Offline Reinforcement Learning

We aim at understanding the principles underpinning learning and generalization, to build reliable AI systems that can learn more efficiently from available data, intelligently gather additional relevant data, and quickly adapt to and reason about unusual scenarios when deployed in…

MSR Montréal focuses on improving the understanding of fundamental concepts in (deep) Reinforcement Learning (RL) and addressing the open problems that need to be overcome to employ RL on a large scale in the real world.