Learning Manipulation Skills through Robot Chain-of-Thought with Sparse Failure Guidance

TL;DR

This paper introduces a novel approach for robot manipulation skill learning by decomposing tasks into sub-tasks for fine-grained reward guidance using vision-language models, significantly improving success rates.

Contribution

It proposes a task decomposition method combined with VLM-based self imitation learning to enhance reward guidance and accelerate robot skill acquisition.

Findings

Achieves 5.4x higher success rates than RoboCLIP

Outperforms baselines like CLIP and LIV

Provides more informative reward signals through task decomposition

Abstract

Defining reward functions for skill learning has been a long-standing challenge in robotics. Recently, vision-language models (VLMs) have shown promise in defining reward signals for teaching robots manipulation skills. However, existing work often provides reward guidance that is too coarse, leading to insufficient learning processes. In this paper, we address this issue by implementing more fine-grained reward guidance. We decompose tasks into simpler sub-tasks, using this decomposition to offer more informative reward guidance with VLMs. We also propose a VLM-based self imitation learning process to speed up learning. Empirical evidence demonstrates that our algorithm consistently outperforms baselines such as CLIP, LIV, and RoboCLIP. Specifically, our algorithm achieves a $5.4 \times$ higher average success rates compared to the best baseline, RoboCLIP, across a series of manipulation tasks.