Toward Information Theoretic Active Inverse Reinforcement Learning

TL;DR

This paper develops an information-theoretic framework for active inverse reinforcement learning, enabling efficient selection of demonstration trajectories to better infer human preferences in autonomous systems.

Contribution

It introduces an information-theoretic acquisition function for active IRL, analyzes scenarios with longer trajectory queries, and provides an efficient approximation scheme with gridworld experiments.

Findings

Effective trajectory selection reduces human demonstration effort.

The proposed method outperforms baseline approaches in gridworld tests.

Framework lays groundwork for extending to complex environments.

Abstract

As AI systems become increasingly autonomous, aligning their decision-making to human preferences is essential. In domains like autonomous driving or robotics, it is impossible to write down the reward function representing these preferences by hand. Inverse reinforcement learning (IRL) offers a promising approach to infer the unknown reward from demonstrations. However, obtaining human demonstrations can be costly. Active IRL addresses this challenge by strategically selecting the most informative scenarios for human demonstration, reducing the amount of required human effort. Where most prior work allowed querying the human for an action at one state at a time, we motivate and analyse scenarios where we collect longer trajectories. We provide an information-theoretic acquisition function, propose an efficient approximation scheme, and illustrate its performance through a set of gridworld experiments as groundwork for future work expanding to more general settings.