Offline Multi-Task Reinforcement Learning

Batch Multi-Task Learning

Approach where the agent learns to solve multiple tasks using only pre-collected data, without online exploration during training.

← Wstecz