Waypoint Transformer: Reinforcement Learning via Supervised Learning with Intermediate Targets

Anirudhan Badrinath, Allen Nie, Yannis Flet-Berliac, Emma Brunskill

Confirmation: I have read and confirm that at least one author will be attending the workshop in person if the submission is accepted

Keywords: offline reinforcement learning, reinforcement learning via supervised learning, goal-conditioning, behaviour cloning

TL;DR: Using intermediate target goals as conditioning variables with simple behavioral cloning objectives, we can achieve state-of-the-art performance in offline goal-conditioned reinforcement learning.

Abstract:

Despite the recent advancements in offline reinforcement learning via supervised learning (RvS) methods and the success of the decision transformer (DT) architecture in various domains, DTs have proven to fall short in challenging benchmarks. The root cause of this underperformance lies in their inability to seamlessly connect segments of suboptimal trajectories, i.e., stitch, leading to poor performance. To overcome these limitations, we present a novel approach to enhance RvS methods by integrating intermediate targets. We introduce the waypoint transformer (WT), using an architecture that builds upon the DT framework and is further conditioned on dynamically-generated waypoints. The results show a significant improvement in the final return compared to existing RvS methods, with performance on par or greater than existing temporal difference learning-based methods. Additionally, the performance and stability is significantly improved
in the most challenging environments and data configurations, including AntMaze Large Play/Diverse and Kitchen Mixed/Partial.

Submission Number: 10