Adversarial Counterfactual Environment Model Learning

TL;DR

This paper introduces GALILEO, an adversarial learning method for environment models that improves generalization to counterfactual data, enabling more sample-efficient decision-making in various domains.

Contribution

It proposes an adversarial counterfactual-query risk minimization framework and a tractable solution, GALILEO, for better generalization of environment models to unseen, counterfactual scenarios.

Findings

GALILEO accurately predicts counterfactual data.

It significantly improves policy performance in real-world tests.

The method is effective in synthetic and real-world applications.

Abstract

A good model for action-effect prediction, named environment model, is important to achieve sample-efficient decision-making policy learning in many domains like robot control, recommender systems, and patients' treatment selection. We can take unlimited trials with such a model to identify the appropriate actions so that the costs of queries in the real world can be saved. It requires the model to handle unseen data correctly, also called counterfactual data. However, standard data fitting techniques do not automatically achieve such generalization ability and commonly result in unreliable models. In this work, we introduce counterfactual-query risk minimization (CQRM) in model learning for generalizing to a counterfactual dataset queried by a specific target policy. Since the target policies can be various and unknown in policy learning, we propose an adversarial CQRM objective in which the model learns on counterfactual data queried by adversarial policies, and finally derive a tractable solution GALILEO. We also discover that adversarial CQRM is closely related to the adversarial model learning, explaining the effectiveness of the latter. We apply GALILEO in synthetic tasks and a real-world application. The results show that GALILEO makes accurate predictions on counterfactual data and thus significantly improves policies in real-world testing.