Diffusion Meets DAgger: Supercharging Eye-in-hand Imitation Learning

TL;DR

This paper introduces Diffusion Meets DAgger (DMD), a novel method that synthesizes out-of-distribution samples using diffusion models to improve eye-in-hand imitation learning, reducing the need for extensive data collection.

Contribution

DMD combines diffusion models with DAgger to synthesize failure states, enabling robust imitation learning with fewer demonstrations in robotic tasks.

Findings

DMD achieves 80% success with 8 demonstrations in pushing.

DMD outperforms behavior cloning in stacking and pouring tasks.

DMD attains high success rates across multiple robotic manipulation tasks.

Abstract

A common failure mode for policies trained with imitation is compounding execution errors at test time. When the learned policy encounters states that are not present in the expert demonstrations, the policy fails, leading to degenerate behavior. The Dataset Aggregation, or DAgger approach to this problem simply collects more data to cover these failure states. However, in practice, this is often prohibitively expensive. In this work, we propose Diffusion Meets DAgger (DMD), a method to reap the benefits of DAgger without the cost for eye-in-hand imitation learning problems. Instead of collecting new samples to cover out-of-distribution states, DMD uses recent advances in diffusion models to synthesize these samples. This leads to robust performance from few demonstrations. We compare DMD against behavior cloning baseline across four tasks: pushing, stacking, pouring, and shirt hanging. In pushing, DMD achieves 80% success rate with as few as 8 expert demonstrations, where naive behavior cloning reaches only 20%. In stacking, DMD succeeds on average 92% of the time across 5 cups, versus 40% for BC. When pouring coffee beans, DMD transfers to another cup successfully 80% of the time. Finally, DMD attains 90% success rate for hanging shirt on a clothing rack.