Publications

2019

• Boosting decision stumps for dynamic feature selection on data streams
  
  • Barddal Jean Paul
  • Enembreck Fabrició
  • Gomes Heitor Murilo
  • Bifet Albert
  • Pfahringer Bernhard
  Information Systems, Elsevier, 2019, 83, pp.13-29. Feature selection targets the identification of which features of a dataset are relevant to the learning task. It is also widely known and used to improve computation times, reduce computation requirements, and to decrease the impact of the curse of dimensionality and enhancing the generalization rates of classifiers. In data streams, classifiers shall benefit from all the items above, but more importantly, from the fact that the relevant subset of features may drift over time. In this paper, we propose a novel dynamic feature selection method for data streams called Adaptive Boosting for Feature Selection (ABFS). ABFS chains decision stumps and drift detectors, and as a result, identifies which features are relevant to the learning task as the stream progresses with reasonable success. In addition to our proposed algorithm, we bring feature selection-specific metrics from batch learning to streaming scenarios. Next, we evaluate ABFS according to these metrics in both synthetic and real-world scenarios. As a result, ABFS improves the classification rates of different types of learners and eventually enhances computational resources usage. (10.1016/j.is.2019.02.003)
  DOI : 10.1016/j.is.2019.02.003
• Learning and Data Selection in Big Datasets
  
  • Ghadikolaei Hossein S
  • Ghauch Hadi
  • Fischione Carlo
  • Skoglund Mikael
  , 2019. Finding a dataset of minimal cardinality to characterize the optimal parameters of a model is of paramount importance in machine learning and distributed optimization over a network. This paper investigates the compressibility of large datasets. More specifically, we propose a framework that jointly learns the input-output mapping as well as the most representative samples of the dataset (sufficient dataset). Our analytical results show that the cardinality of the sufficient dataset increases sub-linearly with respect to the original dataset size. Numerical evaluations of real datasets reveal a large compressibility, up to 95%, without a noticeable drop in the learnability performance, measured by the generalization error.
• Hypothesis Testing Over the Two-Hop Relay Network
  
  • Salehkalaibar Sadaf
  • Wigger Michèle
  • Wang Ligong
  IEEE Transactions on Information Theory, Institute of Electrical and Electronics Engineers, 2019, 65 (7). Coding and testing schemes and the corresponding achievable type-II error exponents are presented for binary hypothesis testing over two-hop relay networks. The schemes are based on cascade source coding techniques and unanimous decision-forwarding, the latter meaning that a terminal decides on the null hypothesis only if all previous terminals have decided on the null hypothesis. If the observations at the transmitter, the relay, and the receiver form a Markov chain in this order, then, without loss in performance, the proposed cascade source code can be replaced by two independent point-to-point source codes, one for each hop. The decoupled scheme (combined with decision-forwarding) is shown to attain the optimal type-II error exponents for various instances of "testing against conditional independence." The same decoupling is shown to be optimal also for some instances of "testing against independence," when the observations at the transmitter, the receiver, and the relay form a Markov chain in this order and when the relay-to-receiver link is of sufficiently high rate. For completeness, this paper also presents an analysis of the Shimokawa-Han-Amari binning scheme for the point-to-point hypothesis testing setup. (10.1109/TIT.2019.2897698)
  DOI : 10.1109/TIT.2019.2897698
• GÉNÉRATION D'ONDES CARRÉES DANS UN LASER À CASCADES QUANTIQUES SOUMIS À UNE CONTRE-RÉACTION OPTIQUE EXTERNE ET À UNE ROTATION DE POLARISATION
  
  • Spitz O
  • Herdt A
  • Carras M.
  • Elsässer W
  • Grillot F.
  , 2019. Les lasers à cascades quantiques sont connus pour leur dynamique non-linéaire lorsqu'ils sont soumis à une contre-réaction optique. Nous montrons que l'ajout d'une rotation de polarisation peut provoquer un autre phénomène non-linéaire, à savoir une onde carrée, avec l'apparition d'une onde transverse électrique en plus de celle transverse magnétique.
• DC coefficient recovery for JPEG images in ubiquitous communication systems
  
  • Qiu Han
  • Memmi Gérard
  • Chen Xuan
  • Xiong Jian
  Future Generation Computer Systems, Elsevier, 2019, 96, pp.23-31. (10.1016/j.future.2019.01.037)
  DOI : 10.1016/j.future.2019.01.037
• Real-time scene analysis for 3D interaction
  
  • Kaiser Adrien
  , 2019. This PhD thesis focuses on the problem of visual scene analysis captured by commodity depth sensors to convert their data into high level understanding of the scene. It explores the use of 3D geometry analysis tools on visual depth data in terms of enhancement, registration and consolidation. In particular, we aim to show how shape abstraction can generate lightweight representations of the data for fast analysis with low hardware requirements. This last property is important as one of our goals is to design algorithms suitable for live embedded operation in e.g., wearable devices, smartphones or mobile robots. The context of this thesis is the live operation of 3D interaction on a mobile device, which raises numerous issues including placing 3D interaction zones with relation to real surrounding objects, tracking the interaction zones in space when the sensor moves and providing a meaningful and understandable experience to non-expert users. Towards solving these problems, we make contributions where scene abstraction leads to fast and robust sensor localization as well as efficient frame data representation, enhancement and consolidation. While simple geometric surface shapes are not as faithful as heavy point sets or volumes to represent observed scenes, we show that they are an acceptable approximation and their light weight makes them well balanced between accuracy and performance.
• Energy Efficient Resource Allocation for Type-I HARQ Under the Rician Channel
  
  • Leturc Xavier
  • Ciblat Philippe
  • Le Martret Christophe
  IEEE Transactions on Wireless Communications, Institute of Electrical and Electronics Engineers, 2019. This paper addresses the per-link power and bandwidth allocation problem with the objective of maximizing energy efficiency (EE) related metrics under a per-link minimum goodput constraint when only statistical channel state information is available. We consider a parallel (i.e., without multiuser interference) Rician channel model, which encompasses both Rayleigh and additive white Gaussian noise channels as special cases. We also consider Type-I hybrid automatic repeat request with practical modulation and coding schemes. The addressed problems are the maximization of the sum of the user's EE, the maximization of the EE of the user with the lowest EE and the maximization of the EE of the network. We derive the optimal solutions of these problems in closed form using fractional programming and a convex optimization framework. We show that substantial gains can be achieved by taking into account the line of sight between the transmitter and the receiver instead of only considering the average channel power. (10.1109/TWC.2019.2918534)
  DOI : 10.1109/TWC.2019.2918534
• DEVICES AND METHODS FOR LATTICE POINTS ENUMERATION
  
  • Rekaya-Ben Othman Ghaya
  • Askri Aymen
  , 2019, pp.29. A lattice prediction device (200) for predicting a number of lattice points falling inside a bounded region in a given vector space. The bounded region is defined by a radius value, 5 a lattice point representing a digital signal in a lattice constructed over the vector space. The lattice is defined by a lattice generator matrix comprising components. The lattice prediction device (200) comprises a computation unit (201) configured to determine a predicted number of lattice points by applying a machine learning algorithm to input data derived from the radius value and the components of lattice generator matrix.
• Quantum cascade laser technology and applications at mirSense, from spectroscopy to chaotic communication
  
  • Carras Mathieu
  • Aoust Guillaume
  • Maisons Grégory
  • Brun Mickael
  • Spitz Olivier
  • Grillot Frederic
  , 2019. Quantum cascade lasers (QCLs) have been deeply studied structures. We’ll discuss the level of maturity of this technology. In particular we will highlight the main challenges in material production, including growth and process. We’ll show latest results in development of gas sensors at mirSense, as well as products for defense applications. We’ll show why photoacoustic is a perfect match for QCLs, and explore the fundamentals of this sensing technology. Then we’ll say a word on more advanced concepts using such laser technology like chaotic communication for which QCLs have been recently studied.
• Enumeration on Trees with Tractable Combined Complexity and Efficient Updates
  
  • Amarilli Antoine
  • Bourhis Pierre
  • Mengel Stefan
  • Niewerth Matthias
  , 2019, pp.89-103. (10.1145/3294052.3319702)
  DOI : 10.1145/3294052.3319702
• Projet Data&Musée Représentations sémantiques et leur exploitation pour le traitement de données collectées dans des musées et monuments
  
  • Moissinac Jean-Claude Jc
  • Mimouni Nada
  , 2019. Data&Musée collecte de multiple données d'une centaines de partenaires, musées et monuments: billetterie, livres d'or, événements,... Nous avons fait le choix d'une représentation RDF de ces données pour faciliter les liens avec divers jeux de données: DBpedia, Wikidata, Joconde, DataTourisme ... Le but est d'entreprendre des explorations guidées par la sémantique des grands graphes que nous constituons. Nous présentons ici le contexte de Data&Musée et les principaux choix que nous faisons liés aux représentations sémantiques.
• Let there be Chaining: How to Augment your IGP to Chain your Services
  
  • Wion Adrien
  • Bouet Mathieu
  • Iannone Luigi
  • Conan Vania
  , 2019. Ever since Network Functions Virtualization has replaced dedicated appliances, ISPs have been able to add a degree of flexibility in their traffic engineering. However, it also has increased the complexity of the optimization problem, because it is now necessary to place virtual functions and route traffic jointly. Insofar, a logically centralized approach has been taken, where a so-called orchestrator, having full knowledge of the network, the virtual functions, and the traffic, run complex algorithms to find a suitable solution to the problem. The outcome of the algorithms are then translated to network configurations to be pushed to all of the appliances. We argue that there is no need to fully centralize every decision, rather we can leverage existing network intelligence to achieve the same goal. In particular we propose to augment the routing layer with the notion of services, so to rely on the robustness and scalability of Interior Gateway Protocols (IGP). Our solution leverages on existing distributed routing protocols where, in addition, autonomous nodes announce information about the virtual services they provide. Our design is modular and incrementally deployable and has been implemented in what we call a NFV Router. In our evaluation, we show that (i) NFV Routers distributed chaining decisions are close to optimal centrally-computed paths, (ii) on a large scale testbed deployment, NFV Routers efficiently steer traffic through chains and only add a small overhead to control traffic and (iii) our distributed system, because of its local control loop, has a faster reaction to network events than centralized solutions.
• The effect of ramp constraints on coalitional storage games
  
  • Kiedanski Diego
  • Orda Ariel
  • Kofman Daniel
  , 2019, pp.226-238. (10.1145/3307772.3328300)
  DOI : 10.1145/3307772.3328300
• Main effects and interactions in mixed and incomplete data frames
  
  • Robin Geneviève
  • Klopp Olga
  • Josse Julie
  • Moulines Éric
  • Tibshirani Robert
  Journal of the American Statistical Association, Taylor & Francis, 2019. A mixed data frame (MDF) is a table collecting categorical, numerical and count observations. The use of MDF is widespread in statistics and the applications are numerous from abundance data in ecology to recommender systems. In many cases, an MDF exhibits simultaneously main effects, such as row, column or group effects and interactions, for which a low-rank model has often been suggested. Although the literature on low-rank approximations is very substantial, with few exceptions, existing methods do not allow to incorporate main effects and interactions while providing statistical guarantees. The present work fills this gap. * This work has been funded by the DataScience Inititiative (Ecole Polytechnique) and the Russian Academic Excellence Project '5-100 (10.1080/01621459.2019.1623041)
  DOI : 10.1080/01621459.2019.1623041
• A Dual-polarization Rayleigh Backscatter Model for Phasesensitive OTDR Applications
  
  • Guerrier Sterenn
  • Dorize Christian
  • Awwad Elie
  • Renaudier Jeremie
  , 2019, pp.ETu3A.4. (10.1364/ES.2019.ETu3A.4)
  DOI : 10.1364/ES.2019.ETu3A.4
• Principal Component Analysis for Multivariate Extremes
  
  • Drees Holger
  • Sabourin Anne
  , 2019. The first order behavior of multivariate heavy-tailed random vectors above large radial thresholds is ruled by a limit measure in a regular variation framework. For a high dimensional vector, a reasonable assumption is that the support of this measure is concentrated on a lower dimensional subspace, meaning that certain linear combinations of the components are much likelier to be large than others. Identifying this subspace and thus reducing the dimension will facilitate a refined statistical analysis. In this work we apply Principal Component Analysis (PCA) to a re-scaled version of radially thresholded observations. Within the statistical learning framework of empirical risk minimization, our main focus is to analyze the squared reconstruction error for the exceedances over large radial thresholds. We prove that the empirical risk converges to the true risk, uniformly over all projection subspaces. As a consequence, the best projection subspace is shown to converge in probability to the optimal one, in terms of the Hausdorff distance between their intersections with the unit sphere. In addition, if the exceedances are re-scaled to the unit ball, we obtain finite sample uniform guarantees to the reconstruction error pertaining to the estimated projection sub-space. Numerical experiments illustrate the relevance of the proposed framework for practical purposes.
• Optimizing System Architecture Cost and Security Countermeasures
  
  • Berro Sahar
  • Apvrille Ludovic
  • Duc Guillaume
  , 2019. The design of an embedded system is built on a trade-off between its performance and its cost. Nowadays, the security threats that target most of the embedded systems introduce a new factor in this trade-off: the security level of the system. So system architects must consider , during the design, the different attacks that target the system and the possible countermeasures, and their costs. In this article, we present a methodology to help designers explore different countermeasures and evaluate their impact on the cost of the architecture and the probability of success of an adversary. This methodology is based on extended and formalized Attack-Defense Trees that allow to assess the impact of countermeasures on system components and attacks. We use propagation rules to characterize a main attack from its different steps, and we formalize the trade-off between security and cost by an optimization problem between attack probability and total architecture cost.
• Temporal characterization of an urban horizontal atmospheric telecom channel
  
  • Sauvage Chloé
  • Robert Clélia
  • Sorrente Béatrice
  • Erasme Didier
  , 2019, pp.PW4C.5. Free Space Optics (FSO) are breakable under some climatic conditions. However characterization of the propagation channel by studying wavelength transmisttance and data coming from a wavefront experiment could improve FSO's performance. (10.1364/PCAOP.2019.PW4C.5)
  DOI : 10.1364/PCAOP.2019.PW4C.5
• Introducing Innovative Bare Metal Crypto Terminal for Blockchains and BigBang Paradigm
  
  • Urien Pascal
  , 2019, pp.1-4. (10.1109/NTMS.2019.8763823)
  DOI : 10.1109/NTMS.2019.8763823
• A benchmarking methodology for evaluating software switch performance for NFV
  
  • Zhang Tianzhu
  • Linguaglossa Leonardo
  • Roberts James
  • Iannone Luigi
  • Gallo Massimo
  • Giaccone Paolo
  , 2019. Interest in software networking has grown significantly since the introduction of Network Function Virtualization (NFV). Software switches are used in NFV to steer traffic between different virtualized network functions and physical Network Interface Cards (NICs). It is becoming more and more important to objectively evaluate and compare the performance of the multiple alternative implementations that have recently been proposed. A comprehensive performance analysis is still missing for two main reasons: (i) the amount of time required to configure and compare all such tools is enormous; (ii) it is very difficult to define a proper methodology to compare different solutions in a fair manner. In this paper, we propose a methodology based on four simple yet representative test scenarios used to evaluate the performance of software switches. We apply this methodology to measure throughput and latency metrics for 6 state-of-the-art software switches namely, OVS-DPDK, snabb, BESS, FastClick, VPP and netmap VALE. Our work constitutes a first step to building a better understanding of design tradeoffs and identifying performance bottlenecks.
• A statistical detection mechanism for node misbehaviours in wireless mesh networks
  
  • Chbib Fadlallah
  • Khoukhi Lyes
  • Fahs Walid
  • Khatoun Rida
  • Haydar Jamal
  , 2019, 31 (1), pp.23.
• Wave Performance Analysis and Enhancement for Safety Applications in Vehicular Networks
  
  • Chbib Fadlallah
  • Khoukhi Lyes
  • Fahs Walid
  • Khatoun Rida
  • Haydar Jamal
  , 2019, pp.1-7. (10.1109/NTMS.2019.8763783)
  DOI : 10.1109/NTMS.2019.8763783
• Teaching C Programming Interactively at Scale Using Taskgrader
  
  • Sharrock Rémi
  • Bonfert-Taylor Petra
  • Hiron Mathias
  • Blockelet Michel
  • Miller Chris
  • Goudzwaard Mike
  • Hamonic Ella
  , 2019, pp.1-2. This demo paper introduces a tool and a method to provide a barriers-free, rich, interactive learning experience for students of all levels of preparation in programming courses. Taskgrader is an open-source autograding tool providing instant feedback in large-scale online programming classes. This in-browser tool offers extensive feedback to student code submissions right within any LMS and pass data back to the gradebook. (10.1145/3330430.3333670)
  DOI : 10.1145/3330430.3333670
• Secure Data Sharing with Fast Access Revocation through Untrusted Clouds
  
  • Kapusta Katarzyna
  • Qiu Han
  • Memmi Gérard
  , 2019, pp.1-5. (10.1109/NTMS.2019.8763850)
  DOI : 10.1109/NTMS.2019.8763850
• TMA 2019 - Proceedings of the 3rd Network Traffic Measurement and Analysis Conference
  
  • Secci Stefano
  • Chrisment Isabelle
  • Fiore Marco
  • Tabourier Lionel
  • Lim Keun-Woo
  , 2019.