Publications

2018

• Hybrid Cryptographic Protocol for Secure Vehicle Data Sharing Over a Consortium Blockchain
  
  • Leo Brousmiche Kei
  • Durand Antoine
  • Heno Thomas
  • Poulain Christian
  • Dalmieres Antoine
  • Ben Hamida Elyes
  , 2018, pp.1281-1286. The blockchain technology has recently attracted increasing interests in a wide range of use-cases. Among those, the management of vehicles' data and life cycle over a blockchain has sparked various research initiatives on a global scale, with the promise to prevent automobile frauds and to enable more collaborations between the involved stakeholders. In this paper, we investigate the problem of securing and sharing vehicles' data over a consortium blockchain, and we describe the architecture of the implemented proof-of-concept. Then, we introduce a novel hybrid cryptographic protocol to secure the access to vehicles' data between the involved stakeholders. Finally, we discuss the lessons learned acquired from the preliminary trials and we highlight the future research challenges and opportunities. (10.1109/Cybermatics_2018.2018.00223)
  DOI : 10.1109/Cybermatics_2018.2018.00223
• Brief Announcement: Performance Prediction for Coarse-Grained Locking
  
  • Aksenov Vitalii
  • Alistarh Dan
  • Kuznetsov Petr
  , 2018. A standard design pattern found in many concurrent data structures , such as hash tables or ordered containers, is an alternation of parallelizable sections that incur no data conflicts and critical sections that must run sequentially and are protected with locks. A lock can be viewed as a queue that arbitrates the order in which the critical sections are executed, and a natural question is whether we can use stochastic analysis to predict the resulting throughput. As a preliminary evidence to the affirmative, we describe a simple model that can be used to predict the throughput of coarse-grained lock-based algorithms. We show that our model works well for CLH lock, and we expect it to work for other popular lock designs such as TTAS, MCS, etc. (10.1145/3212734.3212785)
  DOI : 10.1145/3212734.3212785
• Holonic Cellular Automata: Modelling Multi-level Self-organisation of Structure and Behaviour
  
  • Diaconescu Ada
  • Tomforde Sven
  • Müller-Schloer Christian
  , 2018, pp.186-193. Complex organisms, such as multi-cellular ones, have neither emerged spontaneously, nor evolved directly, from a disorganised mass of quarks. Stable intermediary sub-systems, like atoms and uni-cellular organisms, had to occur first and serve as reusable blocks for more complex systems to build upon. The occurrence of structured systems, featuring internal diversity, from uniform self-adaptive sub-systems is a key phenomenon to study in this context. We believe this phenomenon relies on the interactions among self-adaptive sub-systems, both at the micro-level (directly between sub-systems) but most importantly via macro-levels (indirectly via aggregate information and control from/to all sub-systems). To study this, we have developed a hierarchical control simulator based on self-adaptive cellular automata (CA). This paper presents our Holonic Cellular Automata (HCA) simulator, and the preliminary results showing the occurrence of structure / diversity from micro-macro feedback loops among self-adaptive CAs starting in the same states. This provides a promising basis for further investigations into the range of possibilities concerning structure creation, as a key enabler for the emergence of complex systems. (10.1162/isal_a_00040)
  DOI : 10.1162/isal_a_00040
• Semi-Supervised Deep Attribute Networks for Fine-Grained Ship Category Recognition
  
  • Oliveau Quentin
  • Sahbi Hichem
  , 2018, pp.6871-6874. Classifying ships in satellite or aerial images is a challenging problem in remote sensing imagery. This task requires data-hungry learning algorithms (in particular deep models) that build discriminative representations which capture highly variable ship categories. As labeled training data are scarce and expensive, ship category recognition should also rely on abundant unlabeled data in order to enhance its effectiveness. In this paper, we introduce a novel representation learning algorithm based on semi-supervised attributes. Our method allows us to learn deep and discriminative image characteristics shared among different categories while taking into account both labeled and unlabeled data. We demonstrate the effectiveness of this method on the challenging ship category recognition problem and we show its out-performance with respect to related baselines, especially under the regime of scarce labeled data and abundant unlabeled ones. (10.1109/IGARSS.2018.8517589)
  DOI : 10.1109/IGARSS.2018.8517589
• RABASAR: A FAST RATIO BASED MULTI-TEMPORAL SAR DESPECKLING
  
  • Zhao Weiying
  • Deledalle Charles-Alban
  • Denis Loïc
  • Maître Henri
  • Nicolas Jean-Marie
  • Tupin Florence
  , 2018. In this paper, a generic method is proposed to reduce speckle in multi-temporal stacks of SAR images. The method is based on the computation of a " super-image " , with a large number of looks, by temporal averaging. Then, ratio images are formed by dividing each image of the multi-temporal stack by the " super-image ". In the absence of changes of the ra-diometry, the temporal fluctuations of the intensity at a given spatial location are due to the speckle phenomenon. In areas affected by temporal changes, fluctuations cannot be ascribed to speckle only but also to radiometric changes. The overall effect of the division by the " super-image " is the spatial sta-tionarity improvement: ratio images are much more homogeneous than the original images. Therefore, filtering these ratio images with a speckle-reduction method is more effective, in terms of speckle suppression, than filtering the original multi-temporal stack. After denoising of the ratio image, the de-speckled multi-temporal stack is obtained by multiplication with the " super-image ". Results are presented and analyzed both on synthetic and real SAR data and show the interest of the proposed approach.
• Urban Change Detection for Multispectral Earth Observation Using Convolutional Neural Networks
  
  • Daudt Rodrigo Caye
  • Le Saux Bertrand
  • Boulch Alexandre
  • Gousseau Yann
  , 2018. The Copernicus Sentinel-2 program now provides mul-tispectral images at a global scale with a high revisit rate. In this paper we explore the usage of convolutional neural networks for urban change detection using such multispectral images. We first present the new change detection dataset that was used for training the proposed networks, which will be openly available to serve as a benchmark. The Onera Satellite Change Detection (OSCD) dataset is composed of pairs of multispectral aerial images, and the changes were manually annotated at pixel level. We then propose two architectures to detect changes, Siamese and Early Fusion, and compare the impact of using different numbers of spectral channels as inputs. These architectures are trained from scratch using the provided dataset.
• Challenge codes for physically unclonable functions with Gaussian delays: A maximum entropy problem
  
  • Schaub Alexander
  • Rioul Olivier
  • Boutros Joseph
  • Danger Jean-Luc
  • Guilley Sylvain
  , 2018.
• SAR TOMOGRAPHY OF URBAN AREAS: 3D REGULARIZED INVERSION IN THE SCENE GEOMETRY
  
  • Rambour Clément
  • Denis Loïc
  • Tupin Florence
  • Nicolas Jean-Marie
  • Oriot Hélène
  , 2018. Starting from a stack of co-registered SAR images in inter-ferometric configuration, SAR tomography performs a reconstruction of the reflectivity of scatterers in 3-D. Scatterers seen within the same resolution cell in each SAR image can be separated by jointly unmixing the SAR complex amplitude observed throughout the stack. In urban areas, Compress Sensing (CS) approaches have been applied to achieve super-resolution in the estimation of the position of the scat-terers. However, even if all the local information coming from a stack at a given pixel is used, the structural information that is inherent to the image is not directly used to improve the rendering of the scene. This paper addresses the problem of adding structural constraints to sparse tomographic reconstructions of urban areas. We derive an algorithm allowing the inversion of tomographic data under structural constraints and illustrate its performances on a stack of Spotlight TerraSAR-X images.
• Reinforcement Learning Approaches in Dynamic Environments
  
  • Han Miyoung
  , 2018. Reinforcement learning is learning from interaction with an environment to achieve a goal. It is an efficient framework to solve sequential decision-making problems, using Markov decision processes (MDPs) as a general problem formulation. In this thesis, we apply reinforcement learning to sequential decision-making problems in dynamic environments. We first present an algorithm based on Q-learning with a customized exploration and exploitation strategy to solve a real taxi routing problem. Our algorithm is able to progressively learn optimal actions for routing an autonomous taxi to passenger pick-up points. Then, we address the factored MDP problem in a non-deterministic setting. We propose an algorithm that learns transition functions using the Dynamic Bayesian Network formalism. We demonstrate that factorization methods allow to efficiently learn correct models; through the learned models, the agent can accrue higher cumulative rewards. We extend our work to very large domains. In the focused crawling problem, we propose a new scoring mechanism taking into account long-term effects of selecting a link, and present new feature representations of states for Web pages and actions for next link selection. This approach allowed us to improve on the efficiency of focused crawling. In the influence maximization (IM) problem, we extend the classical IM problem with incomplete knowledge of graph structure and topic-based user interest. Our algorithm finds the most influential seeds to maximize topic-based influence by learning action values for each probed node.
• Descent with Mutations Applied to the Linear Ordering Problem
  
  • Hudry Olivier
  , 2018, pp.253-264.
• Completeness-aware Rule Learning from Knowledge Graphs
  
  • Pellissier Tanon Thomas
  • Stepanova Daria
  • Razniewski Simon
  • Mirza Paramita
  • Weikum Gerhard
  , 2018. Knowledge graphs (KGs) are huge collections of primarily encyclopedic facts that are widely used in entity recognition, structured search, question answering , and other tasks. Rule mining is commonly applied to discover patterns in KGs. However, unlike in traditional association rule mining, KGs provide a setting with a high degree of incompleteness, which may result in the wrong estimation of the quality of mined rules, leading to erroneous beliefs such as all artists have won an award. In this paper we propose to use (in-)completeness meta-information to better assess the quality of rules learned from incomplete KGs. We introduce completeness-aware scoring functions for relational association rules. Experimental evaluation both on real and synthetic datasets shows that the proposed rule ranking approaches have remarkably higher accuracy than the state-of-the-art methods in uncovering missing facts. (10.24963/ijcai.2018/749)
  DOI : 10.24963/ijcai.2018/749
• An Information Theory based Approach to Multisource Clustering
  
  • Murena Pierre-Alexandre
  • Sublime Jérémie
  • Matei Basarab
  • Cornuéjols Antoine
  , 2018. (10.24963/ijcai.2018/358)
  DOI : 10.24963/ijcai.2018/358
• Semiring Provenance over Graph Databases
  
  • Ramusat Yann
  • Maniu Silviu
  • Senellart Pierre
  , 2018. We generalize three existing graph algorithms to compute the provenance of regular path queries over graph databases, in the framework of provenance semirings – algebraic structures that can capture different forms of provenance. Each algorithm yields a different trade-off between time complexity and generality, as each requires different properties over the semiring. Together, these algorithms cover a large class of semirings used for provenance (top-k, security, etc.). Experimental results suggest these approaches are complementary and practical for various kinds of provenance indications, even on a relatively large transport network.
• Celer: a Fast Solver for the Lasso with Dual Extrapolation
  
  • Massias Mathurin
  • Gramfort Alexandre
  • Salmon Joseph
  , 2018, 80, pp.3321-3330. Convex sparsity-inducing regularizations are ubiquitous in high-dimensional machine learning, but solving the resulting optimization problems can be slow. To accelerate solvers, state-of-the-art approaches consist in reducing the size of the optimization problem at hand. In the context of regression, this can be achieved either by discarding irrelevant features (screening techniques) or by prioritizing features likely to be included in the support of the solution (working set techniques). Duality comes into play at several steps in these techniques. Here, we propose an extrapolation technique starting from a sequence of iterates in the dual that leads to the construction of improved dual points. This enables a tighter control of op-timality as used in stopping criterion, as well as better screening performance of Gap Safe rules. Finally, we propose a working set strategy based on an aggressive use of Gap Safe screening rules. Thanks to our new dual point construction, we show significant computational speedups on multiple real-world problems.
• A Probabilistic Theory of Supervised Similarity Learning for Pointwise ROC Curve Optimization
  
  • Vogel Robin
  • Bellet Aurélien
  • Clémençon Stéphan
  , 2018. The performance of many machine learning techniques depends on the choice of an appropriate similarity or distance measure on the input space. Similarity learning (or metric learning) aims at building such a measure from training data so that observations with the same (resp. different) label are as close (resp. far) as possible. In this paper, similarity learning is investigated from the perspective of pairwise bipartite ranking, where the goal is to rank the elements of a database by decreasing order of the probability that they share the same label with some query data point, based on the similarity scores. A natural performance criterion in this setting is pointwise ROC optimization: maximize the true positive rate under a fixed false positive rate. We study this novel perspective on similarity learning through a rigorous probabilistic framework. The empirical version of the problem gives rise to a constrained optimization formulation involving U-statistics, for which we derive universal learning rates as well as faster rates under a noise assumption on the data distribution. We also address the large-scale setting by analyzing the effect of sampling-based approximations. Our theoretical results are supported by illustrative numerical experiments.
• A Probabilistic Theory of Supervised Similarity Learning: Pairwise Bipartite Ranking and Pointwise ROC Curve Optimization
  
  • Clémençon Stéphan
  • Vogel Robin
  • Bellet Aurélien
  , 2018.
• Structured Output Learning with Abstention: Application to Accurate Opinion Prediction
  
  • Garcia Alexandre
  • Essid Slim
  • Clavel Chloé
  • d'Alché-Buc Florence
  , 2018. Motivated by Supervised Opinion Analysis, we propose a novel framework devoted to Structured Output Learning with Abstention (SOLA). The structure prediction model is able to abstain from predicting some labels in the structured output at a cost chosen by the user in a flexible way. For that purpose, we decompose the problem into the learning of a pair of predictors, one devoted to structured abstention and the other, to struc-tured output prediction. To compare fully labeled training data with predictions potentially containing abstentions, we define a wide class of asymmetric abstention-aware losses. Learning is achieved by surrogate regression in an appropriate feature space while prediction with abstention is performed by solving a new pre-image problem. Thus, SOLA extends recent ideas about Struc-tured Output Prediction via surrogate problems and calibration theory and enjoys statistical guarantees on the resulting excess risk. Instantiated on a hierarchical abstention-aware loss, SOLA is shown to be relevant for fine-grained opinion mining and gives state-of-the-art results on this task. Moreover, the abstention-aware representations can be used to competitively predict user-review ratings based on a sentence-level opinion predic-tor.
• Channel Surveillance Strategy and Interference Reduction in Future Wireless Networks
  
  • Ta Duc-Tuyen
  , 2018. The wireless revolution is creating a huge demand for accessing to the radio frequency spectrum with the explosion of the number of connected devices and the large diversity of use cases and requirements. However, the conflict between the spectrum scarcity and the spectrum underutilization leads to significant inefficiencies of wireless communications and impedes the deployment of new applications.Recently, Cognitive Radio (CR) has emerged as a promising technology to address to alleviate the spectrum scarcity and better utilize the spectrum resources by enabling the network users to detect and exploit the spectrum opportunities. The successful deployment of CR networks, however, depends not only on the efficient exploitation of the spectrum opportunities but also on the self-coexistence mechanisms between cognitive users (SUs). The objective of this thesis, therefore, is to provide a systematic study of self-coexistence mechanisms for the cognitive users in both centralized and distributed CR network architecture, which directly address the unaddressed technical challenges of the threat caused by the misbehaving users in the centralized infrastructure networks and the resource allocation issues in the distributed infrastructure networks.
• System-Level Design and Virtual Prototyping of a Telecommunication Application on a NUMA Platform
  
  • Genius Daniela
  • Apvrille Ludovic
  , 2018, pp.1-8. The use of model-driven approaches for embedded system design has become a common practice. Among these model-driven approaches, only a few of them include the generation of a full-system simulation comprising operating system, code generation for tasks and hardware simulation models. Even less common is the extension to massively parallel, NoC based designs, such as required for high performance streaming applications where dozens of tasks are replicated onto identical general purpose processor cores of a Multi-processor System-on-chip (MP-SoC). We present the extension of a system-level tool to handle clustered Network-on-Chip (NoC) with virtual prototyping platforms. On the one hand, the automatic generation of the virtual prototype becomes more complex as topcell, address mapping and linker script have to be adapted. On the other hand, the exploration of the design space is particularly important for this class of applications, as performance may strongly be impacted by Non Uniform Memory Access (NUMA). (10.1109/ReCoSoC.2018.8449375)
  DOI : 10.1109/ReCoSoC.2018.8449375
• Topological Sorting with Regular Constraints
  
  • Amarilli Antoine
  • Paperman Charles
  , 2018, 45 (1), pp.115:1--115:14. We introduce the constrained topological sorting problem (CTS): given a regular language K and a directed acyclic graph G with labeled vertices, determine if G has a topological sort that forms a word in K. This natural problem applies to several settings, e.g., scheduling with costs or verifying concurrent programs. We consider the problem CTS[K] where the target language K is fixed, and study its complexity depending on K. We show that CTS[K] is tractable when K falls in several language families, e.g., unions of monomials, which can be used for pattern matching. However, we show that CTS[K] is NP-hard for K = (ab) * and introduce a shuffle reduction technique to show hardness for more languages. We also study the special case of the constrained shuffle problem (CSh), where the input graph is a disjoint union of strings, and show that CSh[K] is additionally tractable when K is a group language or a union of district group monomials. We conjecture that a dichotomy should hold on the complexity of CTS[K] or CSh[K] depending on K, and substantiate this by proving a coarser dichotomy under a different problem phrasing which ensures that tractable languages are closed under common operators.
• Public Privacy and Brick Houses Made of Glass
  
  • Marsh Stephen
  • Diaconescu Ada
  • Evans David
  • Kosa Tracy Ann
  • Lewis Peter R.
  • Mahbub Habib Sheikh
  , 2018, AICT-528, pp.137-148. In this work in progress paper, we present a description of a new view of privacy in public, examining how it is possible to ascertain the privacy levels of individuals in context and in groups, and different ways of visualising these Public Privacy levels. We examine how awareness of one’s Public Privacy may have an impact on behaviour and privacy protection in general, and propose future work to examine the concept in more detail. (10.1007/978-3-319-95276-5_10)
  DOI : 10.1007/978-3-319-95276-5_10
• Automating the Production of Communicative Gestures in Embodied Characters
  
  • Ravenet Brian
  • Pelachaud Catherine
  • Clavel Chloé
  • Marsella Stacy
  Frontiers in Psychology, Frontiers Media, 2018, 9, pp.1-12. In this paper we highlight the different challenges in modeling communicative gestures for Embodied Conversational Agents (ECAs). We describe models whose aim is to capture and understand the specific characteristics of communicative gestures in order to envision how an automatic communicative gesture production mechanism could be built. The work is inspired by research on how human gesture characteristics (e.g., shape of the hand, movement, orientation and timing with respect to the speech) convey meaning. We present approaches to computing where to place a gesture, which shape the gesture takes and how gesture shapes evolve through time. We focus on a particular model based on theoretical frameworks on metaphors and embodied cognition that argue that people can represent, reason about and convey abstract concepts using physical representations and processes, which can be conveyed through physical gestures. (10.3389/fpsyg.2018.01144)
  DOI : 10.3389/fpsyg.2018.01144
• Adaptive Multiplicity Codes based PIR Protocol for Multi-Cloud Plateform Services
  
  • Salaun Lou
  • Alloum Amira
  • Jacquet Philippe
  , 2018, pp.1-8. Our contribution consists in deriving an adaptive multiplicity code based PIR protocol based on a code selection algorithm which guarantees minimal communication overhead for a given system architecture. We formulate the related constrained optimization problem, analyze it and introduce an algorithm for enabling the adaptive Information Theoretical secure PIR protocol to operate in highly dynamic multi cloud platform services. In addition, we prove that this algorithm also solves the feasibility problem and achieves optimal solution. (10.1109/5GWF.2018.8517057)
  DOI : 10.1109/5GWF.2018.8517057
• On-Wafer CPW Standards for S-Parameter Measurements of Balanced Circuits Up to 40 GHz
  
  • Pham Thi Dao
  • Allal Djamel
  • Ziade Francois
  • Bergeault Eric
  , 2018, pp.1-2. The Multimode Thru-Reflect-Line (TRL) calibration technique is applied to perform mixed-mode scattering parameter measurements of on-wafer differential circuits. The paper describes the design and the realization of coupled coplanar waveguide (CCPW) standards on quartz (Si O2 ) substrate in the Ground-Signal-Ground-Signal-Ground (GSGSG) configuration. Simulation and measurement results for a mismatched transmission line demonstrate the validity of the method up to 40 GHz. (10.1109/CPEM.2018.8500997)
  DOI : 10.1109/CPEM.2018.8500997
• MRAM-on-FDSOI Integration: A Bit-Cell Perspective
  
  • Cai Hao
  • Wang You
  • Kang Wang
  • Naviner Lirida
  • Liu Xinning
  • Yang Jun
  • Zhao Weisheng
  , 2018, pp.263-268. In this paper we discuss the potential foundry announced hybrid integration of magnetic random access memory (MRAM) on fully depleted silicon-on-insulator (FD-SOI) technology. The spin transfer torque magnetic tunnel junction (STT-MTJ) and the next generation voltage-controlled magnetic anisotropy (VCMA) MTJ are separately integrated into a 28 nm FD-SOI process. Circuit-level design strategies are explored that use FD-SOI leverage and spin-device characteristic to realize writing and reading power-delay efficiency, robust and reliable performance in a 1-transistor 1-MTJ (1T1M) bit cell. Process variation aware strategies for MTJ-FDSOI integration are proposed to compensate failure operations, by using the dynamic step-wise back-bias and the flip-well back-bias. A qualitative summary demonstrates that the MRAM-on-FDSOI integration offers attractive performance for future non-volatile CMOS integration. (10.1109/ISVLSI.2018.00056)
  DOI : 10.1109/ISVLSI.2018.00056