Publications

2021

• Brief Announcement: Accountability and Reconfiguration -Self-Healing Lattice Agreement
  
  • Freitas de Souza Luciano
  • Kuznetsov Petr
  • Rieutord Thibault
  • Tucci-Piergiovanni Sara
  , 2021. An accountable distributed system provides means to detect deviations of system components from their expected behavior. It is natural to complement fault detection with a reconfiguration mechanism, so that the system could heal itself, by replacing malfunctioning parts with new ones. In this paper, we describe a framework that can be used to implement a large class of accountable and reconfigurable replicated services. We build atop the fundamental lattice agreement abstraction lying at the core of storage systems and cryptocurrencies. Our asynchronous implementation of accountable lattice agreement ensures that every violation of consistency is followed by an undeniable evidence of misbehavior of a faulty replica. The system can then be seamlessly reconfigured by evicting faulty replicas, adding new ones and merging inconsistent states. We believe that this paper opens a direction towards asynchronous "self-healing" systems that combine accountability and reconfiguration. (10.4230/LIPIcs.DISC.2021.54)
  DOI : 10.4230/LIPIcs.DISC.2021.54
• Coloring the Voronoi tessellation of lattices
  
  • Dutour Sikirić Mathieu
  • Madore David
  • Moustrou Philippe
  • Vallentin Frank
  Journal of the London Mathematical Society, London Mathematical Society ; Wiley, 2021, 104 (3), pp.1135-1171. This material is based upon work supported by the National Science Foundation under Grant No. DMS-1439786 while M.D.S. and P.M. were in residence at the Institute for Computational and Experimental Research in Mathematics in Providence, RI, during the “Point Configurations in Geometry, Physics and Computer Science” semester program. P.M. received support from the Tromsø Research Foundation grant 17_matte_CR. F.V. was partially supported by the SFB/TRR 191 “Symplectic Structures in Geometry, Algebra and Dynamics.” F.V. also gratefully acknowledges support by DFG grant VA 359/1-1. This project has received funding from the European Union's Horizon 2020 research and innovation program under the Marie Skłodowska-Curie agreement number 764759. (10.1112/jlms.12456)
  DOI : 10.1112/jlms.12456
• A Novel Application of Boolean Functions With High Algebraic Immunity in Minimal Codes
  
  • Chen Hang
  • Ding Cunsheng
  • Mesnager Sihem
  • Tang Chunming
  IEEE Transactions on Information Theory, Institute of Electrical and Electronics Engineers, 2021, 67 (10), pp.6856-6867. (10.1109/TIT.2021.3102314)
  DOI : 10.1109/TIT.2021.3102314
• Pour une approche par le design des délibérations
  
  • Frappier Tallulah
  • Huron Samuel
  , 2021. Depuis plusieurs années, de nombreuses plateformes en ligne sont conçues pour accueillir spécifiquement des échanges d’arguments et sont utilisées dans des situations aussi diverses qu’un projet d’urbanisme citadin ou au sein de partis politiques et d’initiatives citoyennes pour afin de définir des prises de décisions collectives. Ces plateformes représentent une partie des technologies civiques, ou civic techs, conçues dans le but de transformer les règles du jeu politique ou d’intensifier les engagements des citoyens dans le cadre des règles existantes de la démocratie représentative. Interroger le design de ces pratiques nous semble nécessaire pour comprendre ces initiatives actuelles qui tentent de modifier notre vie démocratique et influencent nos conceptions politiques ainsi que certaines prises de décisions institutionnelles. Nous proposons ici d’interroger de possibles grilles d’analyse du design de ces outils tout en soulignant l’importance de considérer le design comme partie prenante des processus de délibération.
• Post-Quantum Secure Inner Product Functional Encryption Using Multivariate Public Key Cryptography
  
  • Debnath Sumit Kumar
  • Mesnager Sihem
  • Dey Kunal
  • Kundu Nibedita
  Mediterranean Journal of Mathematics, Springer Verlag, 2021, 18 (5), pp.204. (10.1007/s00009-021-01841-2)
  DOI : 10.1007/s00009-021-01841-2
• On the Capacity Enlargement of Gaussian Broadcast Channels With Passive Noisy Feedback
  
  • Ravi Aditya Narayan
  • Pillai Sibi Raj B.
  • Prabhakaran Vinod M.
  • Wigger Michèle
  IEEE Transactions on Information Theory, Institute of Electrical and Electronics Engineers, 2021, 67 (10), pp.6356-6367. It is well known that the capacity region of an average transmit power constrained Gaussian Broadcast Channel (GBC) with independent noise realizations at the receivers is enlarged by the presence of causal noiseless feedback. When the noise variances at the receivers are identical, even passive feedback via independent memoryless Gaussian links can lead to a capacity region enlargement. The last fact remains true even when the feedback noise variance is very high, and available only from one of the receivers. While such capacity enlargements are feasible for several other feedback models in the Gaussian BC setting, it is also known that feedback does not change the capacity region for physically degraded broadcast channels. In this paper, we consider a two user GBC with independent noise realizations at the receivers, where the feedback links from the receivers are corrupted by independent additive Gaussian noise processes. We investigate the set of four noise variances, two forward and two feedback, for which no capacity enlargement is possible. A sharp characterization of this region is derived, i.e., any quadruple outside the presented region will lead to a capacity enlargement, whereas quadruples inside will leave the capacity region unchanged. Our results lead to the conclusion that when the forward noise variances are different, too noisy a feedback from one of the receivers alone is not always beneficial for enlarging the capacity region, be it from the stronger user or the weaker one, in sharp contrast to the case of equal forward noise variances. (10.1109/TIT.2021.3096639)
  DOI : 10.1109/TIT.2021.3096639
• Chewing-games
  
  • Rabiet Victor
  • Zayana Karim
  • Boyer Ivan
  CultureMath, ENS, 2021. Tire sur l'élastique, la mathématique cherra. D'aucun verra dans l'allongement ou la contraction d'un chewing-gum des taux d'évolution, directs ou réciproques. Par exemple, +16 % dans un sens ou-6,25 % dans l'autre. D'aucun jouera les experts en stéganographie, en inscrivant un message privé sur un strap bien tendu, puis en le relâchant : devenu secret le texte se cache dans les plis ! D'aucun déduira l'aire de l'ellipse de demi-axes a et b, πab, de celle du disque de rayon a, πa 2 , d'une simple affinité. Mais il y a plus sérieux, comme le montreront les deux exemples ci-après abordables en classes de spécialité Mathématiques au lycée : la densité des rationnels parmi les réels d'une part, la propriété fondamentale du logarithme comme transformateur de produits en sommes.
• Making with Data
  
  • Huron Samuel
  • Nagel Till
  • Oehlberg Lora
  • Willett Wesley
  , 2021. While digital technologies have revolutionized how we collect and visually represent data, humans continue the thousand-year tradition of producing physical representations of data. Physical representations of data can range from the mundane (hourglass egg timers in an everyday kitchen) to the spectacular (large-scale data sculptures in a museum). Physical representations of data are experiencing a dramatic renaissance, driven by new fabrication technologies, materials, and processes as well as a growing enthusiasm for all things data. The artists, designers, and scientists who create physical representations of data draw from a range of domains and traditions, and represent a fascinating, inspiring, and revealing cross-section of contemporary maker and data culture. To highlight the diversity of approaches, we are currently curating a collection of first-hand accounts from 25+ artists, designers, and researchers that document the process of designing and creating new physical representations and experiences with data. Each story describes the creators’ motivation and inspiration, their approaches for sourcing and encoding data, and their experience navigating the design and fabrication process. In our talk, we will present five themes that capture how people are “making with data” today: Data craft highlights artists and designers whose hand-crafted pieces manually (and sometimes painstakingly) integrate data into objects – from the extraordinaryexotic and bespoke to the personal and everyday. Digital production examines how digital fabrication techniques like 3D printing and digital milling can produce unique and expressive data-driven physical forms. Data automation introduces new physical platforms that use automation and robotics to dynamically and interactively encode data physically. Participatory showcases ways in which designers invite viewers into the creation process, allowing them to encode or reveal data through their interactions with a piece, material, or other people. Environmental projects, meanwhile, reveal data in the context of our surroundingsnatural environments, often exploring the use of natural processes to create new and compelling representations. Each of these approaches entails profound design choices and considerations which impact the design and production process, the tools and skills required to create the works, and ultimately the connection that is created between the creator, the viewer, and the data In this lighting talk, we will present highlights from our rich and exciting set of art pieces, projects, and installations. We will illustrate each theme with a case study featuring a particularly compelling work, and also provide our own first-hand reflections on creating and experiencing physical representations of data. Finally, we will turn the discussion back to the broader community, examining additional approaches not captured by our themes, and highlighting aspects of the creation process that are of particular interest to the information plus audience.
• EEG-based Decoding of Auditory Attention to a Target Instrument for Neuro-steered Music Source Separation
  
  • Cantisani Giorgia
  • Essid Slim
  • Richard Gael
  , 2021. This paper describes a novel neuro-steered music source separation framework and conducts an extensive evaluation of the proposed system on MAD-EEG, a dataset composed of EEG recordings of subjects attending to a particular in duo and trio music excerpts. We propose an unsupervised non-negative matrix factorisation (NMF) variant, named Contrastive-NMF, that separates a target instrument, guided by the user's selective auditory attention to that instrument, which is tracked in his/her electroencephalographic (EEG) response to music. We analyse the impact of multiple aspects of the musical stimuli, such as the number and type of instruments in the mixture, the spatial rendering and the music genre, obtaining encouraging results, especially in difficult cases where non-informed models struggle. We believe that this unsupervised NMF variant is advantageous for neuro-steered music source separation as it allows us to incorporate additional information in a principled optimisation fashion and does not need training data, which is particularly difficult to acquire for applications involving EEG recording.
• TURIN: A coding system for Trust in hUman Robot INteraction
  
  • Hulcelle Marc
  • Varni Giovanna
  • Rollet Nicolas
  • Clavel Chloé
  , 2021, pp.1-8. (10.1109/ACII52823.2021.9597448)
  DOI : 10.1109/ACII52823.2021.9597448
• Adversarial Attacks in a Multi-view Setting: An Empirical Study of the Adversarial Patches Inter-view Transferability
  
  • Tarchoun Bilel
  • Alouani Ihsen
  • Ben Khalifa Anouar
  • Mahjoub Mohamed Ali
  , 2021, pp.299-302. While machine learning applications are getting mainstream owing to a demonstrated efficiency in solving complex problems, they suffer from inherent vulnerability to adversarial attacks. Adversarial attacks consist of additive noise to an input which can fool a detector. Recently, successful real-world printable adversarial “patches” were proven efficient against state-of-the-art neural networks. In the transition from digital noise based attacks to real-world physical attacks, the myriad of factors affecting object detection will also affect adversarial patches. Among these factors, view angle is one of the most influential, yet under-explored. In this paper, we study the effect of view angle on the effectiveness of an adversarial patch. To this aim, we propose the first approach that considers a multi-view context by combining existing adversarial patches with a perspective geometric transformation in order to simulate the effect of view angle changes. Our approach has been evaluated on two datasets: the first dataset which contains most real world constraints of a multi-view context, and the second dataset which empirically isolates the effect of view angle. The experiments show that view angle significantly affects the performance of adversarial patches, where in some cases the patch loses most of its effectiveness. We believe that these results motivate taking into account the effect of view angles in future adversarial attacks, and open up new opportunities for adversarial defenses. (10.1109/CW52790.2021.00057)
  DOI : 10.1109/CW52790.2021.00057
• Visualizations as Data Input?
  
  • Huron Samuel
  • Willett Wesley
  , 2021. We examine ``input visualizations'', visual representations that are designed to collect (and represent) new data rather than encode pre-existing datasets. Information visualization is commonly used to reveal insights and stories within existing data. As a result, most contemporary visualization approaches assume existing data sets or data structures as the starting point for design, through which that data will be mapped to visual encodings to produce final visualizations. Meanwhile, the implications of visualizations as inputs and as data sources have received extremely little attention---despite the existence of visual and physical examples stretching back centuries---and the benefits, trades-offs, design patterns, and even the language necessary to describe them remain unexplored. In this paper we argue for the deeper examination of input visualizations, highlighting a set of recent examples and introducing vocabulary for describing them. Finally, we present a series of provocations which examine some of the challenges posed by input visualizations and suggest opportunities for better understanding this type of visual representations and their potential.
• How ECA vs Human Leaders Affect the Perception of Transactive Memory System (TMS) in a Team
  
  • Biancardi Beatrice
  • O'Toole Patrick
  • Giaccaglia Ivan
  • Ravenet Brian
  • Pitt Ian
  • Mancini Maurizio
  • Varni Giovanna
  , 2021, pp.1-8. Transactive Memory System (TMS) is a mental representation of the distribution of knowledge between the members of a team. Can an Embodied Conversational Agent perform as well as a Human when intervening as a leader to support the development of the team’s TMS? And, if yes, are there differences in the way the team perceives their respective interventions? In this paper, a perceptive online study is conducted on how Human leader interventions affect the perception of a team’s TMS. The results are compared to the ones from a previous study evaluating an Embodied Conversational agent leader rather than a human one. Both the agent and the human adopt nonverbal behaviors characterizing 2 leadership styles: Transformational (TFL) and Transactional (TAL). TFL is expected to stimulate team members curiosity and creativity in problem-solving; instead, TAL emphasizes the role of the leader in supervising the team, providing it with feedback when needed. The results show that the intervention from both the agent and the human are perceived to potentially improve the perceived TMS of a team. Another interesting insight is that the TFL style works better when performed by the Human, where both the TAL and TFL style perform well when realized by the agent. (10.1109/ACII52823.2021.9597454)
  DOI : 10.1109/ACII52823.2021.9597454
• Don’t Judge Me by My Face: An Indirect Adversarial Approach to Remove Sensitive Information From Multimodal Neural Representation in Asynchronous Job Video Interviews
  
  • Hemamou Leo
  • Guillon Arthur
  • Martin Jean-Claude
  • Clavel Chloé
  , 2021, pp.1-8. Use of machine learning for automatic analysis of job interview videos has recently seen increased interest. Despite claims of fair output regarding sensitive information such as gender or ethnicity of the candidates, the current approaches rarely provide proof of unbiased decision-making, or that sensitive information is not used. Recently, adversarial methods have been proved to effectively remove sensitive information from the latent representation of neural networks. However, these methods rely on the use of explicitly labeled protected variables (e.g. gender), which cannot be collected in the context of recruiting in some countries (e.g. France). In this article, we propose a new adversarial approach to remove sensitive information from the latent representation of neural networks without the need to collect any sensitive variable. Using only a few frames of the interview, we train our model to not be able to find the face of the candidate related to the job interview in the inner layers of the model. This, in turn, allows us to remove relevant private information from these layers. Comparing our approach to a standard baseline on a public dataset with gender and ethnicity annotations, we show that it effectively removes sensitive information from the main network. Moreover, to the best of our knowledge, this is the first application of adversarial techniques for obtaining a multimodal fair representation in the context of video job interviews. In summary, our contributions aim at improving fairness of the upcoming automatic systems processing videos of job interviews for equality in job selection. (10.1109/ACII52823.2021.9597443)
  DOI : 10.1109/ACII52823.2021.9597443
• Using Valence Emotion to Predict Group Cohesion’s Dynamics: Top-down and Bottom-up Approaches
  
  • Maman Lucien
  • Chetouani Mohamed
  • Likforman-Sulem Laurence
  • Varni Giovanna
  , 2021.
• Power allocation in Uplink Multiband Satellite System with Nonlinearity-Aware Receiver
  
  • Louchart Arthur
  • Ciblat Philippe
  • Poulliat Charly
  , 2021. In this paper, we address power allocation for an uplink multiband single-beam satellite system taking into account on-board nonlinearities. These nonlinearities are generated by the high power amplifier. Based on the closed-form expression of the capacity associated with the optimal receiver exploiting the structure of the nonlinear effects, three power allocations are studied: maximization of the sum-rate, maximization of the minimum of the rates, and minimization of the sum of the powers. The related optimization problems boil down nonconvex problems that can be cast and solved by using signomial programming. We propose practical and scalable algorithms for fixing these power allocation problems.
• To do or not to do: finding causal relations in smart homes
  
  • Fadiga Kanvaly
  • Diaconescu Ada
  • Dessalles Jean-Louis
  • Houze Etienne
  , 2021, pp.110-119. Research in Cognitive Science suggests that humans understand and represent knowledge of the world through causal relationships. In addition to observations, they can rely on experimenting and counterfactual reasoning – i.e. referring to an alternative course of events – to identify causal relations and explain atypical situations. Different instances of control systems, such as smart homes, would benefit from having a similar causal model, as it would help the user understand the logic of the system and better react when needed. However, while data-driven methods achieve high levels of correlation detection, they mainly fall short of finding causal relations, notably being limited to observations only. Notably, they struggle to identify the cause from the effect when detecting a correlation between two variables. This paper introduces a new way to learn causal models from a mixture of experiments on the environment and observational data. The core of our method is the use of selected interventions, especially our learning takes into account the variables where it is impossible to intervene, unlike other approaches. The causal model we obtain is then used to generate Causal Bayesian Networks, which can be later used to perform diagnostic and predictive inference. We use our method on a smart home simulation, a use case where knowing causal relations pave the way towards explainable systems. Our algorithm succeeds in generating a Causal Bayesian Network close to the simulation’s ground truth causal interactions, showing encouraging prospects for application in real-life systems. (10.1109/ACSOS52086.2021.00030)
  DOI : 10.1109/ACSOS52086.2021.00030
• Concentration Inequalities for Two-Sample Rank Processes with Application to Bipartite Ranking
  
  • Clémençon Stéphan
  • Limnios Myrto
  • Vayatis Nicolas
  Electronic Journal of Statistics, Shaker Heights, OH : Institute of Mathematical Statistics, 2021, 15 (2), pp.4659 -- 4717. The ROC curve is the gold standard for measuring the performance of a test/scoring statistic regarding its capacity to discriminate between two statistical populations in a wide variety of applications, ranging from anomaly detection in signal processing to information retrieval, through medical diagnosis. Most practical performance measures used in scoring/ranking applications such as the AUC, the local AUC, the p-norm push, the DCG and others, can be viewed as summaries of the ROC curve. In this paper, the fact that most of these empirical criteria can be expressed as two-sample linear rank statistics is highlighted and concentration inequalities for collections of such random variables, referred to as two-sample rank processes here, are proved, when indexed by VC classes of scoring functions. Based on these nonasymptotic bounds, the generalization capacity of empirical maximizers of a wide class of ranking performance criteria is next investigated from a theoretical perspective. It is also supported by empirical evidence through convincing numerical experiments. (10.1214/21-EJS1907)
  DOI : 10.1214/21-EJS1907
• Detection of abnormal folding patterns with unsupervised deep generative models
  
  • Guillon Louise
  • Cagna Bastien
  • Dufumier Benoit
  • Chavas Joël
  • Rivière Denis
  • Mangin Jean-François
  , 2021. Although the main structures of cortical folding are present in each human brain, the folding pattern is unique to each individual. Because of this large normal variability, the identification of abnormal patterns associated to developmental disorders is a complex open challenge. In this paper, we tackle this problem as an anomaly detection task and explore the potential of deep generative models using benchmarks made up of synthetic anomalies. To focus learning on the folding geometry, brain MRI are preprocessed first to deal only with a skeleton-based negative cast of the cortex. A variational auto-encoder is trained to get a representation of the regional variability of the folding pattern of the general population. Then several synthetic benchmark datasets of abnormalities are designed. The latent space expressivity is assessed through classification experiments between control's and abnormal's latent codes. Finally, the properties encoded in the latent space are analyzed through perturbation of specific latent dimensions and observation of the resulting modification of the reconstructed images. The results have shown that the latent representation is rich enough to distinguish subtle differences like asymmetries between the right and left hemispheres.
• Timing configurations affect the macro-properties of multi-scale feedback systems
  
  • Mellodge Patricia
  • Diaconescu Ada
  • Di Felice Louisa Jane
  , 2021, 2021, pp.100-109. Multi-scale feedback systems, where information cycles through micro- and macro-scales leading to adaptation, are ubiquitous across domains, from animal societies and human organisations to electric grids and neural networks. Studies on the effects of timing on system properties are often domain specific. The Multi-Scale Abstraction Feedbacks (MSAF) design pattern aims to generalise the modelling of multi-scale systems where feedback occurs across scales. We expand on MSAF to include timing concerns and illustrate their effects via two models: a hierarchical oscillator (HO) and a hierarchical cellular automata (HCA). Results show how (i) different timing configurations significantly affect system macro-properties and (ii) different regions of time configurations can lead to the same macro-properties. These results contribute to theory, while also providing useful insights for designing and controlling such systems. (10.1109/ACSOS52086.2021.00032)
  DOI : 10.1109/ACSOS52086.2021.00032
• Contrastive Learning with Continuous Proxy Meta-Data for 3D MRI Classification
  
  • Dufumier Benoit
  • Gori Pietro
  • Victor Julie
  • Grigis Antoine
  • Wessa Michel
  • Brambilla Paolo
  • Favre Pauline
  • Polosan Mircea
  • Mcdonald Colm
  • Piguet Camille Marie
  • Duchesnay Edouard
  , 2021. Traditional supervised learning with deep neural networks requires a tremendous amount of labelled data to converge to a good solution. For 3D medical images, it is often impractical to build a large homogeneous annotated dataset for a specific pathology. Self-supervised methods offer a new way to learn a representation of the images in an unsupervised manner with a neural network. In particular, contrastive learning has shown great promises by (almost) matching the performance of fully-supervised CNN on vision tasks. Nonetheless, this method does not take advantage of available meta-data, such as participant's age, viewed as prior knowledge. Here, we propose to leverage continuous proxy metadata, in the contrastive learning framework, by introducing a new loss called y-Aware InfoNCE loss. Specifically, we improve the positive sampling during pre-training by adding more positive examples with similar proxy meta-data with the anchor, assuming they share similar discriminative semantic features.With our method, a 3D CNN model pre-trained on 10 4 multi-site healthy brain MRI scans can extract relevant features for three classification tasks: schizophrenia, bipolar diagnosis and Alzheimer's detection. When fine-tuned, it also outperforms 3D CNN trained from scratch on these tasks, as well as state-of-the-art self-supervised methods. Our code is made publicly available here.
• Innovative Wallet Using Trusted On-Line Keystore
  
  • Urien Pascal
  , 2021, pp.12-14. This demonstration presents a prototype wallet, secured by an innovative on-line keystore, based on TLS1.3 server running in secure element. The main idea is to use via internet, keystore dedicated to blockchain, thanks to command shell, running over TLS1.3 server embedded in secure element. TLS1.3 exchanges are secured thanks to 32 bytes pre-shared-key (PSK), which acts as a strong PIN for secure element. A cheap keystore board is detailed, based on Wi-Fi SoC and Javacard 3.04. On the wallet side, a TLS 1.3 client uses a dedicated smartcard that protects PSK from eavesdropping. (10.1109/BRAINS52497.2021.9569783)
  DOI : 10.1109/BRAINS52497.2021.9569783
• ORSUM 2021 - 4th Workshop on Online Recommender Systems and User Modeling
  
  • Vinagre João
  • Jorge Alípio Mário
  • Al-Ghossein Marie
  • Bifet Albert
  , 2021, pp.792--793. Modern online services continuously generate data at very fast rates. This continuous flow of data encompasses content – e.g. posts, news, products, comments –, but also user feedback – e.g. ratings, views, reads, clicks –, together with context data – user device, spacial or temporal data, user task or activity, weather. This can be overwhelming for systems and algorithms designed to train in batches, given the continuous and potentially fast change of content, context and user preferences or intents. Therefore, it is important to investigate online methods able to transparently adapt to the inherent dynamics of online services. Incremental models that learn from data streams are gaining attention in the recommender systems community, given their natural ability to deal with the continuous flows of data generated in dynamic, complex environments. User modeling and personalization can particularly benefit from algorithms capable of maintaining models incrementally and online. The objective of this workshop is to foster contributions and bring together a growing community of researchers and practitioners interested in online, adaptive approaches to user modeling, recommendation and personalization, and their implications regarding multiple dimensions, such as evaluation, reproducibility, privacy and explainability. (10.1145/3460231.3470940)
  DOI : 10.1145/3460231.3470940
• Quantum cryptography and its application frontiers
  
  • Alleaume Romain
  , 2021.
• Towards Compositional Verification of Synchronous Reactive System
  
  • Chabane Sarah
  • Ameur-Boulifa Rabéa
  • Mezghiche Mohamed
  International Journal of Critical Computer-Based Systems, Inderscience, 2021, 10, pp.120-142.