References

ODRL is motivated and built on top of some commonly adopted environments and benchmarks, including:

• OpenAI Gym
• D4RL
• Meta-World

ODRL is also motivated by the following off-dynamics RL papers. We highly recommend the users to read these papers:

• Off-dynamics reinforcement learning: Training for transfer with domain classifiers
• Cross-domain policy adaptation via value-guided data filtering
• Cross-domain policy adaptation by capturing representation mismatch
• When to trust your simulator: Dynamics-aware hybrid offline-and-online reinforcement learning
• Dara: Dynamics-aware reward augmentation in offline reinforcement learning