Decision Awareness in Reinforcement Learning

Explore decision awareness in reinforcement learning through this comprehensive lecture by Pierre-Luc Bacon from Université de Montréal. Delve into the learning principle that optimizes system components to produce optimal decisions, focusing on recent advances in model-based reinforcement learning. Examine control-oriented transition models learned through implicit differentiation and the end-to-end learning of neural ordinary differential equations for nonlinear trajectory optimization. Gain insights into computational challenges and scaling techniques, including efficient Jacobian factorization in forward mode automatic differentiation and novel constrained optimizers inspired by adversarial learning. This DS4DM Coffee Talk, presented at the GERAD Research Center, offers a deep dive into cutting-edge reinforcement learning concepts and methodologies.

Decision Awareness in Reinforcement Learning, Pierre-Luc Bacon