Loris Gaven

About

I am a second-year PhD student at Inria, France, in the Flowers team supervised by Pierre-Yves Oudeyer. My research focuses on Reinforcement Learning, open-ended learning, and curiosity-driven exploration, with the goal of building autonomous agents capable of learning without human supervision, solely through interaction with their environment.

News

• Jan 2026I gave a talk titled "Toward Artificial Curiosity" at the University of Padua.
• Oct 2025DebateQD was accepted at the Multi-Turn Interactions in Large Language Models Workshop at NeurIPS.
• Sep 2025We presented MAGELLAN and HERAKLES at the IMOL Workshop in Hatfield.
• Aug 2025I gave a talk on MAGELLAN at the Metacognitive Satellite of CCN in Amsterdam.
• Jul 2025I presented MAGELLAN at the ICML poster session in Vancouver.
• Jul 2025I gave an online talk on MAGELLAN (YouTube link).
• May 2025MAGELLAN was accepted at ICML.
• Dec 2024I presented SAC-GLAM at the IMOL Workshop at NeurIPS in Vancouver.
• Oct 2024I started my PhD with the Inria Flowers team.

Selected Publications

• MAGELLAN: Metacognitive predictions of learning progress guide autotelic LLM agents in large goal spaces L. Gaven, T. Carta, C. Romac, C. Colas, S. Lamprier, O. Sigaud, P-Y. Oudeyer ICML, 2025

All Publications

• MAGELLAN: Metacognitive predictions of learning progress guide autotelic LLM agents in large goal spaces L. Gaven, T. Carta, C. Romac, C. Colas, S. Lamprier, O. Sigaud, P-Y. Oudeyer ICML, 2025
• HERAKLES: Hierarchical Skill Compilation for Open-ended LLM Agents T. Carta, C. Romac, L. Gaven, P-Y. Oudeyer, O. Sigaud, S. Lamprier arXiv preprint, 2025
• Optimizing for Persuasion Improves LLM Generalization: Evidence from Quality-Diversity Evolution of Debate Strategies A.J. Reedi, C. Léger, J. Pourcel, L. Gaven, P. Charriau, G. Pourcel arXiv preprint, 2025
• SAC-GLAM: Improving Online RL for LLM agents with Soft Actor-Critic and Hindsight Relabeling L. Gaven, C. Romac, T. Carta, S. Lamprier, O. Sigaud, P-Y. Oudeyer IMOL Workshop - NeurIPS, 2024

Links