Publications

2019

• Random Forest Oriented Fast QTBT Frame Partitioning
  
  • Amestoy Thomas
  • Mercat Alexandre
  • Hamidouche Wassim
  • Bergeron Cyril
  • Menard Daniel
  , 2019, pp.1837-1841. (10.1109/icassp.2019.8683413)
  DOI : 10.1109/icassp.2019.8683413
• Compression Improvement via Reference Organization for 2D-multiview Content
  
  • Nikitin Pavel
  • Cagnazzo Marco
  • Jung Joel
  , 2019, pp.1612-1616. (10.1109/ICASSP.2019.8682999)
  DOI : 10.1109/ICASSP.2019.8682999
• Speech enhancement with variational autoencoders and alpha-stable distributions
  
  • Leglaive Simon
  • Şimşekli Umut
  • Liutkus Antoine
  • Girin Laurent
  • Horaud Radu
  , 2019, pp.541-545. his paper focuses on single-channel semi-supervised speech en-hancement. We learn a speaker-independent deep generative speechmodel using the framework of variational autoencoders. The noisemodel remains unsupervised because we do not assume prior knowl-edge of the noisy recording environment. In this context, our con-tribution is to propose a noise model based on alpha-stable distribu-tions, instead of the more conventional Gaussian non-negative ma-trix factorization approach found in previous studies. We develop aMonte Carlo expectation-maximization algorithm for estimating themodel parameters at test time. Experimental results show the supe-riority of the proposed approach both in terms of perceptual qualityand intelligibility of the enhanced speech signal. (10.1109/ICASSP.2019.8682546)
  DOI : 10.1109/ICASSP.2019.8682546
• Enhancing HEVC Spatial Prediction by Context-based Learning
  
  • Wang Li
  • Fiandrotti Attilio
  • Purica Andrei
  • Valenzise Giuseppe
  • Cagnazzo Marco
  , 2019, pp.4035-4039. Deep generative models have been recently employed to compress images, image residuals or to predict image regions. Based on the observation that state-of-the-art spatial prediction is highly optimized from a rate-distortion point of view, in this work we study how learning-based approaches might be used to further enhance this prediction. To this end, we propose an encoder-decoder convolutional network able to reduce the energy of the residuals of HEVC intra prediction, by leveraging the available context of previously decoded neighboring blocks. The proposed context-based prediction enhancement (CBPE) scheme enables to reduce the mean square error of HEVC prediction by 25% on average, without any additional signalling cost in the bitstream. (10.1109/icassp.2019.8683624)
  DOI : 10.1109/icassp.2019.8683624
• Singing Voice Separation: A Study on Training Data
  
  • Prétet Laure
  • Hennequin Romain
  • Royo-Letelier Jimena
  • Vaglio Andrea
  , 2019, pp.506-510. In the recent years, singing voice separation systems showed increased performance due to the use of supervised training. The design of training datasets is known as a crucial factor in the performance of such systems. We investigate on how the characteristics of the training dataset impacts the separation performances of state-of-the-art singing voice separation algorithms. We show that the separation quality and diversity are two important and complementary assets of a good training dataset. We also provide insights on possible transforms to perform data augmentation for this task. (10.1109/ICASSP.2019.8683555)
  DOI : 10.1109/ICASSP.2019.8683555
• Architecture Models Refinements for Software Development of Critical Real-time Embedded Systems
  
  • Borde Etienne
  , 2019. Cyber Physical Systems are systems controlled or monitored by computer-based programs, tightly integrated networks, sensors, and actuators. Trains, aircrafts, cars, and some medical equipments are examples of complex CPS. Software development of complex CPS has become so difficult that it represents most of the cost of CPS production. According to domain experts, this trend is going to reach a point where software development would represent the main source of cost of a CPS production. In addition, it is interesting to note that the integration, verification and validation of software in CPS require more efforts than the analysis, design, and implementation activities. The main reason is that these activities are conducted late in the development process and issues discovered at this stage of the process will require to rework artifacts produced in the previous activities (i.e. analysis, design and/or implementation). In this document, we present our work aiming to improve the reliability of software development in the domain of CPS. In this context, we define the reliability of the development process as its capacity to deliver intermediate artifacts for which the rework effort would be as small as possible. This problem is very difficult for general purpose software (i.e. used on desktop computers or servers), and even more difficult for software in CPS. The main reason is that software in CPS is often critical, real-time and embedded on domain specific execution platforms. As a consequence, non-functional properties (also called quality attributes) of software applications in CPS are often as important and difficult to satisfy as the logical correctness of these applications. In order to the improve the reliability of software development in the domain of CPS, we propose a Model Driven Engineering (MDE) method based on step-wise refinements of software architecture descriptions (also called architectural models). An architecture description being an abstraction of the software being developed, the implementation of this software (i.e. source or binary code) is an interpretation of the architecture model. In the framework we propose, such interpretations are automated using model refinements, i.e. model to model transformations lowering the abstraction level of the architecture description. However, models interpretation may introduce faults such as bugs or invalidation of nonfunctional requirements. It is hence necessary to control as much as possible the correctness, consistency, and optimality of artifacts produced along the model refinement steps. To reach this objective, we propose to 1. define model transformations so as to interleave refinement steps with analysis of the resulting artifacts. We thus improve the consistency between the analysis results and the software implementation by analyzing models as close as possible to the implementation. 2. define timing analysis and real-time scheduling techniques to ensure the correctness of software architectures from a timing perspective. 3. formalize model transformations in order to ensure their correctness using formal verification techniques. 4. compose model transformations in order to automate the search for optimal (or nearoptimal) architectures. The work presented in this document is thus at the frontier among different research domains: MDE, real-time systems scheduling, formal verification, and operational research. In this work, we chose to rely and extend the Architecture Analysis and Design Language (AADL) to model the cyber part of CPS. The reasons for this choice are simple: Firstly, AADL is a standard and a domain specific language for real-time embedded systems. Secondly, It allows to represent software architectures with different abstraction levels. Last but not least, AADL supports different types of models of computations communications, some of which being deterministic. As a guideline for our work, we developed the methodology we propose in a MDE framework called RAMSES (Refinement of AADL Models for the Synthesis of Embedded Systems). This document presents both the methodology and some illustrations of its implementation in RAMSES.
• Efficient approximate unitary t-designs from partially invertible universal sets and their application to quantum speedup
  
  • Mezher Rawad
  • Ghalbouni Joe
  • Dgheim Joseph
  • Markham Damian
  , 2019. At its core a $t$-design is a method for sampling from a set of unitaries in a way which mimics sampling randomly from the Haar measure on the unitary group, with applications across quantum information processing and physics. We construct new families of quantum circuits on $n$-qubits giving rise to $\varepsilon$-approximate unitary $t$-designs efficiently in $O(n^3t^2)$ depth. These quantum circuits are based on a relaxation of technical requirements in previous constructions. In particular, the construction of circuits which give efficient approximate $t$-designs by Brandao, Harrow, and Horodecki (F.G.S.L Brandao, A.W Harrow, and M. Horodecki, Commun. Math. Phys. (2016).) required choosing gates from ensembles which contained inverses for all elements, and that the entries of the unitaries are algebraic. We reduce these requirements, to sets that contain elements without inverses in the set, and non-algebraic entries, which we dub partially invertible universal sets. We then adapt this circuit construction to the framework of measurement based quantum computation(MBQC) and give new explicit examples of $n$-qubit graph states with fixed assignments of measurements (graph gadgets) giving rise to unitary $t$-designs based on partially invertible universal sets, in a natural way. We further show that these graph gadgets demonstrate a quantum speedup, up to standard complexity theoretic conjectures. We provide numerical and analytical evidence that almost any assignment of fixed measurement angles on an $n$-qubit cluster state give efficient $t$-designs and demonstrate a quantum speedup.
• Laser-induced Single-bit Faults in Flash Memory: Instructions Corruption on a 32-bit Microcontroller
  
  • Colombier Brice
  • Menu Alexandre
  • Dutertre Jean-Max
  • Moëllic Pierre-Alain
  • Rigaud Jean-Baptiste
  • Danger Jean-Luc
  , 2019, pp.1-10. Physical attacks are a known threat posed against secure embedded systems. Notable among these is laser fault injection, which is often considered as the most effective fault injection technique. Indeed, laser fault injection provides a high spatial accuracy, which enables an attacker to induce bit-level faults. However, experience gained from attacking 8-bit targets might not be relevant on more advanced micro-architectures, and these attacks become increasingly challenging on 32-bit microcontrollers. In this article, we show that the flash memory area of a 32-bit microcontroller is sensitive to laser fault injection. These faults occur during the instruction fetch process, hence the stored value remains unaltered. After a thorough characterisation of the induced faults and the associated fault model, we provide detailed examples of bit- level corruption of instructions and demonstrate practical applications in compromising the security of real-life codes. Based on these experimental results, we formulate a hy- pothesis about the underlying micro-architectural features that explain the observed fault model. (10.1109/HST.2019.8741030)
  DOI : 10.1109/HST.2019.8741030
• Leveraging Body Interactions to Support Immersive Analytics
  
  • Fruchard Bruno
  • Prouzeau Arnaud
  • Chapuis Olivier
  • Lecolinet Eric
  , 2019, pp.10 pages. New immersive devices (e.g., virtual or augmented reality) enable displaying large amounts of data in space to better support data analysis. Manipulating this data efficiently is crucial, but challenging because the user must be able to activate various commands or adjust various values while remaining free to move. Using the whole body offers several valuable advantages: 1) The body provides a physical support as an interactive surface, which improves accuracy and makes it less tiring to interact; 2) Using the body does not impair mobility and avoids handling devices; 3) Proprioception makes it possible to interact eyes-free, including for choosing values in a range; 4) By leveraging spatial memory, the body helps memorizing commands, thus interacting in expert mode (i.e., perform quick actions without visual feedback). In this position paper, we analyze various ways of interacting with the body and discuss their advantages and challenges for immersive analytics.
• Understanding alternatives in data analysis activities
  
  • Liu Jiali
  • Boukhelifa Nadia
  • Eagan James
  , 2019, pp.5. Data workers are non-professional data scientists who engage in data analysis activities as part of their daily work. In this position paper, we share past and on-going work to understand data workers’ sense-making practices. We use multidisciplinary approaches to explore their human-tool partnerships. We introduce our current research on the role of alternatives in data analysis activities. Finally, we conclude with open questions and research directions.
• Information theory: An analysis and design tool for HCI
  
  • Liu Wanyu
  • Oulasvirta Antti
  • Rioul Olivier
  • Beaudouin-Lafon Michel
  • Guiard Yves
  , 2019. Shannon’s information theory, since its first introduction in 1948, has received much attention and successful applications in many domains. Apart from Fitts’ law and Hick’s law, which came out when experimental psychologists were enthusiastic about applying information theory to various areas of psychology, the relation of information theory to human-computer interaction (HCI) has not been clear. Even the two above-mentioned “laws” remain controversial in both psychology and HCI. On the other hand, in recent years, information theory has started to directly inspire or contribute to HCI research.
• A LSTM Approach to Detection of Autonomous Vehicle Hijacking
  
  • Singh Negi Naman
  • Jelassi Ons
  • Clémençon Stéphan
  • Fischmeister Sebastian
  , 2019, 1, pp.475-482. In the recent decades, automotive research has been focused on creating a driverless future. Autonomous vehicles are expected to take over tasks which are dull, dirty and dangerous for humans (3Ds of robotization). However, augmented autonomy increases reliance on the robustness of the system. Autonomous vehicle systems are heavily focused on data acquisition in order to perceive the driving environment accurately. In the future, a typical autonomous vehicle data ecosystem will include data from internal sensors, infrastructure, communication with nearby vehicles, and other sources. Physical faults, malicious attacks or a misbehaving vehicle can result in the incorrect perception of the environment, which can in turn lead to task failure or accidents. Anomaly detection is hence expected to play a critical role improving the security and efficiency of autonomous and connected vehicles. Anomaly detection can be defined as a way of identifying unusual or unexpected events and/or measurements. In this paper, we focus on the specific case of malicious attack/hijacking of the system which results in unpredictable evolution of the autonomous vehicle. We use a Long Short-Term Memory (LSTM) network for anomaly/fault detection. It is, first, trained on non-abnormal data to understand the system's baseline performance and behaviour, monitored through three vehicle control parameters namely velocity, acceleration and jerk. Then, the model is used to predict over a number of future time steps and an alarm is raised as soon as the observed behaviour of the autonomous car significantly deviates from the prediction. The relevance of this approach is supported by numerical experiments based on data produced by an autonomous car simulator, capable of generating attacks on the system. (10.5220/0007726004750482)
  DOI : 10.5220/0007726004750482
• Procédé et système pour traiter un message transmis à un véhicule automobile par une entité communicante distante
  
  • Monteuuis Jean-Philippe
  • Labiod Houda
  • Zhang Jun J.
  • Servel Alain
  • Mafrica Stefano
  , 2019.
• Integrity Issues for IoT: From Experiment to Classification Introducing Integrity Probes
  
  • Urien Pascal
  , 2019, pp.344-350. (10.5220/0007746903440350)
  DOI : 10.5220/0007746903440350
• Anarchic Urban Expansion Detection and Monitoring with Integration of Expert Knowledge
  
  • Chaabane F.
  • Réjichi S.
  • Kefi Chayma
  • Haytem Ismail
  • Tupin Florence
  , 2019.
• From Reflex to Reflection: Two Tricks AI Could Learn from Us
  
  • Dessalles Jean-Louis
  Philosophies, MDPI, 2019, 4 (2), pp.27. Deep learning and other similar machine learning techniques have a huge advantage over other AI methods: they do function when applied to real-world data, ideally from scratch, without human intervention. However, they have several shortcomings that mere quantitative progress is unlikely to overcome. The paper analyses these shortcomings as resulting from the type of compression achieved by these techniques, which is limited to statistical compression. Two directions for qualitative improvement, inspired by comparison with cognitive processes, are proposed here, in the form of two mechanisms: complexity drop and contrast. These mechanisms are supposed to operate dynamically and not through pre-processing as in neural networks. Their introduction may bring the functioning of AI away from mere reflex and closer to reflection. (10.3390/philosophies4020027)
  DOI : 10.3390/philosophies4020027
• Antenne sinueuse à ondes de fuite pour applications millimétriques en bande W
  
  • Bernabeu-Jimenez Tomas
  • Begaud Xavier
  • Magne François
  • Hadjloum Massinissa
  • Cosson Bruno
  , 2019.
• De Fourier à la 5G
  
  • Ciblat Philippe
  • Debbah Merouane
  Interstices, INRIA, 2019.
• Étude et Dimensionnement d’un Amplificateur de Puissance de Type Outphasing
  
  • Bachi Joe
  • Serhan Ayssar
  • Pham Dang-Kièn Germain
  • Desgreys Patricia
  • Giry Alexandre
  , 2019.
• On the Performance of Spatial Modulations Over Multimode Optical Fiber Transmission Channels
  
  • Damen Mohamed Oussama
  • Rekaya-Ben Othman Ghaya
  IEEE Transactions on Communications, Institute of Electrical and Electronics Engineers, 2019, 67 (5), pp.3470-3481. In this paper we analyze the performance of coded and uncoded spatial modulations over multi-mode fiber optic channels with Mode-Dependent Loss (MDL) under Maximum Likelihood (ML) and Zero-Forcing (ZF) detection schemes. We focus on multi-dimensional modulations that satisfy certain orthogonality criteria such as the TAST codes. In particular, we link the post-detection Signal-to-Noise Ratio (SNR) to the Orthogonality Defect Factor (ODF) of the equivalent channel matrix by deriving their closed-form expressions as well as characterizing their statistical properties. Using the post-detection SNR and ODF properties, we develop upper and ad-hoc tight bounds on the error probabilities, which illustrates how these “orthogonal” spatial modulations mitigate the MDL under both ML and ZF detection. The obtained properties also allow us to propose a modification to the detection algorithm such that it still achieves essentially optimal performance but at a much smaller cost than exhaustive or tree search algorithms. The theoretical results are validated numerically and by simulations.
• Second-Order Asymptotics for Communication Under Strong Asynchronism
  
  • Li Longguang
  • Tchamkerten Aslan
  IEEE Transactions on Information Theory, Institute of Electrical and Electronics Engineers, 2019, 65 (5), pp.2838-2849. The capacity under strong asynchronism was recently shown to be essentially unaffected by the imposed decoding delay-the elapsed time between when information is available at the transmitter and when it is decoded-and the output sampling rate. This paper shows that, in contrast with capacity, the second-order term in the maximum rate expansion is sensitive to both parameters. When the receiver must locate the sent codeword exactly and therefore achieve minimum delay equal to the blocklength n, the second-order term in the maximum rate expansion is of order Θ(1/p) for any sampling rate ρ = O(1/√n) (and ρ = ω(1/n) for otherwise reliable communication is impossible). Instead, if ρ = ω(1/√n), then the second-order term is the same as under full sampling and is given by a standard Θ(√n) term. However, if the delay constraint is only slightly relaxed to n(1+o(1)), then the above order transition (for ρ = O(1/√n) and ρ = w(1/√n)) vanishes and the secondorder term remains the same as under full sampling for any ρ = ω(1/n). (10.1109/TIT.2018.2882488)
  DOI : 10.1109/TIT.2018.2882488
• Characterization of Régnier’s matrices in classification
  
  • Hudry Olivier
  , 2019.
• Throughput Optimization in Ultra-Reliable Low-Latency Communication with Short Packets
  
  • Avranas Apostolos
  • Kountouris Marios
  • Ciblat Philippe
  , 2019. We consider an ultra-reliable low-latency communication (URLLC) system with short packets employing hybrid automatic repeat request (HARQ). Depending on the delay of HARQ feedback and retransmissions, the latency constraint can be either violated or fulfilled at the expense of power consumption. We focus on the energy-latency tradeoff and examine whether it is better to do one-shot transmission or use HARQ. We analyze the energy consumption for incremental redundancy (IR) HARQ and compare it with the no HARQ case. The analysis relies on closed-form expressions for the outage probability of IR-HARQ with variables both the blocklength and the power. Our results show that for a wide range of blocklength, when the feedback delay is more than half the latency constraint, it is beneficial in terms of energy to use one-shot transmission (i.e. no HARQ).
• Channel Impulsive Noise Mitigation for Linear Video Coding Schemes
  
  • Zheng S
  • Cagnazzo M.
  • Kieffer Michel
  , 2019. This paper considers the problem of impulse noise mitigation when video is encoded using a SoftCast-based Linear Video Coding (LVC) scheme and transmitted using an Orthogonal Frequency-Division Multiplexing (OFDM) scheme over a wideband channel prone to impulse noise. In the time domain, the impulse noise is modeled as independent and identically distributed (iid) Bernoulli-Gaussian variables. A Fast Bayesian Matching Pursuit algorithm is employed for impulse noise mitigation. This approach requires the provision-ing of some OFDM subchannels to estimate the impulse noise locations and amplitudes. Provisioned subchannels cannot be used to transmit data and lead to a decrease of the video quality at receivers in absence of impulse noise. Using a phenomenological model (PM) of the residual noise variance after impulse mitigation, we have proposed an algorithms that is able to evaluate the optimal number of subchannels to provision for impulse noise correction. Simulation results show that the PM can accurately predict the number of sub-channels to provision and that impulse noise mitigation can significantly improve the decoded video quality compared to a situation where all subchannels are used for data transmission. (10.1109/icassp.2019.8682699)
  DOI : 10.1109/icassp.2019.8682699
• From ERS to Sentinel: processing and analysis of multi-temporal SAR series and applications to urban areas
  
  • Tupin Florence
  , 2019.