Set-Valued Policy Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Fuentes-Vicente, Laura, Even, Mathieu, Dormion, Gaëlle, Chambaz, Antoine, Shalit, Uri, Josse, Julie
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913146394902528
author Fuentes-Vicente, Laura
Even, Mathieu
Dormion, Gaëlle
Chambaz, Antoine
Shalit, Uri
Josse, Julie
author_facet Fuentes-Vicente, Laura
Even, Mathieu
Dormion, Gaëlle
Chambaz, Antoine
Shalit, Uri
Josse, Julie
contents Conventional treatment policies map patient covariates to a single recommended intervention in order to maximize expected clinical outcomes. Although a rich body of causal inference methods has been developed to estimate such policies, point-valued recommendations can be highly sensitive to estimation uncertainty, model specification, and finite-sample variability, while typically providing little guidance about how confident one should be in the recommended action. In this work, we propose a set-valued policy learning paradigm for the multiple-treatment setting, in which policies output a set of plausible treatments rather than a single recommendation. This formulation enables intrinsic uncertainty quantification, with the size of the predicted set reflecting the degree of decision ambiguity. We extend the learning-to-defer framework to multiple treatments via a novel \textit{greatest Lower Bound} method, and introduce \textit{conformal policy learning}, which bridges the gap between unobserved ground-truth optimal treatments and estimated optimal treatment rules. Drawing on insights from the noisy-label literature, we develop a randomness-injection approach that guarantees marginal coverage without requiring assumptions on underlying black-box optimal treatment rules. Through experiments on synthetic data and a real-world application to In-Vitro Fertilization (IVF), we demonstrate that our methods produce robust and actionable policies that naturally incorporate clinical considerations while effectively balancing performance and reliability.
format Preprint
id arxiv_https___arxiv_org_abs_2605_19830
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Set-Valued Policy Learning
Fuentes-Vicente, Laura
Even, Mathieu
Dormion, Gaëlle
Chambaz, Antoine
Shalit, Uri
Josse, Julie
Machine Learning
Statistics Theory
Conventional treatment policies map patient covariates to a single recommended intervention in order to maximize expected clinical outcomes. Although a rich body of causal inference methods has been developed to estimate such policies, point-valued recommendations can be highly sensitive to estimation uncertainty, model specification, and finite-sample variability, while typically providing little guidance about how confident one should be in the recommended action. In this work, we propose a set-valued policy learning paradigm for the multiple-treatment setting, in which policies output a set of plausible treatments rather than a single recommendation. This formulation enables intrinsic uncertainty quantification, with the size of the predicted set reflecting the degree of decision ambiguity. We extend the learning-to-defer framework to multiple treatments via a novel \textit{greatest Lower Bound} method, and introduce \textit{conformal policy learning}, which bridges the gap between unobserved ground-truth optimal treatments and estimated optimal treatment rules. Drawing on insights from the noisy-label literature, we develop a randomness-injection approach that guarantees marginal coverage without requiring assumptions on underlying black-box optimal treatment rules. Through experiments on synthetic data and a real-world application to In-Vitro Fertilization (IVF), we demonstrate that our methods produce robust and actionable policies that naturally incorporate clinical considerations while effectively balancing performance and reliability.
title Set-Valued Policy Learning
topic Machine Learning
Statistics Theory
url https://arxiv.org/abs/2605.19830