Robust Learning from Observation with Model Misspecification

TL;DR

This paper introduces a robust imitation learning algorithm designed to transfer policies from simulation to real environments despite model misspecification, using only state-only demonstrations from the real environment.

Contribution

It proposes a novel robust IL method that leverages insights from robust RL and adversarial approaches to improve zero-shot transfer from simulation to real-world settings.

Findings

Outperforms state-only IL methods in zero-shot transfer

Demonstrates robustness under various testing conditions

Achieves better real-world performance in continuous-control benchmarks

Abstract

Imitation learning (IL) is a popular paradigm for training policies in robotic systems when specifying the reward function is difficult. However, despite the success of IL algorithms, they impose the somewhat unrealistic requirement that the expert demonstrations must come from the same domain in which a new imitator policy is to be learned. We consider a practical setting, where (i) state-only expert demonstrations from the real (deployment) environment are given to the learner, (ii) the imitation learner policy is trained in a simulation (training) environment whose transition dynamics is slightly different from the real environment, and (iii) the learner does not have any access to the real environment during the training phase beyond the batch of demonstrations given. Most of the current IL methods, such as generative adversarial imitation learning and its state-only variants, fail to imitate the optimal expert behavior under the above setting. By leveraging insights from the Robust reinforcement learning (RL) literature and building on recent adversarial imitation approaches, we propose a robust IL algorithm to learn policies that can effectively transfer to the real environment without fine-tuning. Furthermore, we empirically demonstrate on continuous-control benchmarks that our method outperforms the state-of-the-art state-only IL method in terms of the zero-shot transfer performance in the real environment and robust performance under different testing conditions.