Maximum Entropy Semi-Supervised Inverse Reinforcement Learning

TL;DR

This paper introduces MESSI, a semi-supervised extension of MaxEnt-IRL, which leverages both expert and unsupervised trajectories to improve inverse reinforcement learning performance.

Contribution

The paper proposes MESSI, a novel algorithm that integrates semi-supervised learning principles into MaxEnt-IRL using pairwise trajectory penalties.

Findings

MESSI outperforms MaxEnt-IRL when using unsupervised data.

Empirical results show improved learning in highway driving and grid-world tasks.

Unsupervised trajectories enhance IRL performance with MESSI.

Abstract

A popular approach to apprenticeship learning (AL) is to formulate it as an inverse reinforcement learning (IRL) problem. The MaxEnt-IRL algorithm successfully integrates the maximum entropy principle into IRL and unlike its predecessors, it resolves the ambiguity arising from the fact that a possibly large number of policies could match the expert's behavior. In this paper, we study an AL setting in which in addition to the expert's trajectories, a number of unsupervised trajectories is available. We introduce MESSI, a novel algorithm that combines MaxEnt-IRL with principles coming from semi-supervised learning. In particular, MESSI integrates the unsupervised data into the MaxEnt-IRL framework using a pairwise penalty on trajectories. Empirical results in a highway driving and grid-world problems indicate that MESSI is able to take advantage of the unsupervised trajectories and improve the performance of MaxEnt-IRL.