Publications

2022

• Core Selection for Capacity on Demand in Multi-Core Fiber Transmission System
  
  • Abouseif Akram
  • Rekaya Ben-Othman Ghaya
  • Jaouën Yves
  , 2022. We propose and validate a core selection method on coupled-based MCF transmission system. This method allows to answer the question of capacity on demand of future application, without sacrificing the performance.
• Core Selection for Capacity on Demand in Multi-Core Fiber Transmission System
  
  • Abouseif Akram
  • Othman Ghaya Rekaya-Ben
  • Jaouën Yves
  , 2022, pp.878-880. (10.1109/ACP55869.2022.10088981)
  DOI : 10.1109/ACP55869.2022.10088981
• A unified method to design bridges for OPC UA PubSub networks in the industrial IoT
  
  • Nguyen Quang-Duy
  • Dhouib Saadia
  • Bellot Patrick
  , 2022. Specification part 14 of the Open Platform Communication Unified Architecture (OPC UA) standard provides five different profiles to implement the publish-subscribe messaging pattern. The specification is also called OPC UA PubSub, and its profiles are called PubSub profiles. Two devices deployed with the same PubSub profile can exchange and collaborate; however, two devices deployed with two different PubSub profiles are unable to communicate. It is a limit for the Industry Internet of Things, a complex environment where there would be heterogeneous devices and networks. One approach to overcoming this issue is to use a bridge for the devices deployed with different PubSub profiles. In this sense, this paper provides a unified method to design bridges for OPC UA PubSub networks. The proof-ofconcept experiment, also presented in this paper, is a use case of bridging PubSub broker-less and broker-based networks.
• Proceedings of the 7th Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE 2022)
  
  • Lagrange Mathieu
  • Mesaros Annamaria
  • Pellegrini Thomas
  • Richard Gael
  • Serizel Romain
  • Stowell Dan
  , 2022, pp.1-225.
• Perles du second degré
  
  • Zayana Karim
  • Evci Evrim
  CultureMath, ENS, 2022.
• Strong converses using change of measure and asymptotic markov chains
  
  • Hamad Mustapha
  • Wigger Michèle
  • Sarkiss Mireille
  , 2023, pp.535-540. The main contribution of this paper is a strong converse result for K-hop distributed hypothesis testing against independence with multiple (intermediate) decision centers under a Markov condition. Our result shows that the set of type-II error exponents that can simultaneously be achieved at all the terminals does not depend on the maximum permissible type-I error probabilities. Our strong converse proof is based on a change of measure argument and on the asymptotic proof of specific Markov chains. This proof method seems to be useful also in other applications, and is appealing because it does not require resorting to variational characterizations or blowing-up methods as in previous related proofs. (10.1109/ITW54588.2022.9965790)
  DOI : 10.1109/ITW54588.2022.9965790
• Linear View Change in Optimistically Fast BFT
  
  • Rambaud Matthieu
  • Tonkikh Andrei
  • Abspoel Mark
  , 2022, pp.67-78. To be competitive with centralized applications, consensus protocols in blockchains must provide minimal latency while being able to scale to thousands of participants in order to preserve a high level of decentralization. A common way to minimize latency is to augment a consensus protocol with a fast track, which ensures that a decision is reached in just a couple of message delays in favorable conditions. However, it is a challenging task to preserve safety and good performance when these favorable conditions do not hold. To the best of our knowledge, all existing Byzantine fault-tolerant consensus protocols with fast tracks require view change protocols with quadratic authenticator complexity. In this paper, we provide the first solution to Byzantine consensus with fast track with a linear view change. The protocol incurs no asymptotic overhead over the baseline while reducing the latency in favorable conditions by a factor of 2. Our construction is based on a novel type of cryptographic proofs, which we call Proofs of Exclusivity (or PoE for short), which may be of independent interest. While our protocol for constructing a PoE comes at no extra costs in latency or asymptotic complexities, it does require some extra computation. To make sure that it does not impair the overall performance, we also show how to apply accountability and proofs of misbehavior in order to reduce to zero the overhead incurred by the computation of a PoE. More precisely, our mechanism guarantees that whenever this overhead is not zero, then automatically honest participants obtain a publicly verifiable proof that a well-identified malicious participant openly misbehaved. In this case, the overhead of computing a few extra threshold signatures for the Proof of Exclusivity can be seen as a relatively small price to get rid of a malicious participant (10.1145/3560829.3563562)
  DOI : 10.1145/3560829.3563562
• Latent and Adversarial Data Augmentation for Sound Event Detection and Classification
  
  • Perera David
  • Essid Slim
  • Richard Gaël
  , 2022. Invariance-based learning is a promising approach in deep learning. Among other benefits, it can mitigate the lack of diversity of available datasets and increase the interpretability of trained models. To this end, practitioners often use a consistency cost penalizing the sensitivity of a model to a set of carefully selected data augmentations. However, there is no consensus about how these augmentations should be selected. In this paper, we study the behavior of several augmentation strategies. We consider the task of sound event detection and classification for our experiments. In particular, we show that transformations operating on the internal layers of a deep neural network are beneficial for this task.
• Between Principle and Pragmatism: Reflections on Prototyping Computational Media with Webstrates
  
  • Borowski Marcel
  • Fog Bjarke V
  • Griggio Carla F
  • Eagan James R
  • Klokmose Clemens N
  ACM Transactions on Computer-Human Interaction, Association for Computing Machinery, 2022. Computational media describes a vision of software, which, in contrast to application-centric software, is (1) malleable, so users can modify existing functionality, (2) computable, so users can run custom code, (3) distributable, so users can open documents across different devices, and (4) shareable, so users can easily share and collaborate on documents. Over the last ten years, the Webstrates and Codestrates projects aimed to realize this vision of computational media. Webstrates is a server application that synchronizes the DOM of websites. Codestrates builds on top of Webstrates and adds an authoring environment, which blurs the use and development of applications. Grounded in a chronology of the development of Webstrates and Codestrates, we present eight tensions that we needed to balance during their development. We use these tensions as an analytical lens in three case studies and a game challenge in which participants created games using Codestrates. We discuss the results of the game challenge based on these tensions and present key takeaways for six of them. Finally, we present six lessons learned from our endeavor to realize the vision of computational media, demonstrating the balancing act of weighing the vision against the pragmatics of implementing a working system. CCS Concepts: • Human-centered computing → Interactive systems and tools; Empirical studies in HCI ; Web-based interaction; Collaborative interaction. (10.1145/3569895)
  DOI : 10.1145/3569895
• Optimal Convergence Rate in the Quantum Zeno Effect for Open Quantum Systems in Infinite Dimensions
  
  • Möbus Tim
  • Rouzé Cambyse
  Annales Henri Poincaré, Springer Verlag, 2022, 24 (5), pp.1617-1659. Abstract In open quantum systems, the quantum Zeno effect consists in frequent applications of a given quantum operation, e.g., a measurement, used to restrict the time evolution (due, for example, to decoherence) to states that are invariant under the quantum operation. In an abstract setting, the Zeno sequence is an alternating concatenation of a contraction operator (quantum operation) and a $$C_0$$ C 0 -contraction semigroup (time evolution) on a Banach space. In this paper, we prove the optimal convergence rate $$\mathcal {O}(\tfrac{1}{n})$$ O ( 1 n ) of the Zeno sequence by proving explicit error bounds. For that, we derive a new Chernoff-type $$\sqrt{n}$$ n -Lemma, which we believe to be of independent interest. Moreover, we generalize the convergence result for the Zeno effect in two directions: We weaken the assumptions on the generator, inducing the Zeno dynamics generated by an unbounded generator, and we improve the convergence to the uniform topology. Finally, we provide a large class of examples arising from our assumptions. (10.1007/s00023-022-01241-6)
  DOI : 10.1007/s00023-022-01241-6
• An Embedded AI-Based Smart Intrusion Detection System for Edge-to-Cloud Systems
  
  • Shrivastwa Ritu-Ranjan
  • Bouakka Zakaria
  • Perianin Thomas
  • Dislaire Fabrice
  • Gaudron Tristan
  • Souissi Youssef
  • Karray Khaled
  • Guilley Sylvain
  , 2022, 1747, pp.20-39. This article proposes a general purpose IoT framework usually applicable to all Edge-to-Cloud applications and provides an evaluation study on a use-case involving automotive V2X architecture, tested and verified on a toy smart-car in an emulated smart-car environment. The architecture in study is finely tuned to mimic actual scenarios and therefore the sensors available on the toy car encompasses almost all the sensors that assist a regular ADAS in smart cars of today. The cloud connectivity is maintained through the CoAP protocol which is a standard IoT connectivity protocol. Finally, the security solution proposed is that of a smart Intrusion Detection System (IDS) that is built using Machine Learning (ML) technique and is deployed on the edge. The edge IDS is capable of performing anomaly detection and reporting both detection results as well as sensor collected big data to the cloud. On the cloud side the server stores and maintains the collected data for further retraining of ML models for edge anomaly detection which is differentiated into two categories viz.\@ sensor anomaly detection model and network anomaly detection model. To demonstrate Software update Over The Air (SW-OTA) the cloud in the evaluation setup implements a ML model upgrade capability from the cloud to the connected edge. This implementation and evaluation provides a Proof-of-Concept of the choice of ML as IDS candidate and the framework in general to be applicable to various other IoT scenarios such as Healthcare, Smart-home, Smart-city, Harbour and Industrial environments, and so on, and paves way for future optimization studies. (10.1007/978-3-031-23201-5_2)
  DOI : 10.1007/978-3-031-23201-5_2
• Distributed Randomness from Approximate Agreement
  
  • Freitas Luciano
  • Kuznetsov Petr
  • Tonkikh Andrei
  , 2022. Randomisation is a critical tool in designing distributed systems. The common coin primitive, enabling the system members to agree on an unpredictable random number, has proven to be particularly useful. We observe, however, that it is impossible to implement a truly random common coin protocol in a fault-prone asynchronous system. To circumvent this impossibility, we introduce two relaxations of the perfect common coin: (1) approximate common coin generating random numbers that are close to each other; and (2) Monte Carlo common coin generating a common random number with an arbitrarily small, but non-zero, probability of failure. Building atop the approximate agreement primitive, we obtain efficient asynchronous implementations of the two abstractions, tolerating up to one third of Byzantine processes. Our protocols do not assume trusted setup or public key infrastructure and converge to the perfect coin exponentially fast in the protocol running time. By plugging one of our protocols for Monte Carlo common coin in a well-known consensus algorithm, we manage to get a binary Byzantine agreement protocol with O(n³ log n) communication complexity, resilient against an adaptive adversary, and tolerating the optimal number f < n/3 of failures without trusted setup or PKI. To the best of our knowledge, the best communication complexity for binary Byzantine agreement achieved so far in this setting is O(n⁴). We also show how the approximate common coin, combined with a variant of Gray code, can be used to solve an interesting problem of Intersecting Random Subsets, which we introduce in this paper. (10.4230/LIPICS.DISC.2022.24)
  DOI : 10.4230/LIPICS.DISC.2022.24
• Estimating Speedup Factor For Personal HSMs Based On Secure Elements
  
  • Urien Pascal
  , 2022, pp.1-4. This paper presents experimental results in order to estimate speedup factor for personal HSM. A personal HSM is built over a grid of secure elements, and runs two TLS daemons, one for secure element programming, and another one for service interface. We presents on original personal HSM, working with secure element processors and I2C bus, which supports up to 16 secure elements. We perform tests that open simultaneous TLS sessions and compute signatures. The speedup factor is in the range 50 to 100. Finally we try to find an optimal performance/cost balance, and we argue that personal HSMs may have a better price to performance ratio than traditional HSMs in a not so far future. (10.1109/CSNet56116.2022.9955609)
  DOI : 10.1109/CSNet56116.2022.9955609
• Analysis of The Manhattan Update Rule Algorithm
  
  • Chabane Lylia Thiziri
  • Pham Dang-Kien Germain
  • Desgreys Patricia
  , 2022, pp.1-4. In order to overcome the limitations of traditional computer media, many researchers are turning to analog neural networks to get closer to the functioning of the brain. It is then necessary to use HW-friendly algorithms such as the Manhattan Update Rule (MUR) which is a version of the Back-Propagation (BP) algorithm compatible with the HW implementation. Although many studies use this algorithm for their hardware implementation of neural networks, no article proposes a study upstream of the latter. In this article, we propose an analysis methodology of the Manhattan algorithm allowing us to choose the weight update value ∆ω in order to obtain a minimum of 90% of accuracy. Also, we have answered the questions raised by the state of the art: we have therefore shown that it is possible to achieve the same performance as the BP in terms of precision (3.1% difference at max) and convergence speed. We have shown the link between the dependence of the number of epochs on the initialization of the weights, with the size of the network and the database. Finally, we gave indications for the choice of the version of the MUR (batch or stochastic) according to their speed of convergence. (10.1109/ICECS202256217.2022.9971047)
  DOI : 10.1109/ICECS202256217.2022.9971047
• nnMorpho, a PyTorch library for Mathematical Morphology operators
  
  • Romero-García Gonzalo
  • Agon Carlos
  • Bloch Isabelle
  , 2022.
• Scattering at the angles of polyhedral rooms: application of stress-energy tensor conservation in Riemannian spaces
  
  • Polack Jean-Dominique
  • Meacham Aidan
  • Badeau Roland
  • Valière Jean-Christophe
  , 2022, pp.1-9. Riemannian spaces with negative curvature constitute the proper setting for the distribution of images created by irregular polyhedral rooms with obtuse angles. The crucial parameter is the excess angle that arises around specific edges, called hinges, when first and second order images are considered, as it pilots the metric tensor of the space and all its geometrical properties. With the use of these geometrical properties, and complementing it with the uncertainty principle, we describe the scattering of wave packets around dihedral angles: it is proportional to the excess angle, and is best described in terms of the conservation of the stress-energy tensor. The basic elements for computing the scattering are given.
• The absorptive nature of the scattering coefficient in the stress-energy tensor formalism for room acoustics
  
  • Polack Jean-Dominique
  • Meacham Aidan
  • Badeau Roland
  , 2022. In the stress-energy tensor formalism, the symmetry between absorption and scattering coefficients, as proven by measurements combined with simulations, is counter-intuitive. By introducing the wall admittance, we show that the scattering coefficient is partly created by the real part of the wall admittance combined with the active intensity, that is, is partly due to absorption. However, it also depends on the imaginary part of the wall admittance in combination with the reactive intensity, which confers it genuine scattering properties. In the case of plane waves impinging on planar boundary, the admittance formalism shows that reactive intensity vanishes in directions parallel to the wall; when the source is at finite distance from the wall, a residual reactive intensity subsists. However, for curved boundaries, the velocity in directions parallel to the wall is no longer proportional to the pressure, and scattering occurs.
• OSATE-DIM solves the instance model-view update problem in AADL
  
  • Mittal Rakshit
  • Blouin Dominique
  , 2022, pp.1-6. AADL (Architecture Analysis and Design Language) is a rich modeling language for embedded systems through several constructs such as component extension and refinement to promote modularity of declarative specifications. To ease the processing of AADL models by tools, OSATE, the reference tool for AADL, defines another model computed from declarative models which results in a single tree system where all information is readily available. Tools can efficiently use this readily available information to analyze the system. An automated backward transformation (deinstantiation) from instance models to declarative models is missing to update the corresponding declarative specification given changes that have been performed on the instance model. Since the instance model is a 'view' of the declarative model, this is a view-update problem. We demonstrate the OSATE Declarative-Instance Mapping Tool (OSATE-DIM), to perform incremental deinstantiation in AADL. This tool significantly eases the development of AADL analysis and code generation tools. (10.1145/3550356.3559083)
  DOI : 10.1145/3550356.3559083
• Solving the instance model-view update problem in AADL
  
  • Mittal Rakshit
  • Blouin Dominique
  • Bhobe Anish
  • Bandyopadhyay Soumyadip
  , 2022, pp.55-65. (10.1145/3550355.3552396)
  DOI : 10.1145/3550355.3552396
• Multi-paradigm modeling for early analysis of ROS-based robotic applications using a library of AADL models
  
  • Senn Eric
  • Bourdon Lucie
  • Blouin Dominique
  , 2022, pp.677-683. (10.1145/3550356.3563129)
  DOI : 10.1145/3550356.3563129
• Multi-scale model-based explanations for cyber-physical systems
  
  • Diaconescu Ada
  • Houze Etienne
  • Dessalles Jean-Louis
  • Vangheluwe Hans
  • Franceschini Romain
  , 2022, pp.684-691. Automated control in Cyber-Physical Systems (CPS) generates behaviours that may surprise non-expert users. Relevant explanations are required to maintain user trust. Large CPS (e.g., autonomous car networks and smart grids) raise additional scaleability issues for the explanatory processes and complexity issues for generated explanations. We propose a multi-scale system modelling and explanation technique to address these concerns. The idea is to increase the scale, or abstraction level, of the modelled CPS, whenever possible without loss of salient information, so as to produce smaller system representations and hence to reduce the complexity of the explanatory process and of the generated explanations. We illustrate our proposal via an urban traffic case study, modelling traffic at two different scales (i.e., modelling individual cars at a lower-scale; and traffic jams at a higher-scale). We show how a multi-scale explanatory process can use the lower- and higher-scale models to generate either longer (more detailed) explanations, or shorter (more abstract) explanations, respectively. This proof-of-concept illustration offers a basis for further research towards a comprehensive multi-scale explanatory solution for CPS. (10.1145/3550356.3561554)
  DOI : 10.1145/3550356.3561554
• Centralized architecture for ECU security management in connected and autonomous vehicles
  
  • Khemissa Hamza
  • Urien Pascal
  , 2022, pp.1409-1414. The strong development of automotive industry is changing traditional perceptions towards a vision of connected and autonomous vehicles (CAVs), such as each vehicle consists of a number of networked computer components, called Electronic Control Units (ECUs) in order to achieve numerous automotive services. Controller Area Network (CAN) is primarily designed for automotive networking with little regard to security. Indeed, the lack of authentication and confidentiality features could lead to automotive cyberattacks putting at risk the safety of the driver, the pedestrians and other vehicles. Therefore, identity management, authentication and data confidentiality must be handled efficiently. In this paper, we propose a centralized architecture for ECU security management in CAVs. First, we present a lightweight symmetric cryptography based session key agreement scheme between each ECU and the manufacturer data center, which uses a random nonce, concatenation operator, a simple hash function and a keyed-hash message authentication code (HMAC). Then, we define the configuration and security parameters on the CAN bus. Finally, we discuss our proposal. To the best of our knowledge, no prior works have been proposed for the establishment of a session key between each ECU and the data center. (10.1109/ICTC55196.2022.9952757)
  DOI : 10.1109/ICTC55196.2022.9952757
• Everlasting secure key agreement from the quantum computational timelock
  
  • Alleaume Romain
  • Vyas Nilesh
  , 2022.
• Domain Adaptation for Stance Detection towards Unseen Target on Social Media
  
  • Deng Ruofan
  • Panl Li
  • Clavel Chloé
  , 2022, pp.1-8. Stance detection aims at identifying people's stand-point towards a given target. New targets are constantly appearing on social media, making previous annotated data unusable by stance detection models relying on classical supervised machine learning. Thus, cross-target stance detection which uses labeled data from source targets to learn a model that can be adapted to the destination new target, has become a prevailing research direction. However, previous methods rely on manually chosen similar source-destination target pairs and lack generalization to unseen targets with no explicit relation to known ones. To this end, we investigate the problem from a domain adaptation perspective and further propose a novel Unified Target-aware Domain Adaptation method (UTDA) that leverages knowledge transfer capability of transformer-based language model. The proposed method can effectively extract critical target-shared features for detecting stance by feature disentanglement and automatically learn to identify target relations. UTDA can easily be applied to a new unseen target since it does not rely on any pre-defined target pairs. Experimental results on two benchmark stance datasets demonstrate that our method achieves better performance than strong baselines (10.1109/ACII55700.2022.9953818)
  DOI : 10.1109/ACII55700.2022.9953818
• Dynamic Defenses for Improved Resilience of Connected Cars
  
  • Ayrault Maxime
  , 2022. With the advent of connected cars, new security threats need to be faced. There are mainly two factors that make up the severity of these threats: Firstly, the attack surface is growing with the ever increasing use of software-driven electronic components in the car and especially with every new interface that connects the car to the internet and the outside world. Secondly, the potential impact of security vulnerabilities is growing with the car electronics taking over more and more safety critical functionalities, such as “brake-by-wire” or advanced driver-assistance systems. In the recent years, new attacks have been published that make use of wireless connections in order to take over the control of a car. With these new attack vectors and the growing complexity of the on-board units, safety and security are becoming a major design objective of new automotive systems. The term resilience by design refers to the goal of securing the overall system architecture instead of deploying local security patches. This includes the detection of intrusions or defects and a coordinated protection against these threats. Bio-inspired approaches use the natural resitance of biological organism as a blueprint to propose technical solutions to this challenge. As an example, the principle of a moving target defense is to change the configuration of a system so as to make deterministic attacks impractical. This defense pattern can be observed in many viruses – such as HIV – which are constantly changing the surface proteins exposed to the outside world so as to escape attacks from the immune system. In this project, we would like to study the potential of architectural reconfigurations as means of a moving target defense against cyber attacks and defects in a connected car environment. As a prerequisite for this work, a precise and formal architectural model will be constructed. The main idea is to examine the space of functionally correct configurations of the architecture. In case of a detected intrusion, the system switches in a non-deterministic manner to some remaining configuration. The decision of when and how to trigger a reconfiguration is based on the attack tree: A possible strategy is to keep the system in a state where a maximum number of legal configurations remains in order to be able to react to evolving attacks.