Offline-to-Online Reinforcement Learning with Classifier-Free Diffusion Generation
This addresses the challenge of costly online interactions in reinforcement learning for researchers and practitioners, though it is incremental as it builds on existing offline-to-online methods.
The paper tackles the problem of offline-to-online reinforcement learning by proposing a new data augmentation approach called Classifier-Free Diffusion Generation (CFDG), which improves generation quality and alignment with online data, resulting in a 15% average performance improvement on the D4RL benchmark.
Offline-to-online Reinforcement Learning (O2O RL) aims to perform online fine-tuning on an offline pre-trained policy to minimize costly online interactions. Existing work used offline datasets to generate data that conform to the online data distribution for data augmentation. However, generated data still exhibits a gap with the online data, limiting overall performance. To address this, we propose a new data augmentation approach, Classifier-Free Diffusion Generation (CFDG). Without introducing additional classifier training overhead, CFDG leverages classifier-free guidance diffusion to significantly enhance the generation quality of offline and online data with different distributions. Additionally, it employs a reweighting method to enable more generated data to align with the online data, enhancing performance while maintaining the agent's stability. Experimental results show that CFDG outperforms replaying the two data types or using a standard diffusion model to generate new data. Our method is versatile and can be integrated with existing offline-to-online RL algorithms. By implementing CFDG to popular methods IQL, PEX and APL, we achieve a notable 15% average improvement in empirical performance on the D4RL benchmark such as MuJoCo and AntMaze.