Learning Robust Decision Policies from Observational Data

TL;DR

This paper introduces a method for learning decision policies from observational data that minimizes tail risks and provides valid cost bounds, even with limited or overlapping data, using conformal prediction techniques.

Contribution

It develops a novel approach for robust policy learning that ensures risk reduction and statistical validity under finite samples and challenging data conditions.

Findings

Method effectively reduces tail risks in decision policies.

Provides statistically valid bounds on decision costs.

Demonstrates robustness on real and synthetic datasets.

Abstract

We address the problem of learning a decision policy from observational data of past decisions in contexts with features and associated outcomes. The past policy maybe unknown and in safety-critical applications, such as medical decision support, it is of interest to learn robust policies that reduce the risk of outcomes with high costs. In this paper, we develop a method for learning policies that reduce tails of the cost distribution at a specified level and, moreover, provide a statistically valid bound on the cost of each decision. These properties are valid under finite samples -- even in scenarios with uneven or no overlap between features for different decisions in the observed data -- by building on recent results in conformal prediction. The performance and statistical properties of the proposed method are illustrated using both real and synthetic data.