Misspecification in Inverse Reinforcement Learning
This addresses the problem of unreliable reward inference in IRL for applications like human preference learning, though it is incremental as it builds on existing models without proposing a new paradigm.
The paper analyzes the robustness of different Inverse Reinforcement Learning (IRL) models to misspecification, quantifying how much demonstrator policies can deviate from standard models before leading to faulty reward inferences, and introduces a framework with formal tools for assessing misspecification in IRL.
The aim of Inverse Reinforcement Learning (IRL) is to infer a reward function $R$ from a policy $π$. To do this, we need a model of how $π$ relates to $R$. In the current literature, the most common models are optimality, Boltzmann rationality, and causal entropy maximisation. One of the primary motivations behind IRL is to infer human preferences from human behaviour. However, the true relationship between human preferences and human behaviour is much more complex than any of the models currently used in IRL. This means that they are misspecified, which raises the worry that they might lead to unsound inferences if applied to real-world data. In this paper, we provide a mathematical analysis of how robust different IRL models are to misspecification, and answer precisely how the demonstrator policy may differ from each of the standard models before that model leads to faulty inferences about the reward function $R$. We also introduce a framework for reasoning about misspecification in IRL, together with formal tools that can be used to easily derive the misspecification robustness of new IRL models.