Efficient Online RL Fine Tuning with Offline Pre-trained Policy Only
This addresses a problem for RL practitioners by enabling efficient fine-tuning in scenarios where only pre-trained policies are available, such as after imitation learning, though it appears incremental as it builds on existing offline-to-online RL methods.
The paper tackles the challenge of improving pre-trained policies through online RL fine-tuning without requiring pre-trained Q-functions, which often hinder exploration due to conservatism, and proposes PORL, a method that initializes Q-functions from scratch to achieve competitive performance with existing algorithms.
Improving the performance of pre-trained policies through online reinforcement learning (RL) is a critical yet challenging topic. Existing online RL fine-tuning methods require continued training with offline pretrained Q-functions for stability and performance. However, these offline pretrained Q-functions commonly underestimate state-action pairs beyond the offline dataset due to the conservatism in most offline RL methods, which hinders further exploration when transitioning from the offline to the online setting. Additionally, this requirement limits their applicability in scenarios where only pre-trained policies are available but pre-trained Q-functions are absent, such as in imitation learning (IL) pre-training. To address these challenges, we propose a method for efficient online RL fine-tuning using solely the offline pre-trained policy, eliminating reliance on pre-trained Q-functions. We introduce PORL (Policy-Only Reinforcement Learning Fine-Tuning), which rapidly initializes the Q-function from scratch during the online phase to avoid detrimental pessimism. Our method not only achieves competitive performance with advanced offline-to-online RL algorithms and online RL approaches that leverage data or policies prior, but also pioneers a new path for directly fine-tuning behavior cloning (BC) policies.