In appendix, we Reward OptimizationWepropose .

We additionally assume that the agent cannot interact with the environment but has access to the action-labeled transition data collected by some agents with unknown qualities.

Inthissection compare compare learned theprior environment 6. Experimental Tostudy proposed [54], where froman learning.

Intelligent agents must be generalists, capable of quickly adapting to various tasks.

However, existing offline RL methods tend to behave poorly during fine-tuning. In this paper, we study the fine-tuning problem in the context of conservative offline RL methods and we devise an approach for learning an effective initialization from offline data that also enables fast online fine-tuning capabilities.

Model-free deepreinforcement learning (RL)algorithms havebeenwidely used for a range of complex control tasks.

In this paper,we propose aprobabilisticmeta-learning algorithm that can sample models for a new task from a modeldistribution.