Regret-based Reward Elicitation for Markov Decision Processes

TL;DR

This paper introduces a regret-based reward elicitation method for Markov Decision Processes that reduces the need for precise reward specification by efficiently querying preferences, resulting in near-optimal policies.

Contribution

It proposes a novel approach using minimax regret and bound queries to efficiently elicit reward information, improving policy quality with less detailed reward input.

Findings

Regret-based elicitation effectively produces near-optimal policies.

Using bound queries reduces the amount of reward information needed.

Empirical results show improved efficiency over traditional methods.

Abstract

The specification of aMarkov decision process (MDP) can be difficult. Reward function specification is especially problematic; in practice, it is often cognitively complex and time-consuming for users to precisely specify rewards. This work casts the problem of specifying rewards as one of preference elicitation and aims to minimize the degree of precision with which a reward function must be specified while still allowing optimal or near-optimal policies to be produced. We first discuss how robust policies can be computed for MDPs given only partial reward information using the minimax regret criterion. We then demonstrate how regret can be reduced by efficiently eliciting reward information using bound queries, using regret-reduction as a means for choosing suitable queries. Empirical results demonstrate that regret-based reward elicitation offers an effective way to produce near-optimal policies without resorting to the precise specification of the entire reward function.