Occam's razor is insufficient to infer the preferences of irrational agents

TL;DR

This paper demonstrates the fundamental limitations of inferring human preferences from observed behavior due to irrationality and ambiguity, showing that simple assumptions alone are insufficient for accurate inference.

Contribution

It introduces a theoretical framework highlighting the impossibility of uniquely decomposing policies into rewards and planning algorithms without normative assumptions.

Findings

No Free Lunch theorem applies to IRL decomposition

Occam's razor cannot reliably identify true reward functions

Additional normative assumptions are necessary for accurate inference

Abstract

Inverse reinforcement learning (IRL) attempts to infer human rewards or preferences from observed behavior. Since human planning systematically deviates from rationality, several approaches have been tried to account for specific human shortcomings. However, the general problem of inferring the reward function of an agent of unknown rationality has received little attention. Unlike the well-known ambiguity problems in IRL, this one is practically relevant but cannot be resolved by observing the agent's policy in enough environments. This paper shows (1) that a No Free Lunch result implies it is impossible to uniquely decompose a policy into a planning algorithm and reward function, and (2) that even with a reasonable simplicity prior/Occam's razor on the set of decompositions, we cannot distinguish between the true decomposition and others that lead to high regret. To address this, we need simple `normative' assumptions, which cannot be deduced exclusively from observations.