Joelle Pineau

Joelle Pineau challenges the community to rethink the fundamental objectives in reinforcement learning. Moving beyond the classical Bellman equation framework, she discusses alternative formulations of value that may better capture what we truly want agents to optimize for, touching on topics of safety, fairness, multi-objective optimization, and human-aligned reward design.

Joelle Pineau is a Professor at McGill University, Vice President of AI Research at Meta (FAIR), and a core member of Mila – Quebec AI Institute. She is known for her work on partially observable Markov decision processes, dialogue systems, and reproducibility in machine learning.

She has advocated strongly for reproducibility standards in AI research and has received the NSERC E.W.R. Steacie Memorial Fellowship. Her research focuses on developing responsible AI systems with applications in healthcare and robotics.