Learning from Guided Play: Improving Exploration for Adversarial Imitation Learning with Simple Auxiliary Tasks

TL;DR

This paper introduces Learning from Guided Play (LfGP), a framework that enhances exploration in adversarial imitation learning by incorporating auxiliary tasks and expert demonstrations, leading to better performance in robotic manipulation tasks.

Contribution

The paper proposes LfGP, a novel approach that leverages auxiliary tasks and expert data to improve exploration and sample efficiency in adversarial imitation learning.

Findings

LfGP outperforms AIL and behavior cloning in manipulation tasks.

LfGP is more sample-efficient than baseline methods.

Analysis reveals better exploration reduces local maxima issues.

Abstract

Adversarial imitation learning (AIL) has become a popular alternative to supervised imitation learning that reduces the distribution shift suffered by the latter. However, AIL requires effective exploration during an online reinforcement learning phase. In this work, we show that the standard, naive approach to exploration can manifest as a suboptimal local maximum if a policy learned with AIL sufficiently matches the expert distribution without fully learning the desired task. This can be particularly catastrophic for manipulation tasks, where the difference between an expert and a non-expert state-action pair is often subtle. We present Learning from Guided Play (LfGP), a framework in which we leverage expert demonstrations of multiple exploratory, auxiliary tasks in addition to a main task. The addition of these auxiliary tasks forces the agent to explore states and actions that standard AIL may learn to ignore. Additionally, this particular formulation allows for the reusability of expert data between main tasks. Our experimental results in a challenging multitask robotic manipulation domain indicate that LfGP significantly outperforms both AIL and behaviour cloning, while also being more expert sample efficient than these baselines. To explain this performance gap, we provide further analysis of a toy problem that highlights the coupling between a local maximum and poor exploration, and also visualize the differences between the learned models from AIL and LfGP.