Accounting for Human Learning when Inferring Human Preferences

TL;DR

This paper explores how accounting for human learning in inverse reinforcement learning can improve the inference of human preferences, especially in unfamiliar environments, by modeling humans as learning agents rather than stationary demonstrators.

Contribution

It introduces a model that incorporates human learning into IRL, demonstrating that this approach can outperform traditional methods assuming stationarity in certain scenarios.

Findings

Learning-aware IRL can outperform stationary models in some cases.

Model misspecification can lead to poor inference results.

Accounting for human learning is crucial in unfamiliar environments.

Abstract

Inverse reinforcement learning (IRL) is a common technique for inferring human preferences from data. Standard IRL techniques tend to assume that the human demonstrator is stationary, that is that their policy $\pi$ doesn't change over time. In practice, humans interacting with a novel environment or performing well on a novel task will change their demonstrations as they learn more about the environment or task. We investigate the consequences of relaxing this assumption of stationarity, in particular by modelling the human as learning. Surprisingly, we find in some small examples that this can lead to better inference than if the human was stationary. That is, by observing a demonstrator who is themselves learning, a machine can infer more than by observing a demonstrator who is noisily rational. In addition, we find evidence that misspecification can lead to poor inference, suggesting that modelling human learning is important, especially when the human is facing an unfamiliar environment.