Home » Teams » ACIDE » Mehdi Khamassi

  • Mehdi Khamassi

  • Directeur de recherche
  • Équipe: ACIDE
  • Bureau: 65-66/306
  • Email: mehdi (dot) khamassi (at) sorbonne-universite (dot) fr
  • Telephone:+33 6 50 76 44 92
  • Addresse: ISIR, Campus Pierre et Marie Curie, 4 place Jussieu, BC173, 75005 Paris
  • Site web: https://pages2.isir.upmc.fr/mkhamassi/
  • Bio: Mehdi Khamassi est directeur de recherche employé par le Centre National de la Recherche Scientifique (CNRS), affecté à l'Institut des Systèmes Intelligents et de Robotique (ISIR), sur le campus de Sorbonne Université, Paris, France. Il a une formation d'ingénieur (Ecole Nationale Supérieure d'Informatique pour l'Industrie et l'Entreprise / Conservatoire National des Arts et Métiers, Évry, 2003 ; spécialisations en intelligence artificielle et modélisation statistique). Il est également diplômé du Cogmaster (2003) par l'Université Pierre et Marie Curie (UPMC), Paris. Il a ensuite effectué une thèse de neurosciences cognitives entre l'UPMC et le Collège de France (2007). Il a été recruté en 2010 par le CNRS comme chargé de recherches. Il est co-organisateur de la conférence Symposium of Biology of Decision-Making (SBDM) depuis 2012. Depuis 2015, il est également membre du conseil pédagogique, anciennement co-directeur des études, et maintenant co-responsable de la majeure modélisation, pour le Cogmaster (Université Paris Sciences Lettres (Ecole Normale Supérieure) / Université de Paris Cité / EHESS). Il a été chercheur invité au Okinawa Institute of Science and Technology, Japon, en 2008, à University of Trento, Italie, en 2013-2015, à University of Oxford, UK, en 2017-2020, et à la National Technical University of Athens, Grèce depuis 2016. Il est rédacteur en chef d'Intellectica, et éditeur pour plusieurs autres journaux comme Frontiers in Neurorobotics, Frontiers in Decision Neuroscience, ReScience X, et Neurons, Behavior, Data analysis and Theory. Ses thèmes de recherche principaux incluent la prise de décision et l'apprentissage par renforcement chez les animaux et les robots, le rôle des différents types de récompenses (sociales, non-sociales, informationnelles) dans l'apprentissage, et les questions éthiques soulevées par la prise de décision autonome des machines. Ses principales méthodes sont la modélisation computationnelle, la conception d'expériences neurobiologiques pour tester les prédictions des modèles, l'analyse de données expérimentales, la conception d'algorithmes d'IA pour les robots, et l'expérimentation comportementale impliquant des humains, des animaux non-humains ou des robots.

Publications

  • Augustin Chartouny, Jean-Pierre Nadal, Mehdi Khamassi. [∼Re] Exploration in Model-based Reinforcement Learning by Empirically Estimating Learning Progress. The ReScience journal, 2024, 73, ⟨10.5281/zenodo.13627804⟩. ⟨hal-04683838⟩
    [ HTTP | PDF ]
  • Mehdi Khamassi, Marceau Nahon, Raja Chatila. Strong and weak alignment of large language models with human values. Scientific Reports, 2024, 14, pp.19399. ⟨10.1038/s41598-024-70031-3⟩. ⟨hal-04670604⟩
    [ HTTP | PDF ]
  • Mehdi Khamassi, Adrien Peyrache, Karim Benchenane, D. A. Hopkins, N. Lebas, et al.. Rat anterior cingulate neurons responsive to rule or strategy changes are modulated by the hippocampal theta rhythm and sharp‐wave ripples. European Journal of Neuroscience, 2024, pp.1-28. ⟨10.1111/ejn.16496⟩. ⟨hal-04673955⟩
    [ HTTP | PDF ]
  • Kathinka Evers, Michele Farisco, Raja Chatila, Brian Earp, Ismael Freire, et al.. Artificial consciousness. Some logical and conceptual preliminaries. 2024. ⟨hal-04670602⟩
    [ HTTP | PDF ]
  • François Cinotti, Etienne Coutureau, Mehdi Khamassi, Alain R Marchand, Benoît Girard. Regulation of reinforcement learning parameters captures long‐term changes in rat behaviour. European Journal of Neuroscience, 2024, ⟨10.1111/ejn.16449⟩. ⟨hal-04630200⟩
    [ HTTP | PDF ]
  • Paris Oikonomou, Athanasios Dometios, Mehdi Khamassi, Costas S Tzafestas. Online prediction of novel trajectories using a library of movement primitives. 2024 IEEE International Conference on Development and Learning (ICDL), May 2024, Austin, Texas (USA), United States. ⟨hal-04542999⟩
    [ HTTP | PDF ]
  • Stefana Broadbent, Florian Forestier, Mehdi Khamassi, Célia Zolynski. Pour une nouvelle culture de l'attention. Odile Jacob, 2024. ⟨hal-04542396⟩
    [ HTTP ]
  • Gianluca Baldassarre, Richard Duro, Emilio Cartoni, Mehdi Khamassi, Alejandro Romero, et al.. Purpose for Open-Ended Learning Robots: A Computational Taxonomy, Definition, and Operationalisation. 2024. ⟨hal-04670576⟩
    [ HTTP | PDF ]
  • Olivier Gapenne, Mehdi Khamassi. Does the enactive approach provide the framework for a new definition of learning?. Intellectica - La revue de l’Association pour la Recherche sur les sciences de la Cognition (ARCo), 2024, Perspective(s) enactive(s) en sciences de l'éducation et de la formation, 80 (1), pp.21-58. ⟨hal-04670326⟩
    [ HTTP | PDF ]
  • Augustin Chartouny, Keivan Amini, Mehdi Khamassi, Benoît Girard. A new paradigm to study social and physical affordances as model-based reinforcement learning. Cognitive Robotics, 2024, 4, pp.142-155. ⟨10.1016/j.cogr.2024.08.001⟩. ⟨hal-04678813⟩
    [ HTTP | PDF ]
  • Paris Oikonomou, Athanasios Dometios, Mehdi Khamassi, Costas Tzafestas. Zero-shot model-free learning of periodic movements for a bio-inspired soft-robotic arm. Frontiers in Robotics and AI, 2023, 10, pp.1256763. ⟨10.3389/frobt.2023.1256763⟩. ⟨hal-04249680⟩
    [ HTTP | PDF ]
  • George Velentzas, Costas S Tzafestas, Mehdi Khamassi. Memory Development with Heteroskedastic Bayesian Last Layer Probabilistic Deep Neural Networks. Workshop on World Models and Predictive Coding in Cognitive Robotics at 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2023), Oct 2023, Detroit (Michigan), United States. ⟨hal-04249890⟩
    [ HTTP | PDF ]
  • Francois Cinotti, Etienne Coutureau, Mehdi Khamassi, Alain Marchand, Benoît Girard. Regulation of the exploration-exploitation trade-off captures long-term changes in rat behaviour. 2023. ⟨hal-04228026⟩
    [ HTTP | PDF ]
  • Rémi Dromnelle, Erwan Renaudo, Mohamed Chetouani, Petros Maragos, Raja Chatila, et al.. Reducing computational cost during robot navigation and human-robot interaction with a human-inspired reinforcement learning architecture. International Journal of Social Robotics, 2023, 15, pp.1297-1323. ⟨10.1007/s12369-022-00942-6⟩. ⟨hal-03829879⟩
    [ HTTP | PDF ]
  • Irene Navarro Lobato, Adrian Aleman-Zapata, Anumita Samanta, Milan Bogers, Shekhar Narayanan, et al.. Increased cortical plasticity leads to memory interference and enhanced hippocampal-cortical interactions. eLife, 2023, ⟨10.7554/eLife.84911⟩. ⟨hal-04112434⟩
    [ HTTP | PDF ]
  • Mehdi Khamassi, Marco Mirolli, Christian Wallraven. Editorial: Neurorobotics explores the human senses. Frontiers in Neurorobotics, 2023, 17, ⟨10.3389/fnbot.2023.1214871⟩. ⟨hal-04102103⟩
    [ HTTP | PDF ]
  • Nasser Soumi, Steve N'Guyen, Mehdi Khamassi. Entre Ciel et Terre. Drôles d'objets 2023, May 2023, Nancy, France. ⟨hal-04700949⟩
    [ HTTP | PDF ]
  • Camille Lakhlifi, François-Xavier Lejeune, Marion Rouault, Mehdi Khamassi, Benjamin Rohaut. Illusion of knowledge in statistics among clinicians: evaluating the alignment between objective accuracy and subjective confidence, an online survey. Cognitive Research: Principles and Implications, 2023, 8, ⟨10.1186/s41235-023-00474-1⟩. ⟨hal-04076913⟩
    [ HTTP | PDF ]
  • Erwan Renaudo, Philipp Zech, Raja Chatila, Mehdi Khamassi. Editorial: Computational Models of Affordance for Robotics. Frontiers in Neurorobotics, 2022, 16, pp.1045355. ⟨10.3389/fnbot.2022.1045355⟩. ⟨hal-03782365⟩
    [ HTTP | PDF ]
  • Mehdi Khamassi, Jean-Gabriel Ganascia. L'IA hier et demain. Audition de Jean-Gabriel Ganascia. Cahiers de TESaCo n°3, 2022, pp.47-72. ⟨hal-03852201⟩
    [ HTTP ]
  • Elisa Massi, Jeanne Barthélemy, Juliane Mailly, Rémi Dromnelle, Julien Canitrot, et al.. Model-Based and Model-Free Replay Mechanisms for Reinforcement Learning in Neurorobotics. Frontiers in Neurorobotics, 2022, 16, pp.864380. ⟨10.3389/fnbot.2022.864380⟩. ⟨hal-03703727⟩
    [ HTTP | PDF ]
  • Paris Oikonomou, Athanasios Dometios, Mehdi Khamassi, Costas Tzafestas. Reproduction of Human Demonstrations with a Soft-Robotic Arm based on a Library of Learned Probabilistic Movement Primitives. 2022 IEEE International Conference on Robotics and Automation (ICRA 2022), May 2022, Philadelphia, PA, United States. ⟨hal-03593923⟩
    [ HTTP | PDF ]
  • Thomas Misiek, Mehdi Khamassi. [Re] A general model of hippocampal and dorsal striatal learning and decision making. ReScience C, 2022, 8 (1), pp.#4. ⟨10.5281/zenodo.6573684⟩. ⟨hal-03676469⟩
    [ HTTP | PDF ]
  • Gilles Bailly, Mehdi Khamassi, Benoît Girard. Computational Model of the Transition from Novice to Expert Interaction Techniques. ACM Transactions on Computer-Human Interaction, 2022, ⟨10.1145/3505557⟩. ⟨hal-03537963⟩
    [ HTTP | PDF ]
  • Neema Moin Afshar, François Cinotti, David A. Martin, Mehdi Khamassi, Donna J Calu, et al.. Reward-mediated, model-free reinforcement-learning mechanisms in Pavlovian and instrumental tasks are related. Journal of Neuroscience, 2022, 43 (3), pp.458-471. ⟨10.1523/JNEUROSCI.1113-22.2022⟩. ⟨hal-03800432⟩
    [ HTTP ]
  • Cyril Monier, Mehdi Khamassi. Liberté & Cognition. Une autre voie est possible. Intellectica - La revue de l’Association pour la Recherche sur les sciences de la Cognition (ARCo), 2021, pp.7-32. ⟨hal-03656536⟩
    [ HTTP | PDF ]
  • Mehdi Khamassi, Jean Lorenceau. Inscription corporelle des dynamiques cognitives et leur impact sur la liberté de l’humain en société. Intellectica - La revue de l’Association pour la Recherche sur les sciences de la Cognition (ARCo), 2021, 75, pp.33-72. ⟨hal-03656562⟩
    [ HTTP | PDF ]
  • Paris Oikonomou, Athanasios Dometios, Mehdi Khamassi, Costas S Tzafestas. Task Driven Skill Learning in a Soft-Robotic Arm. 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, Sep 2021, Prague, Czech Republic. ⟨hal-03275472v2⟩
    [ HTTP | PDF ]
  • Alizée Lopez-Persem, Mehdi Khamassi. Décision et action. Khamassi, M. (Ed.) Neurosciences cognitives, 2021. ⟨hal-03411266⟩
    [ HTTP | PDF ]
  • Mehdi Khamassi (Dir.). Neurosciences cognitives. Éditions de Boeck Supérieur, 2021. ⟨hal-03411280⟩
    [ HTTP ]
  • Anne Collins, Mehdi Khamassi. Initiation à la modélisation computationnelle. Khamassi, M. (Ed.) Neurosciences cognitives, 2021. ⟨hal-03411274⟩
    [ HTTP | PDF ]
  • Lea Roumazeilles, Matthias Schurz, Mathilde Lojkiewiez, Lennart Verhagen, Urs Schüffelgen, et al.. Social prediction modulates activity of macaque superior temporal cortex. Science Advances , 2021, 7 (38). ⟨hal-03346442⟩
    [ HTTP | PDF ]
  • I Rañó, Mehdi Khamassi, K Wong-Lin. Stability Analysis of Bio-inspired Source Seeking with Noisy Sensors. 2021 European Control Conference (ECC), Jul 2021, Kongens Lyngby, Denmark. ⟨hal-03277526⟩
    [ HTTP | PDF ]
  • Mehdi Khamassi. Quelques questions éthiques autour du développement de l’autonomie décisionnelle en intelligence artificielle et en robotique. Cahiers de TESaCo n°2, 2021, pp.23-31. ⟨hal-03854619⟩
    [ HTTP | PDF ]
  • Marios C Panayi, Mehdi Khamassi, Simon Killcross. The rodent lateral orbitofrontal cortex as an arbitrator selecting between model-based and model-free learning systems. Behavioral Neuroscience, 2021, 135 (2), pp.226-244. ⟨10.1037/bne0000454⟩. ⟨hal-03107588v2⟩
    [ HTTP | PDF ]
  • Geoffrey Schoenbaum, Mehdi Khamassi, Mathias Pessiglione, Jay A Gottfried, Elisabeth A Murray. The magical orbitofrontal cortex.. Behavioral Neuroscience, 2021, 135 (2), pp.108 - 108. ⟨10.1037/bne0000470⟩. ⟨hal-03251587⟩
    [ HTTP | PDF ]
  • Mehdi Khamassi. Adaptive coordination of multiple learning strategies in brains and robots. 9th International Conference on the Theory and Practice of Natural Computing (TPNC 2020), Dec 2020, Taoyuan, Taiwan. ⟨10.1007/978-3-030-63000-3_1⟩. ⟨hal-03277525⟩
    [ HTTP | PDF ]
  • Evelien H.S. Schut, Irene Navarro Lobato, Alejandra Alonso, Steven Smits, Mehdi Khamassi, et al.. The Object Space Task reveals increased expression of cumulative memory in a mouse model of Kleefstra syndrome. Neurobiology of Learning and Memory, 2020, 173, pp.107265. ⟨10.1016/j.nlm.2020.107265⟩. ⟨hal-02941978⟩
    [ HTTP | PDF ]
  • Rémi Dromnelle, Benoît Girard, Erwan Renaudo, Raja Chatila, Mehdi Khamassi. Coping with the variability in humans reward during simulated human-robot interactions through the coordination of multiple learning strategies. 29th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN 2020), Aug 2020, Naples (en ligne), Italy. ⟨10.1109/RO-MAN47096.2020.9223451⟩. ⟨hal-02899767v2⟩
    [ HTTP | PDF ]
  • Rémi Dromnelle, Erwan Renaudo, Guillaume Pourcel, Raja Chatila, Benoît Girard, et al.. How to reduce computation time while sparing performance during robot navigation? A neuro-inspired architecture for autonomous shifting between model-based and model-free learning. Living Machines, Jul 2020, Freiburg (on line conference), Germany. ⟨10.1007/978-3-030-64313-3_8⟩. ⟨hal-02883717v3⟩
    [ HTTP | PDF ]
  • Marco K Wittmann, Elsa Fouragnan, Davide Folloni, Miriam C Klein-Flügge, Bolton K H Chau, et al.. Global reward state affects learning and activity in raphe nucleus and anterior insula in monkeys. Nature Communications, 2020, 11 (1), ⟨10.1038/s41467-020-17343-w⟩. ⟨hal-03098645⟩
    [ HTTP | PDF ]
  • Abolfazl Zaraki, Mehdi Khamassi, Luke J Wood, Gabriella Lakatos, Costas S Tzafestas, et al.. A Novel Reinforcement-Based Paradigm for Children to Teach the Humanoid Kaspar Robot. International Journal of Social Robotics, 2020, 12 (3), pp.709-720. ⟨10.1007/s12369-019-00607-x⟩. ⟨hal-02408941⟩
    [ HTTP | PDF ]
  • Mariacarla Staffa, Silvia Rossi, Adriana Tapus, Mehdi Khamassi. Special Issue on Behavior Adaptation, Interaction, and Artificial Perception for Assistive Robotics. International Journal of Social Robotics, 2020, 12, pp.613 - 616. ⟨10.1007/s12369-020-00655-8⟩. ⟨hal-02864719⟩
    [ HTTP | PDF ]
  • Paris Oikonomou, Mehdi Khamassi, Costas S Tzafestas. Periodic movement learning in a soft-robotic arm. IEEE International Conference on Robotics and Automation (ICRA 2020), May 2020, Paris (virtuel), France. ⟨10.1109/ICRA40945.2020.9197035⟩. ⟨hal-03435441⟩
    [ HTTP | PDF ]
  • Stephane Doncieux, Nicolas Bredeche, Léni Kenneth Le Goff, Benoît Girard, Alexandre Coninx, et al.. DREAM Architecture: a Developmental Approach to Open-Ended Learning in Robotics. 2020. ⟨hal-02562103⟩
    [ HTTP | PDF ]
  • Frédéric Alexandre, Peter Ford Dominey, Philippe Gaussier, Benoît Girard, Mehdi Khamassi, et al.. When Artificial Intelligence and Computational Neuroscience meet. Springer. A guided tour of artificial intelligence research, Interfaces and applications of artificial intelligence, 3, 2020, Interfaces and Applications of Artificial Intelligence, 978-3-030-06170-8. ⟨hal-01735123⟩
    [ HTTP | PDF ]
  • Mehdi Khamassi, Benoît Girard. Modeling awake hippocampal reactivations with model-based bidirectional search. Biological Cybernetics (Modeling), 2020, 114 (2), pp.231-248. ⟨10.1007/s00422-020-00817-x⟩. ⟨hal-02504897⟩
    [ HTTP | PDF ]
  • James Gillespie, Iñaki Rañó, Nazmul Siddique, Jose Luis Santos, Mehdi Khamassi. Using Reinforcement Learning to Attenuate for Stochasticity in Robot Navigation Controllers. 2019 IEEE Symposium Series on Computational Intelligence (SSCI 2019), Dec 2019, Xiamen, China. ⟨10.1109/SSCI44817.2019.9002834⟩. ⟨hal-02324129⟩
    [ HTTP | PDF ]
  • François Cinotti, Virginie Fresno, Nassim Aklil, Etienne Coutureau, Benoît Girard, et al.. Dopamine blockade impairs the exploration-exploitation trade-off in rats. Scientific Reports, 2019, 9 (1), ⟨10.1038/s41598-019-43245-z⟩. ⟨hal-02121649⟩
    [ HTTP | PDF ]
  • Jack Hadfield, Georgia Chalvatzaki, Petros Koutras, Mehdi Khamassi, Costas S Tzafestas, et al.. A Deep Learning Approach for Multi-View Engagement Estimation of Children in a Child-Robot Joint Attention Task. 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2019), Nov 2019, Macau, China. ⟨hal-02324118⟩
    [ HTTP | PDF ]