Publications

2018

• Fiber optic monitoring of pipelines in permafrost context
  
  • Clément Pierre
  • Gabet Renaud
  • Jaouën Yves
  • Lanticq V.
  , 2018, paper ThE94.
• Neurophysiological Responses to Different Product Experiences
  
  • Modica Enrica
  • Cartocci Giulia
  • Martinez Levy Ana
  • Cherubino Patrizia
  • Maglione Anton Giulio
  • Di Flumeri Gianluca
  • Mancini Marco
  • Montanari Marco
  • Perrotta Davide
  • Di Feo Paolo
  • Vozzi Alessia
  • Ronca Vincenzo
  • Aricò Pietro
  • Babiloni Fabio
  • Rossi Dario
  Computational Intelligence and Neuroscience, 2018.
• ANALYSIS OF COMMON DESIGN CHOICES IN DEEP LEARNING SYSTEMS FOR DOWNBEAT TRACKING
  
  • Fuentes Magdalena
  • Mcfee Brian
  • Crayencour Hélène C
  • Essid Slim
  • Bello Juan P
  , 2018. Downbeat tracking consists of annotating a piece of musical audio with the estimated position of the first beat of each bar. In recent years, increasing attention has been paid to applying deep learning models to this task, and various architectures have been proposed, leading to a significant improvement in accuracy. However, there are few insights about the role of the various design choices and the delicate interactions between them. In this paper we offer a systematic investigation of the impact of largely adopted variants. We study the effects of the temporal granularity of the input representation (i.e. beat-level vs tatum-level) and the encoding of the networks outputs. We also investigate the potential of convolutional-recurrent networks, which have not been explored in previous downbeat tracking systems. To this end, we exploit a state-of-the-art recurrent neural network where we introduce those variants, while keeping the training data, network learning parameters and post-processing stages fixed. We find that temporal granularity has a significant impact on performance, and we analyze its interaction with the encoding of the networks outputs.
• Combining quantum and computational approaches to upgrade secure networks
  
  • Alleaume Romain
  , 2018.
• Simulation of 2 μm thulium-doped single clad silica fiber amplifiers by characterization of the 3F4-3H6 transition
  
  • Romano Clément
  • Tench Robert
  • Delavaux Jean-Marc
  Optics Express, Optical Society of America - OSA Publishing, 2018.
• Influence of the upper nonlasing state on the route to chaos of InAs/GaAs quantum dot lasers
  
  • Huang Heming
  • Arsenijevi Dejan
  • Bimberg Dieter
  • Grillot Frédéric
  , 2018.
• 10-Gb/s Floor-Free Transmission of a Hybrid III-V on Silicon Distributed Feedback Laser with Optical Feedback
  
  • Gomez Sandra
  • Huang Heming
  • Grillot Frédéric
  , 2018.
• Output Fisher embedding regression
  
  • Djerrab Moussab
  • Garcia Alexandre
  • Sangnier Maxime
  • d'Alché-Buc Florence
  Machine Learning, Springer Verlag, 2018, 107 (8-10), pp.1229-1256. (10.1007/s10994-018-5698-0)
  DOI : 10.1007/s10994-018-5698-0
• Automatic Detection of Depressive States from Speech
  
  • Mendiratta Aditi
  • Scibelli Filomena
  • Esposito Antonietta
  • Capuano Vincenzo
  • Likforman-Sulem Laurence
  • Maldonato Mauro
  • Vinciarelli Alessandro
  • Esposito Anna
  Entropy, MDPI, 2018, 69 (6), pp.301-314. This paper proposes a speech-based method for automatic depression classification. The system is based on ensemble learning for Convolutional Neural Networks (CNNs) and is evaluated using the data and the experimental protocol provided in the Depression Classification Sub-Challenge (DCC) at the 2016 Audio–Visual Emotion Challenge (AVEC-2016). In the pre-processing phase, speech files are represented as a sequence of log-spectrograms and randomly sampled to balance positive and negative samples. For the classification task itself, first, a more suitable architecture for this task, based on One-Dimensional Convolutional Neural Networks, is built. Secondly, several of these CNN-based models are trained with different initializations and then the corresponding individual predictions are fused by using an Ensemble Averaging algorithm and combined per speaker to get an appropriate final decision. The proposed ensemble system achieves satisfactory results on the DCC at the AVEC-2016 in comparison with a reference system based on Support Vector Machines and hand-crafted features, with a CNN+LSTM-based system called DepAudionet, and with the case of a single CNN-based classifier. (10.1007/978-3-319-56904-8_29)
  DOI : 10.1007/978-3-319-56904-8_29
• An improved analysis of reliability and entropy for delay PUFs
  
  • Schaub Alexander
  • Danger Jean-Luc
  • Guilley Sylvain
  • Rioul Olivier
  , 2018. (10.1109/DSD.2018.00096)
  DOI : 10.1109/DSD.2018.00096
• CCFI-Cache: A Transparent and Flexible Hardware Protection for Code and Control-Flow Integrity
  
  • Danger Jean-Luc
  • Facon Adrien
  • Guilley Sylvain
  • Heydemann Karine
  • Kühne Ulrich
  • Si Merabet Abdelmalek
  • Timbert Michaël
  , 2018, pp.529-536. In this paper we present a hardware based solution to verify simultaneously Code and Control-Flow Integrity (CCFI), aiming at protecting microcontrollers against both cyber-and physical attacks. This solution is non-intrusive as it does not require any modification of the CPU core. It relies on two additional hardware blocks external to the CPU: The first one – called CCFI-cache – acts as a dedicated cache for the storage of information to check the code and control-flow integrity, and the second one – CCFI-checker – performs control-flow and code integrity verification. Based on a RISC-V platform implementation, we show that the proposed scheme is able to perform online CCFI validation at the price of a small hardware area overhead and doubling the size of the. text section. In most cases, the impact on the run-time performance is on average 32 percent, offering for the first time a generic and practical hardware-enabled cyber-security solution. (10.1109/DSD.2018.00093)
  DOI : 10.1109/DSD.2018.00093
• OpenSSL Bellcore's Protection Helps Fault Attack
  
  • Carre Sebastien
  • Desjardins Matthieu
  • Facon Adrien
  • Guilley Sylvain
  , 2018, pp.500-507. Faults in software implementations target both data and instructions at different locations. Bellcore attack is a well-known fault attack that is able to break CRT-RSA. In response, cryptographic libraries such as OpenSSL are designed with protections. In this paper, we show two new fault locations on OpenSSL implementation of the CRT-RSA signature that restore the Bellcore attack and break OpenSSL protection against it. Quite surprisingly, one of the fault we found is made possible because of the existence of such protection. (10.1109/DSD.2018.00089)
  DOI : 10.1109/DSD.2018.00089
• Qu'est-ce qu'une belle photo ? Essai sur l'esthétique en photographie numérique
  
  • Maître Henri
  , 2018.
• Analysis of Mixed PUF-TRNG Circuit Based on SR-Latches in FD-SOI Technology
  
  • Danger Jean-Luc
  • Yashiro Risa
  • Graba Tarik
  • Mathieu Yves
  • Si-Merabet Abdelmalek
  • Sakiyama Kazuo
  • Miura Noriyuki
  • Nagata Makoto
  , 2018, pp.508-515.
• Video Quality Evaluation for Tile-Based Spatial Adaptation
  
  • Yousef Hiba
  • Le Feuvre J.
  • Valenzise Giuseppe
  • Hulusic Vedad
  , 2018, pp.1-6. The following topics are dealt with: learning (artificial intelligence); feature extraction; video coding; object detection; data compression; image classification; image coding; image representation; image reconstruction; optimisation. (10.1109/MMSP.2018.8547126)
  DOI : 10.1109/MMSP.2018.8547126
• A Storage-Computation-Communication Tradeoff for Distributed Computing
  
  • Yan Qifa
  • Yang Sheng
  • Wigger Michèle
  , 2018, pp.1-5. This paper investigates distributed computing systems where computations are split into “Map” and “Reduce” functions. A new coded scheme, called distributed computing and coded communication (D3C), is proposed, and its communication load is analyzed as a function of the available storage space and the number of intermediate values (IVA) to be computed. D3C achieves the smallest possible communication load for a given storage space, while a smaller number of IVAs need to be computed compared to Li et al.'s coded distributed computing (CDC) scheme. More generally, our scheme can flexibly trade between storage space and the number of IVAs to be computed. Communication load is then analyzed for any given tradeoff. (10.1109/iswcs.2018.8491052)
  DOI : 10.1109/iswcs.2018.8491052
• ALGeoSPF: A Hierarchical Factorization Model for POI Recommendation
  
  • Griesner Jean-Benoît
  • Abdessalem Talel
  • Naacke Hubert
  • Dosne Pierre
  , 2018, pp.87-90. (10.1109/ASONAM.2018.8508249)
  DOI : 10.1109/ASONAM.2018.8508249
• DoF in Sectored Cellular Systems with BS Cooperation Under a Complexity Constraint
  
  • Gelincik Samet
  • Wigger Michèle
  • Wang Ligong
  , 2018. The paper presents upper and lower bounds on the per-user degrees of freedom (DoF) of a sectored hexagonal cellular model where neighboring basestations (BS) can cooperate during at most κ interaction rounds over backhaul links of capacities μ = μ DoF ·[1/2]log P, with P denoting the transmit power at each mobile user. The lower bound is based on practically implementable beamforming and adapts the way BSs cooperate to the sector structure of the cells. It improves over the naive approach that ignores this sector structure. The upper bound is information-theoretic and holds for all possible coding schemes, including for example ergodic interference alignment whose practical implementation currently seems out of reach. Lower and upper bounds show that the complexity constraint imposed by limiting the number of interaction rounds κ indeed limits the largest achievable DoF. In particular, irrespective of the backhaul capacity μ, the per-user DoF cannot exceed a threshold which depends on κ. (10.1109/ISWCS.2018.8491081)
  DOI : 10.1109/ISWCS.2018.8491081
• Throughput-efficient Relay assisted Hybrid ARQ
  
  • Khreis Alaa
  • Ciblat Philippe
  • Bassi Francesca
  • Duhamel Pierre
  , 2018. Reliable data transmission within wireless communication systems can be obtained via various means, including (i) Hybrid Automatic Repeat reQuest (HARQ) mechanisms which allow retransmission of incorrectly decoded packets; (ii) Additional nodes, called relays, which may also help the transmission by retransmitting these packets. An efficient combination of both techniques is therefore of great interest. This paper investigates a relay assisted HARQ protocol aiming at maximizing the system throughput. The protocol allows the source to transmit a new message during the same time-slot in which the relay is retransmitting a previous message. By using an efficient interference canceler at the destination, the numerical results show significant throughput gain compared to standard approaches. (10.1109/iswcs.2018.8491051)
  DOI : 10.1109/iswcs.2018.8491051
• An empirical investigation of botnet as a service for cyberattacks
  
  • Hammi Badis
  • Zeadally Sherali
  • Khatoun Rida
  Transactions on emerging telecommunications technologies, Wiley-Blackwell, 2018, 30 (3), pp.e3537. (10.1002/ett.3537)
  DOI : 10.1002/ett.3537
• Quality Assessment of Deep-Learning-Based Image Compression
  
  • Valenzise Giuseppe
  • Purica Andrei
  • Hulusic Vedad
  • Cagnazzo Marco
  , 2018. —Image compression standards rely on predictive coding , transform coding, quantization and entropy coding, in order to achieve high compression performance. Very recently, deep generative models have been used to optimize or replace some of these operations, with very promising results. However, so far no systematic and independent study of the coding performance of these algorithms has been carried out. In this paper, for the first time, we conduct a subjective evaluation of two recent deep-learning-based image compression algorithms, comparing them to JPEG 2000 and to the recent BPG image codec based on HEVC Intra. We found that compression approaches based on deep auto-encoders can achieve coding performance higher than JPEG 2000, and sometimes as good as BPG. We also show experimentally that the PSNR metric is to be avoided when evaluating the visual quality of deep-learning-based methods, as their artifacts have different characteristics from those of DCT or wavelet-based codecs. In particular, images compressed at low bitrate appear more natural than JPEG 2000 coded pictures, according to a no-reference naturalness measure. Our study indicates that deep generative models are likely to bring huge innovation into the video coding arena in the coming years. (10.1109/mmsp.2018.8547064)
  DOI : 10.1109/mmsp.2018.8547064
• A Repetition Scheme for MBSFN Based Mission-Critical Communications
  
  • Daher Alaa
  • Ali Mohammed Shabbir
  • Coupechoux Marceau
  • Godlewski Philippe
  • Ngouat Pierre
  • Minot Pierre
  , 2018, pp.1-6. (10.1109/VTCFall.2018.8690595)
  DOI : 10.1109/VTCFall.2018.8690595
• A new analytical model for the performance evaluation of the uplink transmission in NB-IoT networks
  
  • Nguyen Tuan Anh
  • Martins Philippe
  • Nguyen van Tam
  • Nguyen T. Mai Trang
  , 2019, pp.1-5. NB-IoT has been defined in 3GPP release 13. It modifies LTE radio access procedures and protocols to fulfill energy consumption, low data rate and latency requirements associated with IoT applications. One of the major issues of NB-IoT deployment is the design of performance evaluation models that make it possible to plan and dimension this new system. This work proposes a new analytical model based on stochastic geometry. It provides a closed form expression of the successful transmission of an NB-IoT system in the uplink for a given sensor spatial intensity. The accuracy of the proposed analytical model is validated by simulation. (10.1109/VTCFall.2018.8690677)
  DOI : 10.1109/VTCFall.2018.8690677
• A survey on game-theoretic approaches for intrusion detection and response optimization
  
  • Kiennert Christophe
  • Ziad Ismail
  • Debar Hervé
  • Leneutre Jean
  ACM Computing Surveys, Association for Computing Machinery, 2018, 51 (5), pp.1-31. Intrusion Detection Systems (IDS) are key components for securing critical infrastructures, capable of detecting malicious activities on networks or hosts. However, the efficiency of an IDS depends primarily on both its configuration and its precision. The large amount of network traffic that needs to be analyzed, in addition to the increase in attacks' sophistication, renders the optimization of intrusion detection an important requirement for infrastructure security, and a very active research subject. In the state of the art, a number of approaches have been proposed to improve the efficiency of intrusion detection and response systems. In this article, we review the works relying on decision-making techniques focused on game theory and Markov decision processes to analyze the interactions between the attacker and the defender, and classify them according to the type of the optimization problem they address. While these works provide valuable insights for decision-making, we discuss the limitations of these solutions as a whole, in particular regarding the hypotheses in the models and the validation methods. We also propose future research directions to improve the integration of game-theoretic approaches into IDS optimization techniques (10.1145/3232848)
  DOI : 10.1145/3232848
• Learning with Noise-Contrastive Estimation: Easing training by learning to scale
  
  • Labeau Matthieu
  • Allauzen Alexandre
  , 2018, pp.3090-3101. Noise-Contrastive Estimation (NCE) is a learning criterion that is regularly used to train neural language models in place of Maximum Likelihood Estimation, since it avoids the computational bottleneck caused by the output softmax. In this paper, we analyse and explain some of the weaknesses of this objective function, linked to the mechanism of self-normalization, by closely monitoring comparative experiments. We then explore several remedies and modifications to propose tractable and efficient NCE training strategies. In particular, we propose to make the scaling factor a trainable parameter of the model, and to use the noise distribution to initialize the output bias. These solutions, yet simple, yield stable and competitive performances in either small and large scale language modelling tasks.