Jiaheng Feng

SUF: Stabilized Unconstrained Fine-Tuning for Offline-to-Online Reinforcement Learning

Offline-to-online reinforcement learning (RL) provides a promising solution to improving suboptimal offline pre-trained policies through online fine-tuning. However, one efficient method, unconstrained fine-tuning, often suffers from severe policy collapse due to excessive distribution shift. To ens…