Counterfactual Strategies for Markov Decision Processes

TL;DR

This paper introduces counterfactual strategies for Markov Decision Processes, enabling minimal modifications to decision strategies to reduce undesired outcomes in sequential decision-making tasks.

Contribution

It extends counterfactual reasoning to MDPs by encoding minimal strategy changes as solutions to non-linear optimization problems.

Findings

Successfully reduces undesired outcome probabilities in real-world datasets

Demonstrates practical viability in complex sequential decision tasks

Provides a method for synthesizing diverse counterfactual strategies

Abstract

Counterfactuals are widely used in AI to explain how minimal changes to a model's input can lead to a different output. However, established methods for computing counterfactuals typically focus on one-step decision-making, and are not directly applicable to sequential decision-making tasks. This paper fills this gap by introducing counterfactual strategies for Markov Decision Processes (MDPs). During MDP execution, a strategy decides which of the enabled actions (with known probabilistic effects) to execute next. Given an initial strategy that reaches an undesired outcome with a probability above some limit, we identify minimal changes to the initial strategy to reduce that probability below the limit. We encode such counterfactual strategies as solutions to non-linear optimization problems, and further extend our encoding to synthesize diverse counterfactual strategies. We evaluate our approach on four real-world datasets and demonstrate its practical viability in sophisticated sequential decision-making tasks.