GT A: Generative Trajectory Augmentation with Guidance for Offline Reinforcement Learning

Offline Reinforcement Learning (Offline RL) presents challenges of learning effective decision-making policies from static datasets without any online interactions.