Utility-Directed Conformal Prediction: A Decision-Aware Framework for Actionable Uncertainty Quantification

TL;DR

This paper introduces a conformal prediction framework that incorporates downstream decision costs and utility functions, improving uncertainty quantification for high-stakes decision-making in fields like healthcare.

Contribution

It develops a decision-aware conformal prediction method that accounts for downstream costs, maintaining coverage guarantees while reducing decision costs compared to standard methods.

Findings

Achieves significantly lower costs than standard conformal methods.

Maintains statistical coverage guarantees.

Effectively incorporates hierarchical structure in healthcare diagnosis.

Abstract

Interest has been growing in decision-focused machine learning methods which train models to account for how their predictions are used in downstream optimization problems. Doing so can often improve performance on subsequent decision problems. However, current methods for uncertainty quantification do not incorporate any information about downstream decisions. We develop a methodology based on conformal prediction to identify prediction sets that account for a downstream cost function, making them more appropriate to inform high-stakes decision-making. Our approach harnesses the strengths of conformal methods -- modularity, model-agnosticism, and statistical coverage guarantees -- while incorporating downstream decisions and user-specified utility functions. We prove that our methods retain standard coverage guarantees. Empirical evaluation across a range of datasets and utility metrics demonstrates that our methods achieve significantly lower costs than standard conformal methods. We present a real-world use case in healthcare diagnosis, where our method effectively incorporates the hierarchical structure of dermatological diseases. The method successfully generates sets with coherent diagnostic meaning, potentially aiding triage for dermatology diagnosis and illustrating how our method can ground high-stakes decision-making employing domain knowledge.