Inverse Reinforcement Learning in a Continuous State Space with Formal Guarantees

TL;DR

This paper introduces a new inverse reinforcement learning algorithm for continuous state spaces with unknown dynamics, providing theoretical guarantees and validating its effectiveness through synthetic experiments.

Contribution

The work presents a novel IRL algorithm for continuous states using orthonormal basis functions, with proven correctness and complexity guarantees.

Findings

Algorithm achieves formal correctness guarantees.

Sample and time complexity are rigorously analyzed.

Synthetic experiments support theoretical results.

Abstract

Inverse Reinforcement Learning (IRL) is the problem of finding a reward function which describes observed/known expert behavior. The IRL setting is remarkably useful for automated control, in situations where the reward function is difficult to specify manually or as a means to extract agent preference. In this work, we provide a new IRL algorithm for the continuous state space setting with unknown transition dynamics by modeling the system using a basis of orthonormal functions. Moreover, we provide a proof of correctness and formal guarantees on the sample and time complexity of our algorithm. Finally, we present synthetic experiments to corroborate our theoretical guarantees.