with linear operator
  
  • Nguyen Quang Van
  • Fercoq Olivier
  • Cevher Volkan
  , 2019.
• A Proximal Approach for Sparse Multiclass SVM
  
  • Chierchia Giovanni
  • Pustelnik Nelly
  • Pesquet Jean-Christophe
  • Pesquet-Popescu Beatrice
  , 2015.
• Quantum attacks against iterated block ciphers
  
  • Kaplan Marc
  , 2014.
• Multi-Valued Routing Tracks for FPGAs in 28nm FDSOI Technology
  
  • Chaudhuri Sumanta
  • Graba Tarik
  • Mathieu Yves
  , 2019. In this paper we present quaternary and ternary routing tracks for FPGAs, and their implementation in 28nm FDSOI technology. We discuss the transistor level design of multi-valued repeaters, multiplexers and translators, and specific features of FDSOI technology which make it possible. Next we compare the multi-valued routing architectures with equivalent single driver two-valued routing architectures. We show that for long tracks, it is possible to achieve upto 3x reduction in dynamic switching energy, upto 2x reduction in routing wire area and 10% reduction in area dedicated to routing resources. The multi-valued tracks are slightly more susceptible to process variation. We present a layout method for multivalued standard cells and determine the layout overhead.We conclude with various usage scenarios of these tracks.
• Learning Graph Representations by Dendrograms
  
  • Charpentier Bertrand
  • Bonald Thomas
  , 2018.
• Stochastic simulation of urban environments: Application to Path-loss in wireless systems
  
  • Courtat Thomas
  • Decreusefond Laurent
  • Martins Philippe
  , 2016.
• A MQTT OPC UA Configuration Tool
  
  • Liu Zepeng
  , 2019.
• Conception et développement d’étalons pour la mesure des paramètres S en mode mixte de circuits intégrés et méthodes associées
  
  • Pham Thi Dao
  , 2019. Des circuits différentiels sont largement utilisés pour la conception de composants hyperfréquences principalement en raison de leur meilleure immunité au bruit. Ces circuits doivent être caractérisés au moyen de paramètres S en mode mixte (mode différentiel, mode commun et conversion entre les deux modes). De plus, la tendance à la miniaturisation et à l’intégration des dispositifs hyperfréquences conduit à l’utilisation de structures planaires ou coplanaires telles que les lignes micro-ruban ou les lignes coplanaires. La structure coplanaire avec les conducteurs déposés à la surface supérieure du substrat évite de réaliser des trous métallisés, et donc simplifie la fabrication et empêche l’apparition d’éléments parasites. Du point de vue de la métrologie électrique, il est nécessaire d’établir la traçabilité des mesures de paramètres S en mode mixte au Système International d’unités (SI). La méthode d’étalonnage Multimode Thru – Reflect – Line (TRL), dérivée de l’étalonnage TRL couramment utilisée pour les mesures de paramètres S de circuits asymétriques, est bien adaptée à cette problématique. En effet, l’impédance caractéristique, qui définit l’impédance de référence du système de mesure, peut être obtenue à partir des constantes de propagation déterminées lors de la procédure Multimode TRL et des capacités linéiques en DC.Nous présentons la première conception et la réalisation d’un kit d’étalonnage Multimode TRL et d’un kit de vérification à base des lignes coplanaires couplées en configuration « Ground – Signal – Ground – Signal – Ground » sur un substrat de quartz (SiO2) à faibles pertes diélectriques pour des mesures de paramètres S en mode mixte sur wafer de 1 GHz à 40 GHz. Les mesures sont effectuées à l’aide de deux méthodes : l’approche « one-tier » basée sur la procédure d’étalonnage Multimode TRL afin de déterminer et de corriger l’ensemble des erreurs systématiques ou bien l’approche « two-tier » qui fractionne la détermination et la correction des termes d’erreur en deux étapes dont la deuxième est associée à la méthode Multimode TRL. La faisabilité et la validation de ces techniques sont démontrées par des mesures d’éléments de vérification, constitués de lignes (adaptées, désadaptées et déséquilibrées) et d’atténuateurs en T, qui montrent un très bon accord entre les valeurs mesurées et simulées.La propagation des incertitudes est évaluée soit à partir du calcul des dérivées partielles à l’aide de l’outil Metas.Unclib ou bien par simulation numérique basée sur la méthode de Monte Carlo. La précision des mesures de paramètres S sous pointes dépend des sources d’influence attribuées aux mesures et aux imperfections des étalons telles que le bruit et la non-linéarité de l’analyseur de réseaux vectoriel, la stabilité des câbles, la répétabilité des mesures et la sensibilité dans la réalisation des étalons. Faute de temps, nous nous limitons à estimer la propagation d’incertitudes liées à la répétabilité de mesure des étalons et du dispositif sous test (DST) aux valeurs des paramètres S corrigés de la ligne désadaptée. Les résultats montrent que l’approche des dérivées partielles basée sur une approximation de la série de Taylor au premier ordre ne peut pas être utilisée avec précision à cause de l’influence significative de la non-linéarité des fonctions mathématiques de l’algorithme Multimode TRL. La méthode Monte Carlo s’avère alors plus précise bien qu’elle nécessite des temps de calcul très longs.
• Enhancing video applications through timed metadata
  
  • Potetsianakis Emmanouil
  , 2019. Video recording devices are often equipped with sensors (smartphones for example, with GPS receiver, gyroscope etc.), or used in settings where sensors are present (e.g. monitor cameras, in areas with temperature and/or humidity sensors). As a result, many systems process and distribute video together with timed metadata streams, often sourced as User-Generated Content. Video delivery has been thoroughly studied, however timed metadata streams have varying characteristics and forms, thus a consistent and effective way to handle them in conjunction with the video streams does not exist. In this Thesis we study ways to enhance video applications through timed metadata. We define as timed metadata all the non-audiovisual data recorded or produced, that are relevant to a specific time on the media timeline. ”Enhancing” video applications has a double meaning, and this work consists of two respective parts. First, using the timed metadata to extend the capabilities of multimedia applications, by introducing novel functionalities. Second, using the timed metadata to improve the content delivery for such applications. To extend multimedia applications, we have taken an exploratory approach, and we demonstrate two use cases with application examples. In the first case, timed metadata is used as input for generating content, and in the second, it is used to extend the navigational capabilities for the underlying multimedia content. By designing and implementing two different application scenarios we were able to identify the potential and limitations of video systems with timed metadata. We use the findings from the first part, to work from the perspective of enhancing video applications, by using the timed metadata to improve delivery of the content. More specifically, we study the use of timed metadata for multi-variable adaptation in multi-view video delivery - and we test our proposals on one of the platforms developed previously. Our final contribution is a buffering scheme for synchronous and lowlatency playback in live streaming systems.
• Automated Detection and Segmentation of Mitochondrial Images based on Gradient Enhancement and Adaptive Gabor Filter
  
  • Nguyen-Thanh Nhan
  • Pham Tuan
  • Ichikawa Kazuhisa
  , 2019. Information of cellular organelles location and morphology is essential for cancer simulation. In order to obtain such information, the segmentation of the organelles from electronic microscopy intracellular image is crucial. In this paper, we focus on the automatic segmentation of mitochondria organelle which is one of the most important organelles tightly related to the form of cancer. A simple three-stage strategy which includes coarse segmentation, detection and fine segmentation is proposed for fully automatic mitochondria segmentation. The local gradient calculation provides a weight factor matrix and a orientation matrix. The weight factor matrix will improve the contrast of organelle boundary of intracellular images and hence facilitate both coarse and fine mitochondria segmentation. The orientation matrix will be used for enhancing the Gabor feature extraction which make the mitochodrial detection process more accurate. Machine learning-based classifiers including k-nearest neighbor (k-NN), support vector machine (SVM)-based and neural network (NN)-based classifiers, are considered to learn with eight extracted features for mitochondrial detection. Experimental results on focused ion beam (FIB) and scanning electron microscope (SEM) images of cancer cellular of human head and neck squamous cell carcinoma (SCC-61) have shown the effectiveness of proposed method.
• Design of a flexible analog-to-feature converter for smart acquisition of biological signals
  
  • Back Antoine
  • Chollet Paul
  • Fercoq Olivier
  • Desgreys Patricia
  , 2019.
• Analysing meaning making of social touch in computer-mediated interaction
  
  • Heron Robin
  • Detienne Françoise
  • Safin Stéphane
  • Baker Michael J
  • Zang Zhuoming
  • Lecolinet Eric
  , 2019.
• On Tree-based Methods for Similarity Learning
  
  • Clémençon Stéphan
  • Vogel Robin
  , 2019, 11943, pp.676–688. In many situations, the choice of an adequate similarity measure or metric on the feature space dramatically determines the performance of machine learning methods. Building automatically such measures is the specific purpose of metric/similarity learning. In [21], similarity learning is formulated as a pairwise bipartite ranking problem: ideally, the larger the probability that two observations in the feature space belong to the same class (or share the same label), the higher the similarity measure between them. From this perspective, the ROC curve is an appropriate performance criterion and it is the goal of this article to extend recur-sive tree-based ROC optimization techniques in order to propose efficient similarity learning algorithms. The validity of such iterative partitioning procedures in the pairwise setting is established by means of results pertaining to the theory of U-processes and from a practical angle, it is discussed at length how to implement them by means of splitting rules specifically tailored to the similarity learning task. Beyond these theoret-ical/methodological contributions, numerical experiments are displayed and provide strong empirical evidence of the performance of the algorith-mic approaches we propose.
• Glottal Opening Measurements in VCV and VCCV Sequences
  
  • Elie Benjamin
  • Amelot Angelique
  • Laprie Yves
  • Maeda Shinji
  , 2019. Many studies on speech acoustics and production use articulatory synthesis as a framework to investigate the relationship between articulatory gestures and acoustic features. Although supraglottal articulatory models are available, usually built from vocal tract imaging acquisitions, glottal gestures are commonly modeled with simple geometric primitives which do not necessarily reflect reality. This study is a first step towards the development of a database of realistic glottal gestures which will be used to design the glottal opening dynamics in articulatory synthesis paradigms. In the first part of this paper, we present experimental measurements of glottal opening dynamics in VCV and VCCV sequences uttered by real subjects, thanks to a specifically designed external photoglottographic device (ePGG). The corpus was designed to highlight the differences in glottis opening between fricatives and stops. The existence of different patterns of glottal opening is evidenced according to the class of the consonants. A numerical study is then used to show the influence of these patterns on the production of sounds and on the coarticulation.
• Exploiting View Synthesis for Super-multiview Video Compression
  
  • Nikitin Pavel
  • Cagnazzo Marco
  • Jung Joël
  • Fiandrotti Attilio
  , 2019. Super-multiview video consists in a 2D arrangement of cameras acquiring the same scene and it is a well-suited format for immersive and free navigation video services. However, the large number of acquired viewpoints calls for extremely effective compression tools. View synthesis allows to reconstruct a viewpoint using nearby cameras texture and depth information. In this work we explore the potential of recent advances in view synthesis algorithms to enhance the compression performances of super-multiview video. Towards this end we consider five methods that replace one viewpoint with a synthesized view, possibly enhanced with some side information. Our experiments suggest that, if the geometry information (i.e. depth map) is reliable, these methods have the potential to improve rate-distortion performance with respect to traditional approaches, at least for some specific content and configuration. Moreover, our results shed some light about how to further improve compression performance by integrating new view-synthesis prediction tools within a 3D video encoder. (10.1145/3349801.3349820)
  DOI : 10.1145/3349801.3349820
• Explanatory AI for Pertinent Communication in Autonomic Systems
  
  • Pol Marius
  • Dessalles Jean-Louis
  • Diaconescu Ada
  , 2020, 1037, pp.212-227. Autonomic computing systems maintain high-level goals by continuously adapting to a changing environment, yet their internal operations have no comprehensible meaning to humans. This paper proposes a dialogue management system based on a communication interface acting as a bridge between subsymbolic and symbolic knowledge representation levels. The communication interface explains the autonomic system operation in a human comprehensible form by representing the sensed data in conceptual spaces. Based on a knowledge base generated by the communication interface, the dialogue management system produces a pertinent flow of conversation between autonomic systems and humans, activating only when exceptional situations are encountered. Our approach is incremental, with the objective of enabling pertinent communication with any artificial system. We build a proof-of-concept implementation of the proposed solution in a smart home platform managed by an autonomic system. (10.1007/978-3-030-29516-5_16)
  DOI : 10.1007/978-3-030-29516-5_16
• Représentations fonctionnelles et modélisation paramétriques et statistiques des antennes, du canal de propagation radio et de leurs interactions
  
  • Roblin Christophe
  , 2019.