Offline Imitation Learning Through Graph Search and Retrieval

TL;DR

The paper introduces GSR, a novel offline imitation learning algorithm that organizes interaction data into a graph, enabling effective learning from suboptimal demonstrations for complex robotic manipulation tasks.

Contribution

GSR combines graph search and retrieval with behavior cloning to improve learning stability and success rates from suboptimal data in robotic manipulation.

Findings

GSR achieves 10-30% higher success rates than baselines.

GSR demonstrates effectiveness on complex visual manipulation tasks.

GSR improves proficiency by over 30% in various experiments.

Abstract

Imitation learning is a powerful machine learning algorithm for a robot to acquire manipulation skills. Nevertheless, many real-world manipulation tasks involve precise and dexterous robot-object interactions, which make it difficult for humans to collect high-quality expert demonstrations. As a result, a robot has to learn skills from suboptimal demonstrations and unstructured interactions, which remains a key challenge. Existing works typically use offline deep reinforcement learning (RL) to solve this challenge, but in practice these algorithms are unstable and fragile due to the deadly triad issue. To overcome this problem, we propose GSR, a simple yet effective algorithm that learns from suboptimal demonstrations through Graph Search and Retrieval. We first use pretrained representation to organize the interaction experience into a graph and perform a graph search to calculate the values of different behaviors. Then, we apply a retrieval-based procedure to identify the best behavior (actions) on each state and use behavior cloning to learn that behavior. We evaluate our method in both simulation and real-world robotic manipulation tasks with complex visual inputs, covering various precise and dexterous manipulation skills with objects of different physical properties. GSR can achieve a 10% to 30% higher success rate and over 30% higher proficiency compared to baselines. Our project page is at https://zhaohengyin.github.io/gsr.