Publications

2017

• Dynamic mitigation of EDFA power excursions with machine learning
  
  • Huang Yishen
  • Gutterman Craig L.
  • Samadi Payman
  • Cho Patricia B.
  • Samoud Wiem
  • Ware Cédric
  • Lourdiane Mounia
  • Zussman Gil
  • Bergman Keren
  Optics Express, Optical Society of America - OSA Publishing, 2017, 25 (3), pp.2245 - 2258. Dynamic optical networking has promising potential to support the rapidly changing traffic demands in metro and long-haul networks. However, the improvement in dynamicity is hindered by wavelength-dependent power excursions in gain-controlled erbium doped fiber amplifiers (EDFA) when channels change rapidly. We introduce a general approach that leverages machine learning (ML) to characterize and mitigate the power excursions of EDFA systems with different equipment and scales. An ML engine is developed and experimentally validated to show accurate predictions of the power dynamics in cascaded EDFAs. Recommended channel provisioning based on the ML predictions achieves within 1% error of the lowest possible power excursion over 94% of the time. We also showcase significant mitigation of EDFA power excursions in super-channel provisioning when compared to the first-fit wavelength assignment algorithm (10.1364/OE.25.002245)
  DOI : 10.1364/OE.25.002245
• On perturbed proximal gradient algorithms
  
  • Atchadé Yves
  • Fort Gersende
  • Moulines Éric
  Journal of Machine Learning Research, Microtome Publishing, 2017, 18 (10), pp.1-33.
• Top-k Querying of Unknown Values under Order Constraints (Extended Version)
  
  • Amarilli Antoine
  • Amsterdamer Yael
  • Milo Tova
  • Senellart Pierre
  , 2017. Many practical scenarios make it necessary to evaluate top-k queries over data items with partially unknown values. This paper considers a setting where the values are taken from a numerical domain, and where some partial order constraints are given over known and unknown values: under these constraints, we assume that all possible worlds are equally likely. Our work is the first to propose a principled scheme to derive the value distributions and expected values of unknown items in this setting, with the goal of computing estimated top-k results by interpolating the unknown values from the known ones. We study the complexity of this general task, and show tight complexity bounds, proving that the problem is intractable, but can be tractably approximated. We then consider the case of tree-shaped partial orders, where we show a constructive PTIME solution. We also compare our problem setting to other top-k definitions on uncertain data.
• Les facettes de l'Open Data : émergence, fondements et travail en coulisses
  
  • Denis Jérôme
  • Goëta Samuel
  , 2017, pp.121-138. Dans ce chapitre, nous revenons sur l'émergence des politiques d'open data en mettant en lumière les grands principes (de la transparence jusqu'à la modernisation de l'administration, en passant par la libre circulation de l'information) que différentes initiatives ont progressivement stabilisés pour faire de l'ouverture des données publiques un enjeu international. Nous montrons ensuite, à partir d'une enquête ethnographique menée dans plusieurs institutions françaises ce que cette ouverture implique concrètement : un travail délicat qui demeure largement invisible et représente le coût caché des principes fondateurs de l'open data.
• Rate Allocation in Predictive Video Coding Using a Convex Optimization Framework
  
  • Fiengo Aniello
  • Chierchia Giovanni
  • Cagnazzo Marco
  • Pesquet-Popescu Béatrice
  IEEE Transactions on Image Processing, Institute of Electrical and Electronics Engineers, 2017, 26 (1), pp.479 - 489. Optimal rate allocation is among the most challenging tasks to perform in the context of predictive video coding, because of the dependencies between frames induced by motion compensation. In this paper, using a recursive rate-distortion model that explicitly takes into account these dependencies, we approach the frame-level rate allocation as a convex optimization problem. This technique is integrated into the recent HEVC encoder, and tested on several standard sequences. Experiments indicate that the proposed rate allocation ensures a better performance (in the rate-distortion sense) than the standard HEVC rate control, and with a little loss w.r.t. an optimal exhaustive research which is largely compensated by a much shorter execution time. (10.1109/TIP.2016.2621666)
  DOI : 10.1109/TIP.2016.2621666
• Proposal of new solution for service advertisement for ETSI ITS environments: CAM-Infrastructure
  
  • Labiod Houda
  • Servel Alain
  • Segarra Gérard
  • Hammi Badis
  • Monteuuis Jean-Philippe
  , 2017, pp.1-4. Collaborative Intelligent Transportation Systems are almost part of our everyday life. A C-ITS environment can provide numerous services that soon will become essential to roads’ users. The latter resides in improvement of road safety, entertainment, and commercial services. However to provide such services, the C-ITS environment needs an advertisement and dissemination service of the latter. Indeed, users have to be aware of the available services in order to request them if needed. Actual standards of service announcement show their limits, especially regarding the support of several communication profiles. For this reason, this paper, describes a new service advertisement message called CAM-Infrastructure. The latter is compliant with ETSI standards and is deployed in a nationwide scale project.
• Towards building 3D individual models from MRI segmentation and tractography to enhance surgical planning for pediatric pelvic tumors and malformations
  
  • Muller Cécile
  • Virzi Alessio
  • Marret Jean-Baptiste
  • Mille Eva
  • Berteloot Laureline
  • Grevent David
  • Blanc Thomas
  • Garcelon Nicolas
  • Buffet Isabelle
  • Hullier-Ammard Elisabeth
  • Gori Pietro
  • Boddaert Nathalie
  • Bloch Isabelle
  • Sarnacki Sabine
  , 2017, pp.113-115.
• Urban area change detection based on generalized likelihood ratio test
  
  • Zhao Weiying
  • Lobry Sylvain
  • Maître Henri
  • Nicolas Jean-Marie
  • Tupin Florence
  , 2017.
• Frame rate vs Resolution: a subjective evaluation of spatio-temporal perceived quality under varying computational budgets
  
  • Debattista Kurt
  • Bugeja Keith
  • Spina Sandro
  • Bashford-Rogers Thomas
  • Hulusic Vedad
  Computer Graphics Forum, Wiley, 2017. Maximising performance for rendered content requires making compromises on quality parameters depending on the computational resources available. Yet, it is currently unclear which parameters best maximise perceived quality. This work investigates perceived quality across computational budgets for the primary spatio-temporal parameters of resolution and frame rate. Three experiments are conducted. Experiment 1 (n = 26) shows that participants prefer fixed frame rates of 60 frames per second (fps) at lower resolutions over 30 fps at higher resolutions. Experiment 2 (n = 24) explores the relationship further with more budgets and quality settings and again finds 60 fps is generally preferred even when more resources are available. Experiment 3 (n = 25) permits the use of adaptive frame rates, and analyses the resource allocation across seven budgets. Results show that while participants allocate more resources to frame rate at lower budgets the situation reverses once higher budgets are available and a frame rate of around 40 fps is achieved. In the overall, the results demonstrate a complex relationship between frame rate and resolution’s effects on perceived quality. This relationship can be harnessed, via the results and models presented, to obtain more cost-effective virtual experiences.
• Uniform bootstrap central limit theorems for Harris chains
  
  • Ciołek Gabriela
  , 2017. The main objective of this talk is to present bootstrap uniform functional central limit theorem for Harris recurrent Markov chains over uniformly bounded classes of functions. We show that the result can be generalized also to the unbounded case. To avoid some complicated mixing conditions, we make use of the well-known regeneration properties of Markov chains. We show that in the atomic case the proof of the bootstrap uniform central limit theorem for Markov chains for functions dominated by a function in L2 space proposed by Radulovic (2004) can be significantly simplified. Regenerative properties of Markov chains can be applied in order to extend some concepts in robust statistics from i.i.d. to a Markovian setting. Bertail and Clémençon (2006) have dened an inuence function and Fréchet dierentiability on the torus what allowed to The main objective of this talk is to present bootstrap uniform functional central limit theorem for Harris recurrent Markov chains over uniformly bounded classes of functions. We show that the result can be generalized also to the unbounded case. To avoid some complicated mixing conditions, we make use of the well-known regeneration properties of Markov chains. We show that in the atomic case the proof of the bootstrap uniform central limit theorem for Markov chains for functions dominated by a function in L2 space proposed by Radulovic (2004) can be signicantly simplified. Regenerative properties of Markov chains can be applied in order to extend some concepts in robust statistics from i.i.d. to a Markovian setting. Bertail and Clémençon (2006) have defined an inluence function and Fréchet differentiability on the torus what allowed to extend the notion of robustness from single observations to the blocks of data instead. In this talk, we present bootstrap uniform central limit theorems for Fréchet differentiable functionals in a Markovian case.The notion of robustness from single observations to the blocks of data instead. In this talk, we present bootstrap uniform central limit theorems for Fréchet differentiable functionals in a Markovian case.
• Semi-automatic teeth segmentation in cone-beam computed tomography by graph-cut with statistical shape prior
  
  • Evain Timothée
  • Ripoche Xavier
  • Atif J.
  • Bloch Isabelle
  , 2017, pp.1197-1200. We propose a new semi-automatic framework for tooth segmentation in Cone-Beam Computed Tomography (CBCT) combining shape priors based on a statistical shape model and graph cut optimization. Poor image quality and similarity between tooth and cortical bone intensities are overcome by strong constraints on the shape and on the targeted area. The segmentation quality was assessed on 64 tooth images for which a reference segmentation was available, with an overall Dice coefficient above 0.95 and a global consistency error less than 0.005.
• Prioritized network coding scheme for multi-layer video streaming
  
  • Baccouch Hana
  • Ageneau Paul-Louis
  • Tizon Nicolas
  • Boukhatem Nadia
  , 2017.
• Foreword to Radio Science for Humanity: URSI-France 2017 Workshop
  
  • Tanzi Tullio
  • Hamelin Joel
  Radio Science Bulletin, Union Radio-Scientifique Internationale (URSI), 2017 (360), pp.60-61.
• Hyperparameter optimization of deep neural networks: combining Hperband with Bayesian model selection
  
  • Bertrand Hadrien
  • Ardon Roberto
  • Perrot Matthieu
  • Bloch Isabelle
  , 2017. One common problem in building deep learning architectures is the choice of the hyper-parameters. Among the various existing strategies, we propose to combine two complementary ones. On the one hand, the Hyperband method formalizes hyper-parameter optimization as a resource allocation problem, where the resource is the time to be distributed between many configurations to test. On the other hand, Bayesian optimization tries to model the hyper-parameter space as efficiently as possible to select the next model to train. Our approach is to model the space with a Gaussian process and sample the next group of models to evaluate with Hyperband. Preliminary results show a slight improvement over each method individually, suggesting the need and interest for further experiments.
• Règles d'Associations Temporelles de signaux sociaux pour la synthèse d'Agents Conversationnels Animés : Application aux attitudes sociales
  
  • Janssoone Thomas
  • Clavel Chloé
  • Bailly Kevin
  • Richard Gael
  Revue des Sciences et Technologies de l'Information - Série RIA : Revue d'Intelligence Artificielle, Lavoisier, 2017. Afin d'améliorer l'interaction entre des humains et des agents conversationnels animés (ACA), l'un des enjeux majeurs du domaine est de générer des agents crédibles socialement. Dans cet article, nous présentons une méthode, intitulée SMART pour social multimodal association rules with timing, capable de trouver automatiquement des associations temporelles entre l'utilisation de signaux sociaux (mouvements de tête, expressions faciales, prosodie. . .) issues de vidéos d'interactions d'humains exprimant différents états affectifs (comportement, attitude, émotions,. . .). Notre système est basé sur un algorithme de fouille de séquences qui lui permet de trouver des règles d'associations temporelles entre des signaux sociaux extraits automatiquement de flux audio-vidéo. SMART va également analyser le lien de ces règles avec chaque état affectif pour ne conserver que celles qui sont pertinentes. Finalement, SMART va les enrichir afin d'assurer une animation facile d'un ACA pour qu'il exprime l'état voulu. Dans ce papier, nous formalisons donc l'implémentation de SMART et nous justifions son inté-rêt par plusieurs études. Dans un premier temps, nous montrons que les règles calculées sont bien en accord avec la littérature en psychologie et sociologie. Ensuite, nous présentons les résultats d'évaluations perceptives que nous avons conduites suite à des études de corpus pro-posant l'expression d'attitudes sociales marquées. ABSTRACT. In the field of Embodied Conversational Agent (ECA) one of the main challenges is to generate socially believable agents. The long run objective of the present study is to infer rules for the multimodal generation of agents' socio-emotional behaviour. In this paper, we introduce the Social Multimodal Association Rules with Timing (SMART) algorithm. It proposes to Revue d'intelligence artificielle-n o 4/2017, 511-537 512 RIA. Volume 31-n o 4/2017 learn the rules from the analysis of a multimodal corpus composed by audio-video recordings of human-human interactions. The proposed methodology consists in applying a Sequence Mining algorithm using automatically extracted Social Signals such as prosody, head movements and facial muscles activation as an input. This allows us to infer Temporal Association Rules for the behaviour generation. We show that this method can automatically compute Temporal Association Rules coherent with prior results found in the literature especially in the psychology and sociology fields. The results of a perceptive evaluation confirms the ability of a Temporal Association Rules based agent to express a specific stance. (10.3166/RIA.31.511-537)
  DOI : 10.3166/RIA.31.511-537
• Trends in Social Network Analysis - Information Propagation, User Behavior Modeling, Forecasting, and Vulnerability Assessment
  
  • Missaoui Rokia
  • Abdessalem Talel
  • Latapy Mathieu
  , 2017, pp.255. <p>The book collects contributions from experts worldwide addressing recent scholarship in social network analysis such as influence spread, link prediction, dynamic network biclustering, and delurking. It covers both new topics and new solutions to known problems. The contributions rely on established methods and techniques in graph theory, machine learning, stochastic modelling, user behavior analysis and natural language processing, just to name a few. This text provides an understanding of using such methods and techniques in order to manage practical problems and situations. Trends in Social Network Analysis: Information Propagation, User Behavior Modelling, Forecasting, and Vulnerability Assessment appeals to students, researchers, and professionals working in the field.</p> <p> </p> (10.1007/978-3-319-53420-6)
  DOI : 10.1007/978-3-319-53420-6
• La fabrique des données brutes. Le travail en coulisses de l'open data
  
  • Denis Jérôme
  • Goëta Samuel
  , 2017. Depuis quelques années, les initiatives d’open data se sont multipliées à travers le monde. Présentées jusque dans la presse grand public comme une ressource inexploitée, le « pétrole » sur lequel le monde serait assis, les données publiques sont devenues objet de toutes les attentions et leur ouverture porteuse de toutes les promesses, à la fois terreau d’un renouveau démocratique et moteur d’une innovation distribuée. Comme dans les sciences, qui ont connu un mouvement de focalisation similaire sur les données et leur partage, l’injonction à l’ouverture opère une certaine mise en invisibilité. Le vocabulaire de la « libération », de la « transparence » et plus encore celui de la « donnée brute » efface toute trace des conditions de production des données, des contextes de leurs usages initiaux et pose leur universalité comme une évidence. Ce chapitre explore les coulisses de l’open data afin de retrouver les traces de cette production et d’en comprendre les spécificités. À partir d’une série d’entretiens ethnographiques dans diverses institutions, il décrit la fabrique des données brutes, dont l’ouverture ne se résume jamais à une mise à disponibilité immédiate, évidente et universelle. Il montre que trois aspects sont particulièrement sensibles dans le processus d’ouverture : l’identification, l'extraction et la « brutification » des données. Ces trois séries d’opérations donnent à voir l’épaisseur sociotechnique des données brutes dont la production mêle dimensions organisationnelles, politiques et techniques.
• CLEAR: Covariant LEAst-Square Refitting with Applications to Image Restoration
  
  • Deledalle Charles-Alban
  • Papadakis Nicolas
  • Salmon Joseph
  • Vaiter Samuel
  SIAM Journal on Imaging Sciences, Society for Industrial and Applied Mathematics, 2017, 10 (1), pp.243-284. In this paper, we propose a new framework to remove parts of the systematic errors affecting popular restoration algorithms, with a special focus for image processing tasks. Generalizing ideas that emerged for $\ell_1$ regularization, we develop an approach re-fitting the results of standard methods towards the input data. Total variation regularizations and non-local means are special cases of interest. We identify important covariant information that should be preserved by the re-fitting method, and emphasize the importance of preserving the Jacobian (w.r.t. the observed signal) of the original estimator. Then, we provide an approach that has a ``twicing'' flavor and allows re-fitting the restored signal by adding back a local affine transformation of the residual term. We illustrate the benefits of our method on numerical simulations for image restoration tasks. (10.1137/16M1080318)
  DOI : 10.1137/16M1080318
• Closed-form expressions of the eigen decomposition of 2 x 2 and 3 x 3 Hermitian matrices
  
  • Deledalle Charles-Alban
  • Denis Loic
  • Tabti Sonia
  • Tupin Florence
  , 2017. The eigen decomposition of covariance matrices is at the core of many data analysis techniques. The study of 2-components or 3-components vector fields typically requires computing numerous eigen decompositions of 2 x 2 or 3 x 3 matrices. This is, for example, the case in the analysis of interferometric or polarimetric SAR images, see MuLoG algorithm (https://hal.archives-ouvertes.fr/hal-01388858). The closed-form expression of eigen-values and eigenvectors then provides a way to derive faster data processing algorithms. This note gives these expressions in the general case (special cases where some coefficients are zero, or the eigenvalues are not separated may not be covered and then require either to introduce a small perturbation of the initial matrix or to derive other expressions).
• Parameter Sensitivity Analysis of the Energy/Frequency Convexity Rule for Nanometer-scale Application Processors
  
  • de Vogeleer Karel
  • Memmi Gerard
  • Jouvelot Pierre
  Sustainable Computing : Informatics and Systems, Elsevier, 2017, 15, pp.16-27. Both theoretical and experimental evidence are presented in this work in order to validate the existence of an Energy/Frequency Convexity Rule, which relates energy consumption and microprocessor frequency for nanometer-scale microprocessors. Data gathered during several month-long experimental acquisition campaigns, supported by several independent publications, suggest that energy consumed is indeed depending on the microprocessor's clock frequency, and, more interestingly, the curve exhibits a clear minimum over the processor's frequency range. An analytical model for this behavior is presented and motivated, which fits well with the experimental data. A parameter sensitivity analysis shows how parameters affect the energy minimum in the clock frequency space. The conditions are discussed under which this convexity rule can be exploited, and when other methods are more effective, with the aim of improving the computer system's energy management efficiency. We show that the power requirements of the computer system, besides the microprocessor, and the overhead affect the location of the energy minimum the most. The sensitivity analysis of the Energy/Frequency Convexity Rule puts forward a number of simple guidelines especially for by low-power systems, such as battery-powered and embedded systems, and less likely by high-performance computer systems. (10.1016/j.suscom.2017.05.001)
  DOI : 10.1016/j.suscom.2017.05.001
• FDOPA Patterns in Adrenal Glands
  
  • Moreau Aurélie
  • Giraudet Anne Laure
  • Kryza David
  • Borson-Chazot Françoise
  • Bournaud-Salinas Claire
  • Mognetti Thomas
  • Lifante Jean-Christophe
  • Combemale Patrick
  • Giammarile Francesco
  • Houzard Claire
  Clinical Nuclear Medicine, Lippincott, Williams & Wilkins, 2017, 42 (5), pp.379-382. (10.1097/RLU.0000000000001636)
  DOI : 10.1097/RLU.0000000000001636
• Convergence to multi-resource fairness under end-to-end window control
  
  • Bonald Thomas
  • Roberts James
  • Vitale Christian
  , 2017. The paper relates to multi-resource sharing between flows with heterogeneous requirements as arises in networks with wireless links or software routers implementing network function virtualization. Bottleneck max fairness (BMF) is a sharing objective in this context with good performance. The paper shows that BMF results when local fairness is imposed at each resource while flow rates are controlled by an end-to-end window. We analytically prove convergence to BMF under a fluid model when flows share a network limited to 2 resources while numerical results confirm BMF convergence for larger networks. Simulation results illustrate the impact of packetized transmission.
• Using modular extension to provably protect Edwards curves against fault attacks
  
  • Dugardin Margaux
  • Guilley Sylvain
  • Moreau Martin
  • Najm Zakaria
  • Rauzy Pablo
  Journal of Cryptographic Engineering, Springer, 2017, vol. 7, nb. 4.
• Towards the Generation of Expressive Co-Speech Gestures
  
  • Ravenet Brian
  • Clavel Chloé
  • Pelachaud Catherine I
  , 2017.
• Acoustic Features for Environmental Sound Analysis
  
  • Serizel Romain
  • Bisot Victor
  • Essid Slim
  • Richard Gael
  , 2017, pp.71-101. Most of the time it is nearly impossible to differentiate between particular type of sound events from a waveform only. Therefore, frequency domain and time-frequency domain representations have been used for years providing representations of the sound signals that are more inline with the human perception. However, these representations are usually too generic and often fail to describe specific content that is present in a sound recording. A lot of work have been devoted to design features that could allow extracting such specific information leading to a wide variety of hand-crafted features. During the past years, owing to the increasing availability of medium scale and large scale sound datasets, an alternative approach to feature extraction has become popular, the so-called feature learning. Finally, processing the amount of data that is at hand nowadays can quickly become overwhelming. It is therefore of paramount importance to be able to reduce the size of the dataset in the feature space. The general processing chain to convert an sound signal to a feature vector that can be efficiently exploited by a classifier and the relation to features used for speech and music processing are described is this chapter. (10.1007/978-3-319-63450-0_4)
  DOI : 10.1007/978-3-319-63450-0_4