Binary Reward Labeling: Bridging Offline Preference and Reward-Based Reinforcement Learning
This work addresses a gap in offline reinforcement learning for researchers and practitioners by enabling preference-based data usage, though it is incremental as it builds on existing reward-based methods.
The authors tackled the problem of applying offline reinforcement learning to preference-based feedback by proposing a framework that transforms preferences into scalar rewards via binary reward labeling, enabling the use of existing reward-based algorithms. The result showed comparable performance to actual reward datasets and outperformed recent baselines in most cases on D4RL benchmarks.
Offline reinforcement learning has become one of the most practical RL settings. However, most existing works on offline RL focus on the standard setting with scalar reward feedback. It remains unknown how to universally transfer the existing rich understanding of offline RL from the reward-based to the preference-based setting. In this work, we propose a general framework to bridge this gap. Our key insight is transforming preference feedback to scalar rewards via binary reward labeling (BRL), and then any reward-based offline RL algorithms can be applied to the dataset with the reward labels. The information loss during the feedback signal transition is minimized with binary reward labeling in the practical learning scenarios. We theoretically show the connection between several recent PBRL techniques and our framework combined with specific offline RL algorithms. By combining reward labeling with different algorithms, our framework can lead to new and potentially more efficient offline PBRL algorithms. We empirically test our framework on preference datasets based on the standard D4RL benchmark. When combined with a variety of efficient reward-based offline RL algorithms, the learning result achieved under our framework is comparable to training the same algorithm on the dataset with actual rewards in many cases and better than the recent PBRL baselines in most cases.