Offline Multi-Task Reinforcement Learning

Multi-Task Batch Constrained Q-Learning

Extension of BCQ to the multi-task context where the Q-function is constrained by batch data while sharing knowledge between tasks.

← Terug