Publications

2023

• A global–local attention network for uncertainty analysis of ground penetrating radar modeling
  
  • Zhao Yunjie
  • Cheng Xi
  • Zhang Taihong
  • Wang Lei
  • Shao Wei
  • Wiart Joe
  Reliability Engineering and System Safety, Elsevier, 2023, 234, pp.109176. (10.1016/j.ress.2023.109176)
  DOI : 10.1016/j.ress.2023.109176
• Dynamic logic-based attack graph for risk assessment in complex computer systems
  
  • Boudermine Antoine
  • Khatoun Rida
  • Choyer Jean-Henri
  Computer Networks, Elsevier, 2023, 228, pp.109730. Nowadays, Information and Communication Technologies (ICT) play a significant role in our modern daily life. Computer networks breakdown can strongly impact everything in our life such as personal data, industrials, banks, oil pipelines, hospitals, nuclear reactors, military platforms, etc. Assessing their security is a necessity to reduce the risk of compromise by an attacker. Nevertheless, the actual solutions are rarely adapted to the high complexity of modern computer systems. They often rely on too much human work and the used algorithms do not scale well. Furthermore, the evolution of the system over time is rarely modeled and is therefore not considered in the evaluation of its security. This paper proposes a dynamic attack graph generation method allowing to model attack paths by considering the evolution of the system over time. We compute the probabilities of compromise of the system components by simulating several cyberattacks from the previously constructed dynamic attack graph. We tested our solution on a use case of several thousand of machines. The measured results demonstrate its ability to assess the threat in complex systems caused by combining exploitation of successive vulnerabilities. (10.1016/j.comnet.2023.109730)
  DOI : 10.1016/j.comnet.2023.109730
• Hi! PARIS: IA et Sciences des données pour la société
  
  • Richard Gael
  • Nicolas Vieille
  • Eric Moulines
  Télécom : revue de l'Association Amicale des ingénieurs de l'Ecole Nationale Supérieure des télécommunications, 2023 (#209).
• Dependable STT-MRAM With Emerging Approximation and Speculation Paradigms
  
  • Cai Hao
  • Hou Yaoru
  • Zhang Mengdi
  • Liu Bo
  • Naviner Lirida
  IEEE Design & Test, IEEE, 2023, 40 (3), pp.17-25. Editor’s notes: STT-MRAMs are a promising candidate for nonvolatile memory (NVM), for example, for caches and implementing embedded NVM. This article proposes a design flow combining concepts of timing speculation and approximate storage to achieve a dependable operation of STT-MRAM devices. —Jürgen Teich, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU). (10.1109/MDAT.2021.3120330)
  DOI : 10.1109/MDAT.2021.3120330
• Functional anomaly detection: a benchmark study
  
  • Staerman Guillaume
  • Adjakossa Eric
  • Mozharovskyi Pavlo
  • Hofer Vera
  • Sen Gupta Jayant
  • Clémençon Stéphan
  International Journal of Data Science and Analytics, Springer Verlag, 2023, 16 (1), pp.101-117. The increasing automation in many areas of the Industry expressly demands to design efficient machine learning solutions for the detection of abnormal events. With the ubiquitous deployment of sensors monitoring nearly continuously the health of complex infrastructures, anomaly detection can now rely on measurements sampled at a very high frequency, providing a very rich representation of the phenomenon under surveillance. In order to exploit fully the information thus collected, the observations cannot be treated as multivariate data anymore and a functional analysis approach is required. It is the purpose of this paper to investigate the performance of recent techniques for anomaly detection in the functional setup on real datasets. After an overview of the state of the art and a visual-descriptive study, a variety of anomaly detection methods are compared. While taxonomies of abnormalities (e.g., shape, location) in the functional setup are documented in the literature, assigning a specific type to the identified anomalies appears to be a challenging task. Thus, strengths and weaknesses of the existing approaches are benchmarked in view of these highlighted types in a simulation study. Anomaly detection methods are next evaluated on two datasets, related to the monitoring of helicopters in flight and to the spectrometry of construction materials namely. The benchmark analysis is concluded by a recommendation guidance for practitioners (10.1007/s41060-022-00366-5)
  DOI : 10.1007/s41060-022-00366-5
• LUVI: Lightweight UWB-VIO based relative positioning for AR-IoT applications
  
  • Choi Hong-Beom
  • Lim Keun-Woo
  • Ko Young-Bae
  Ad Hoc Networks, Elsevier, 2023, 145, pp.103132. In this paper, we propose LUVI, Lightweight UWB-VIO relative positioning method for indoor localization. Recent designs of handheld and embedded devices feature various technologies which have the means to enhance localization performance in indoor environments. These include visual odometry based on cameras and augmented reality, and communication hardware such as UWB. Integration of such technologies to exploit their advantages allows us to compensate for each other's errors in measurement. This improves the overall function of future services, such as visual representation of sensing information from sensors in areas that are not physically visible. However, existing work cannot fully exploit these technologies to high extent, often inducing high errors or wasted resources. LUVI is a novel localization method which estimates the location of a target object using relative coordinates of estimator devices without the aid of definitive coordinates. LUVI focuses on utilization of lightweight management of virtual anchors for localization, with functions that reduce the computing and communication complexity while maintaining the accuracy and improving energy efficiency of the localization. Our work has been fully implemented and tested in several indoor environments, showing robustness to NLOS while significantly reducing computational complexity, and up to 30% lower average error. (10.1016/j.adhoc.2023.103132)
  DOI : 10.1016/j.adhoc.2023.103132
• Procédé et système d’authentification par un équipement vérificateur d’un dispositif à authentifier équipé d’un circuit PUF
  
  • Khalfaoui Sameh
  • Villard Arthur
  • Ma Jingxuan
  • Leneutre Jean
  , 2023.
• Towards the Verification of Strategic Properties in Multi-Agent Systems with Imperfect Information
  
  • Ferrando Angelo
  • Malvone Vadim
  , 2023.
• Lightweight TLS 1.3 Handshake for C-ITS Systems
  
  • Goncharskyi Danylo
  • Kim Sung Yong
  • Gu Pengwenlong
  • Serhrouchni Ahmed
  • Khatoun Rida
  • Nait-Abdesselam Farid
  , 2023, pp.1432-1437. Cooperative Intelligent Transport Systems (C-ITS) Deployment Platform is considered the newest version of vehicular communication systems, which enables the cooperation between two or more ITS sub-systems to provide enhanced services. With the expanded communication range and system complexity, ensuring the credibility of access nodes and protecting users from being monitored has become a difficult problem in network security, especially the services provided by remote servers like navigation. Transport Layer Security (TLS) is widely used for user authentication and encrypted data transmission in all networks. However, although the TLS handshake complexity is significantly reduced in TLS 1.3 the transmission of a full certificate chain during the handshake is still costly, especially for high-mobility vehicles. In this paper, we propose an optional extension named Certificate Get to reduce the TLS handshake overhead in C-ITS. Specifically, with our proposed extension, the revisiting client transmits a hash value of the certificate chain corresponding to a certain server in the ClientHello message, which can reduce the transmission payload of the certificate chain from an average of 4874 bytes to 68 bytes. Simulation results show that our proposed scheme achieves a significant performance gain by greatly reducing the certificate transmission delay by 50% for both TLS 1.3 and TLS 1.2. (10.1109/ICC45041.2023.10279295)
  DOI : 10.1109/ICC45041.2023.10279295
• Defensive Randomization Against Adversarial Attacks in Image-Based Android Malware Detection
  
  • Lan Tianwei
  • Darwaish Asim
  • Nait-Abdesselam Farid
  • Gu Pengwenlong
  , 2023, pp.5072-5077. (10.1109/ICC45041.2023.10279592)
  DOI : 10.1109/ICC45041.2023.10279592
• Cascaded binary classifiers for blind Beam Alignment in mmWave MIMO using one-bit quantization
  
  • Ktari Aymen
  • Ghauch Hadi
  • Rekaya Ben Othman Ghaya
  , 2023 (80-85). This paper proposes a new approach for partial and blind Machine Learning (ML)-based Beam Alignment (BA) for massive mmWave MIMO. It models an uplink scenario using one-bit quantization through a low-complexity fully-analog system architecture. The proposed BA is based on sub-sampled codebooks holding possible beam patterns at UE and BS. We propose to sound a small subset of beams based on instantaneous Received Signal Energies (RSE). These sounded RSE values are then quantized into binary integers. The proposed cascaded structure of Binary Logistic Regression (BLR) aims to iteratively filter the large dataset input-matrix (by deleting low-RSE beams) into a smaller one where our benchmark, the Exhaustive BA is feasible and the signaling overhead remains low. In addition to the theoretical monotonic-convergence guarantees, BLR has good classification quality and low computational complexity. Our extensive numerical simulations illustrate encountering the large signaling overhead problem with high prediction accuracy using one-bit quantization scheme and 14% of the total beam samples. (10.1109/ICCWorkshops57953.2023.10283648)
  DOI : 10.1109/ICCWorkshops57953.2023.10283648
• Non-Invasive Absorbed Power Density Assessment from 5G Millimeter-Wave Mobile Phones Using Method of Moments
  
  • Jafari Seyed Faraz
  • Shirazi Reza Sarraf
  • Moradi Gholamreza
  • Sibille Alain
  • Wiart Joe
  IEEE Transactions on Antennas and Propagation, Institute of Electrical and Electronics Engineers, 2023, pp.1-1. (10.1109/TAP.2023.3278834)
  DOI : 10.1109/TAP.2023.3278834
• The Software Heritage Open Science Ecosystem
  
  • Cosmo Roberto Di
  • Zacchiroli Stefano
  , 2023, pp.33-61. Abstract Software Heritage is the largest public archive of software source code and associated development history, as captured by modern version control systems. As of July 2023, it has archived more than 16 billion unique source code files coming from more than 250 million collaborative development projects. In this chapter, we describe the Software Heritage ecosystem, focusing on research and open science use cases. On the one hand, Software Heritage supports empirical research on software by materializing in a single Merkle direct acyclic graph the development history of public code. This giant graph of source code artifacts (files, directories, and commits) can be used—and has been used—to study repository forks, open source contributors, vulnerability propagation, software provenance tracking, source code indexing, and more. On the other hand, Software Heritage ensures availability and guarantees integrity of the source code of software artifacts used in any field that relies on software to conduct experiments, contributing to making research reproducible. The source code used in scientific experiments can be archived—e.g., via integration with open-access repositories—referenced using persistent identifiers that allow downstream integrity checks and linked to/from other scholarly digital artifacts. (10.1007/978-3-031-36060-2_2)
  DOI : 10.1007/978-3-031-36060-2_2
• An Extrapolation Approach for RF-EMF Exposure Prediction in an Urban Area using Artificial Neural Network
  
  • Chikha Wassim Ben
  • Wang Shanshan
  • Wiart Joe
  IEEE Access, IEEE, 2023, pp.1-1. (10.1109/ACCESS.2023.3280125)
  DOI : 10.1109/ACCESS.2023.3280125
• Impact of non-Lorentzian laser phase noise on Phase-OTDR performance
  
  • Dorize Christian
  • Guerrier Sterenn
  • Awwad Élie
  • Renaudier Jérémie
  , 2023, pp.25. We highlight the importance of the laser source phase noise in sensing applications and show that the standard Lorentzian linewidth criterion is not sufficient to characterize the performance of a sensing system. We then derive a laser linewidth related to the phase noise spectral region of interest, according to the length of the fiber to sense. This is illustrated in a setup based on coded interrogation and with two sensing dedicated laser sources. (10.1117/12.2678119)
  DOI : 10.1117/12.2678119
• Coherent combination method applied to distributed acoustic sensing over deployed multicore fiber
  
  • Orsuti Daniele
  • Guerrier Sterenn
  • Palmieri Luca
  • Awwad Élie
  • Dorize Christian
  • Antonelli Cristian
  • Mecozzi Antonio
  , 2023, 12643, pp.54. From Distributed Acoustic Sensing (DAS) measurements over deployed Multi-Core Fiber (MCF), we discuss several signal processing options to enhance the sensing sensitivity, namely core combination and longitudinal averaging. (10.1117/12.2678315)
  DOI : 10.1117/12.2678315
• Quantifying the Bias of Transformer-Based Language Models for African American English in Masked Language Modeling
  
  • Salutari Flavia
  • Ramos Jerome
  • Rahmani Hosein A
  • Linguaglossa Leonardo
  • Lipani Aldo
  , 2023. In the last three years we witnessed the proliferation of innovative natural language processing (NLP) algorithms attempting at solving different tasks and designed for the most diverse applications. Despite groundbreaking transformer-based language models (LMs) have been proposed and widely adopted, the measurement of their fairness with respect to different social groups still remains unsolved. In this paper, we propose and thoroughly validate an evaluation technique to assess the quality and the bias of the predictions of these LMs on transcripts of both spoken African American English (AAE) and Standard American English (SAE). Our analysis reveals the presence of a bias towards SAE encoded by state-of-the-art LMs, like BERT and DistilBERT, a lower bias in distilled LMs and an opposite bias in RoBERTa and BART. Additionally, we show evidence that this disparity is present across all the LMs when we only consider the grammar and the syntax specific to AAE.
• Signalisation cellulaire pour la detection des fraudes de contournement
  
  • Kouam Anne Josiane
  • Carneiro Viana Aline
  • Martins Philippe
  • Adjih Cédric
  • Tchana Alain
  , 2023. La fraude de contournement, également connue sous le nom de fraude à la SIMBox, est l'une des plus sévères dans les réseaux cellulaires, générant des pertes annuelles de 3,11 milliards de dollars et des menaces pour la sécurité nationale. Un challenge majeur à sa détection est l'évolution constante de la fraude en vue de contourner les solutions publiées dans la littérature. Ce papier explore une nouvelle source de données résiliente à l'évolution de la fraude: la signalisation cellulaire. Au travers d'expérimentations avec des appareils SIMBox nous montrons son potentiel à distinguer les équipements SIMBox des téléphones ordinaires en temps-réel et avant que la fraude ne soit commise.
• A geometrically aware auto-encoder for multi-texture synthesis
  
  • Chatillon Pierrick
  • Gousseau Yann
  • Lefebvre Sidonie
  , 2023, 14009, pp.263-275. We propose an auto-encoder architecture for multi-texture synthesis. The approach relies on both a compact encoder accounting for second order neural statistics and a generator incorporating adaptive periodic content. Images are embedded in a compact and geometrically consistent latent space, where the texture representation and its spatial organisation are disentangled. Texture synthesis and interpolation tasks can be performed directly from these latent codes. Our experiments demonstrate that our model outperforms state-of-the-art feed-forward methods in terms of visual quality and various texture related metrics. The code is available online. (10.1007/978-3-031-31975-4_20)
  DOI : 10.1007/978-3-031-31975-4_20
• Permissionless and asynchronous asset transfer
  
  • Kuznetsov Petr
  • Pignolet Yvonne-Anne
  • Ponomarev Pavel
  • Tonkikh Andrei
  Distributed Computing, Springer Verlag, 2023, 36 (3), pp.349-371. Most modern asset transfer systems use consensus to maintain a totally ordered chain of transactions. It was recently shown that consensus is not always necessary for implementing asset transfer. More efficient, asynchronous solutions can be built using reliable broadcast instead of consensus. This approach has been originally used in the closed (permissioned) setting. In this paper, we extend it to the open (permissionless) environment. We present PASTRO, a permissionless and asynchronous asset-transfer implementation, in which quorum systems, traditionally used in reliable broadcast, are replaced with a weighted Proof-of-Stake mechanism. PASTRO tolerates a dynamic adversary that is able to adaptively corrupt participants. (10.1007/S00446-023-00449-X)
  DOI : 10.1007/S00446-023-00449-X
• EFFLAM Project: from 12-Core Erbium-Ytterbium Fiber amplifier design to SDM transmissions applications
  
  • Jaouën Yves
  • Lebreton Aurélien
  • Pincemin Erwan
  , 2023. Space Division Multiplexing (SDM) is considered as a viable solution to face the upcoming capacity crunch. There has been a lot of work to develop passive multicore fibers (MCF) for long-haul transmission applications. One of the key issues studied in parallel is the availability of power efficient amplification schemes, including active MCFs and fan-in/fan-out solutions. The EFFLAM project concerns the design of a 12-core Er3+/Yb3+ co-doped fiber (12c-EYDF) amplifier prototype and its application in different systems configurations. The comparative advantages/drawbacks of Er3+ and Er3+/Yb3+ doping are relatively well-known. An Er3+/Yb3+ co-doped fiber amplifier provides higher gain and better pump efficiency than a classical Erbium amplifier. The pump absorption is also much higher but, in counterpart, such Er3+/Yb3+ fiber reduces the telecom C-band WDM amplification window. The obtained characteristics of the designed 12c-EYDF amplifier prototype, that includes two 12+1 combiners directly spliced to the active MCF, are described. The 12c-EYDF is fabricated using the stack-and-draw process which offers versality in the preform design. This preform is based on a circular arrangement of doped preforms around a central undoped rod, with all the cores lying on a single circle. Reproducible core-to-core distances of 35±0.5μm are obtained, leading to a satisfactorily accurate matching to the 12c-EYDF geometry and signals through the 12 outer fibers designed to match mode field diameter of active cores after down-tapering. Core contents in erbium and ytterbium strongly impact amplifier properties. A numerical model is used to meet the targeted amplifier requirements (+20dB gain with a per core input power of 0dBm in the C-band). To determine the optimal doping concentrations, we used an algorithm based on PSO (Particle Swarm Optimization). Pump and signals are coupled into a 5.5m long 12c-EYDF using a tapered fiber bundle-based combiner. An averaged insertion loss of 1.7dB is obtained with ±0.3dB core-to-core variations. A 5.3W pump light is launched through the central 195μm/230μm multimode fiber. Individual core performances are measured by launching, into each core, 15 WDM non-modulated channels in the 1535.2-1564.1nm range, with 0dBm per channel input power. The maximal core-to-core output power variation is 1.4dB. Gain and NF at 1550nm are respectively 20dB and 6dB. Gain variation is 6dB from 1535nm to 1562nm.These values are in agreement with PSO simulation. Finally, the performances of the 12c-EYDF amplifier prototype were evaluated in different transmission networks applications scenario such as long-haul (LH), metro-regional and data-center interconnects (DCI).
• From the Standards to Silicon : Formally Proved Memory Controllers
  
  • Malaquias Felipe Lisboa
  • Asavoae Mihail
  • Brandner Florian
  , 2023, 13903, pp.295-311. Recent research in both academia and industry has successfully used deductive verification to design hardware and prove its correctness. While tools and languages to write formally proved hardware have been proposed, applications and use cases are often overlooked. In this work, we focus on Dynamic Random Access Memories (DRAM) controllers and the DRAM itself – which has its expected temporal and functional behaviours described in the standards written by the Joint Electron Device Engineering Council (JEDEC). Concretely, we associate an existing Coq DRAM controller framework – which can be used to write DRAM scheduling algorithms that comply with a variety of correctness criteria – to a back-end system that generates proved logically equivalent hardware. This makes it possible to simultaneously enjoy the trustworthiness provided by the Coq framework and use the generated synthesizable hardware in real systems. We validate the approach by using the generated code as a plug-in replacement in an existing DDR4 controller implementation, which includes a host interface (AXI), a physical layer (PHY) from Xilinx, and a model of a memory part Micron MT40A1G8WE-075E:D. We simulate and synthesise the full system. (10.1007/978-3-031-33170-1_18)
  DOI : 10.1007/978-3-031-33170-1_18
• Cardinal, mais qu’est-ce donc cette diablerie ?
  
  • Rioul Olivier
  • Rabiet Victor
  , 2023. The Whittaker cardinal function was discovered by E. T. Whittaker [1], who wanted to know whether there exists in the class of all functions which take on the same values at the set of points A = ¡/czz}¡j°-_»,h > 0, "a function of royal blood whose distinguished properties set it apart from its bourgeois brethren". This function then played a fundamental role in the development of the theory of central difference processes, a theory which was also originated by E. T. Whittaker [1]. Somewhat later J. M. Whittaker and his co-workers [2], [3]
• StreamAI: Dealing with Challenges of Continual Learning Systems for Serving AI in Production
  
  • Barry Mariam
  • Bifet Albert
  • Billy Jean-Luc
  , 2023, pp.134--137. How to build, deploy, update & maintain dynamic models which continuously learn from streaming data? This paper covers the industrialization aspects of these questions in production systems. In today’s fast-changing environments, organizations are faced with the crucial challenge of predictive analytics in online fashion from big data and deploying Artificial Intelligence models at scale. Applications include cyber-security, cloud infrastructure, social networks and financial markets. Online learning models that learn continuously and adapt to the potentially evolving data distributions have demonstrated efficiency for big data stream learning. However, the challenges of deploying and maintaining such models in production (serving) have stalled their adoption. In this paper, we first categorize key challenges faced by the R&D, MLOps and governance teams for deploying automated and self-training AI models in production. Next, we highlight the challenges related to stream-based online machine-learning systems. Finally, we propose StreamAI, a technology-agnostic architecture to deal with the MLOps journey (learning, serving, maintenance) of online models in production. We conclude with open research questions for AI, MLOps and software engineering to bridge the gaps between industry needs and research-oriented development. (10.1109/ICSE-SEIP58684.2023.00017)
  DOI : 10.1109/ICSE-SEIP58684.2023.00017
• Far-field Surrogate Model of Flexible Antennas Based on Vector Spherical Harmonics and Neural Network
  
  • Hou Guoqing
  • Du Jinxin
  • Roblin Christophe
  • Yang Xue-Xia
  , 2023, pp.1-3. To speed up the stochastic modeling of the farfield (FF) electric field of flexible antennas, a novel method combining vector spherical harmonics (VSH) and neural network (NN) is proposed to construct efficient and effective surrogate models. First, we use VSH to parsimoniously representing the antenna's FF electric field vector with a limited number of modes; then, we use NN to map between the input variables and the VSH mode coefficients. We proposed an improved successive halving (ISH) algorithm to optimize the selection of hyperparameters when constructing the NN model. The results show that compared with the polynomial chaos expansion (PCE) model, the prediction error of the NN model has been reduced by 39.22% at the same modeling cost. (10.1109/ICMMT58241.2023.10277192)
  DOI : 10.1109/ICMMT58241.2023.10277192