Track 2 Session 3: Spotlights

- DualDICE: Behavior-Agnostic Estimation of Discounted Stationary Distribution Corrections - VIREL: A Variational Inference Framework for Reinforcement Learning - Unsupervised Curricula for Visual Meta-Reinforcement Learning - Policy Continuation with Hindsight Inverse Dynamics - Learning Reward Machines for Partially Observable Reinforcement Learning