Articles by altman

Seminar – Matthew Wiesner – 26/03/2026

26 March 2026

Room 5 – 12h00 Title: Modeling Accent and Language at Scale Abstract: State-of-the-art LID models work reliably for ∼100 languages, but there are orders of magnitude more accents and dialects; annotating all of them is intractable. Furthermore, current LID models fail dramatically when applied to accented speech, and very few annotated accented data exist to support accented TTS. This talk explores these challenges and proposes a potential partial solution using massive amounts of diverse and widely available data collected from radio, with soft language labels in the form of geolocations. The talk then explores the link between robustness to accented speech and the capacity to more effectively model sequence-level data. Finally, these models and insights are shown to greatly improve LID on accented speech, and can be used to mine for accented speech to support scalable, controllable accented TTS.Bio: Matthew Wiesner est chercheur à l’Université Johns Hopkins et chercheur invité au Laboratoire Interdisciplinaire des Sciences du Numérique (LISN). Il a obtenu son doctorat et son master en génie électrique à Johns Hopkins en 2021 et 2016 sous la direction de Jan Trmal et de Sanjeev Khudanpur, ainsi que sa licence en génie électrique à l’Université McGill en 2013. Ses recherches Plus d'infos

PhD Defense – Manh Tuan NGUYEN – 25/03/2026

19 March 2026

Date: Wednesday, March 25th at 2:00 PM Place: amphitheater Blaise at CERI.   The presentation will be held in English. The jury members: Title: Considering the inter-judge variability in the perceptual evaluation of speech and voice disorders and its integration into an automatic decision support system Abstract:Perceptual judgments are widely used in domains that lack clear objective criteria or reliable measurement methods, requiring reliance on human expert evaluation. However, such judgments are inherently subjective, often leading to a lack of agreement and, consequently, variability when multiple experts assess the same material. This variability, referred to as inter-rater variability, is typically addressed by aggregating scores or applying majority voting to produce a consensus decision. While effective for obtaining a final decision, this approach leaves the underlying causes of inter-rater variability largely unexplored.This thesis aims to explain inter-rater variability rather than treating consensus decisions as an absolute reference. We argue that such variability may arise from systematic differences between experts, particularly in terms of professional background, training, and the perceptual dimensions emphasized during assessment. By reducing individual judgments to a single consensus score, traditional approaches implicitly discard valuable information about expert reasoning and decision-making strategies. Understanding and explaining this variability is therefore essential.To address this objective, we propose a computational approach to model and interpret inter-rater variability. Leveraging the pattern-recognition capabilities of Plus d'infos

CORNET Seminar – Younes Ben Mazziane – 20/03/2026

16 March 2026

Date: 20/03/2026 à 11h30 Location: Salle C057 Title: Learning in Proportional Allocation Auctions Games Abstract: The Kelly or proportional allocation mechanism is a simple and efficient auction-based scheme that distributes an infinitely divisible resource proportionally to the agents’ bids. When agents are aware of the allocation rule, their interactions form a game, that has been extensively studied. This paper examines the less explored repeated Kelly game, focusing mainly on utilities that are logarithmic in the allocated resource fraction. We first derive this logarithmic form from fairness–throughput trade-offs in wireless network slicing, and then prove that the induced stage game admits a unique Nash equilibrium (NE). For the repeated play, we prove convergence to this NE under three behavioral models: (i) all agents use Online Gradient Descent (OGD), (ii) all agents use Dual Averaging with a quadratic regularizer (DAQ) (a variant of the Follow-the-Regularized leader algorithm), and (iii) all agents play myopic best responses (BR). Our convergence results hold even when agents use personalized learning rates in OGD and DAQ (e.g., tuned to optimize individual regret bounds), and they extend to a broader class of utilities that meet a certain sufficient condition. Finally, we complement our theoretical results with extensive simulations Plus d'infos

SLG Seminar – Tom Labiausse – 16/03/2026

12 March 2026

16 mars à 13h Tom présentera ses derniers travaux sur Hibiki-Zero, un modèle de traduction simultanée de la parole vers la parole. Les améliorations par rapport à ses travaux précédents, Hibiki, sont vraiment très intéressantes et s’appuient sur une technique d’apprentissage par renforcement (GRPO). Utilisée comme le propose Tom, cette technique évite d’avoir à préparer des données d’apprentissage parole-parole alignées au niveau mot comme dans la première mouture d’Hibiki. Vous êtes invités à assister à cette présentation qui pourrait vous donner des idées d’application à certains de vos travaux. Pour plus d’information, vous pouvez consulter cette page web très instructive et accessible (exemples, code, article) :

Habilitation defense – Rosa Figueiredo – 04/03/2026

25 February 2026

Titre: Combinatorial Optimization Methods for Signed Graph Partitioning Date: 04/03/2026, 15h Lieu: Amphi ADA du CERI Composition du jury Rapporteur·e·s : Safia KEDAD, Professor, Conservatoire National des Arts et Métiers, Paris, France Dritan NACE, Professor, University of Technology of Compiègne, Compiègne, France Hande YAMAN, Professor, Katholieke Universiteit Leuven, Leuven, Belgium Examinateur·rice·s : Paula CARROLL, Professor, University College Dublin, Dublin, Ireland Manoel CAMPELO, Professor, Federal University of Ceará, Fortaleza, Brazil

Cornet seminar – Raghupati Vyas – 27/02/2026

13 February 2026

Vendredi 27 Février à 11h30 en C057. Title: Games with Rational and Herding Players Résumé : This talk examines large-population games with heterogeneous decision-making, in which an α-fraction of players are rational, while the remaining agents exhibit herding behaviour. We introduce a new equilibrium notion, called the α-Rational Nash Equilibrium (α-RNE), and discuss its interpretations. We show that some classical equilibria may disappear and new ones may emerge for small values of α>0. Interestingly, rational players benefit from the presence of herding and may even achieve utility exceeding the socially optimum. Even more strikingly, in some cases, herding players also benefit, attaining utility close to the social optimum. Using transportation and bandwidth sharing games as case studies, we analyse the impact of herding on congestion and resource allocation, and quantify system inefficiencies through the Price of Anarchy. Finally, we discuss the mechanism and influence design in the presence of herding. While the expanded set of equilibria creates new opportunities, it also increases the risk of undesirable outcomes when influence cannot be effectively implemented.

PhD thesis defense – Ahmed Njifenjou – 19/12/2025

18 December 2025

Date: Friday, December 19, at 2 p.m. Lieu: Ada amphitheater at CERI. The defense will be presented in English. Title: Open-Domain Conversational Agents with Transformer-Based Language Models: Toward Multilingualism and Personality  The jury will be composed of: – Lina M. Rojas-Barahona , HDR, Orange Innovation, Reviewer – Didier Schwab, Professor, LIG/GETALP, Université de Grenoble, Reviewer – Sophie Rosset, Professor, LISN, Université Paris Saclay, Examiner – David Traum, Professor, ICT, University of Southern California, Examiner – Bassam Jabaian, Associate Professor, LIA, Université d’Avignon, Thesis Co-supervisor – Fabrice Lefèvre, Professor, LIA, Université d’Avignon,  Thesis Director Abstract: Open-Domain Dialogue (ODD) systems are conversational agents designed for natural and open-ended human interaction. The proliferation of Conversational AI tools like ChatGPT has recently reshaped user expectations; beyond grammatical correctness, users now demand agents that demonstrate contextual understanding, cultural awareness, distinct personality, factual consistency, and other human-like conversational abilities. Despite the impressive progress, ODD systems development has long faced key limitations including strong linguistic bias towards English and Chinese, and the Open-Domain Paradox (ODP) (Skantze and Doğruöz, 2023), which constrains genuine conversational diversity and openness. This dissertation tackles these challenges by exploring multilingual and personality-centric strategies for building controllable and culturally adaptive ODD systems using Transformer-based Language Models. The research progresses along the following complementary axes.  First, we investigate Plus d'infos

PhD thesis defense – Grace Tessa – 16/12/2025

8 December 2025

Date: le mardi 16 décembre à 14h00, Lieu: salle SEO7, Avignon Université, Campus Hannah Arendt. Rapporteurs • Rémi Badonnel — Professor, TELECOM Nancy, University of Lorraine• Lyes Khoukhi — Professor, CNAM Paris University Examinateurs • Yezekael Hayel — Professor, LIA, Avignon University• Tooska Dargahi — Assistant Professor, Manchester Metropolitan University• Ahmed Hemida Anwar — Research Scientist, Devcom Research Lab Encadrants • Vianney Kengne Tchendji — Assistant Professor, URIFIA, University of Dschang• Abderrahim Benslimane — Professor, LIA, Avignon University Résumé de thèse: Titre : Cyber tromperie et Résilience avec l’Apprentissage Fédéré Résumé : L’Intelligence Artificielle (IA), en particulier le Machine Learning (ML) et le Deep Learning (DL), a rapidement transformé de nombreux domaines technologiques, notamment lorsqu’elle est intégrée à des dispositifs connectés tels que les smartphones, les capteurs intelligents et les systèmes IoT. Bien que ces technologies améliorent les services numériques et l’expérience utilisateur, elles soulèvent aussi d’importants enjeux en matière de sécurité et de protection des données personnelles. Les approches traditionnelles d’apprentissage automatique reposent sur la centralisation d’importants volumes de données utilisateurs, ce qui accroît les risques de violation de la vie privée et exige des ressources de calcul et de communication élevées dans un contexte réglementaire strict. L’Apprentissage Fédéré (FL) Plus d'infos

1 2 3 … 8

Laboratoire Informatique d'Avignon — Avignon Université