Learning to Defer: A Survey

Fuente: Zenodo
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Strong, Joshua, Sun, Emma, Rogers, Harry, Higham, Helen, Noble, Alison
Format: Recurso digital
Langue:anglais
Publié: Zenodo 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866902233291948032
author Strong, Joshua
Sun, Emma
Rogers, Harry
Higham, Helen
Noble, Alison
author_facet Strong, Joshua
Sun, Emma
Rogers, Harry
Higham, Helen
Noble, Alison
contents <p>Learning to defer (L2D) enables AI systems to choose between autonomous prediction and deferral to experts. This survey consolidates the fast-growing literature through a four-branch taxonomy: methodological frameworks; optimization and theory; task generalizations; and real-world adaptations. We outline contrasts between score-based and predictor–rejector formulations; one-stage, two-stage, and post-hoc training; unify surrogate losses with theoretical guarantees; and synthesize extensions to regression, multi-task prediction, top-$k$ committees, sequential settings, and causal pipelines. Practical considerations include limited annotations, dynamic expert pools, workload/budget control, fairness, interpretability, robustness, and uncertainty handling, concluding with open challenges for reliable human–AI decision systems.<br><br><strong>Note:</strong> This is the preprint version of a manuscript currently under review. It has not undergone peer review.</p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_17843044
institution Zenodo
language eng
publishDate 2025
publisher Zenodo
record_format zenodo
spellingShingle Learning to Defer: A Survey
Strong, Joshua
Sun, Emma
Rogers, Harry
Higham, Helen
Noble, Alison
Machine Learning
Artificial Intelligence
Human-AI Collaboration
Learning to Defer
Survey
<p>Learning to defer (L2D) enables AI systems to choose between autonomous prediction and deferral to experts. This survey consolidates the fast-growing literature through a four-branch taxonomy: methodological frameworks; optimization and theory; task generalizations; and real-world adaptations. We outline contrasts between score-based and predictor–rejector formulations; one-stage, two-stage, and post-hoc training; unify surrogate losses with theoretical guarantees; and synthesize extensions to regression, multi-task prediction, top-$k$ committees, sequential settings, and causal pipelines. Practical considerations include limited annotations, dynamic expert pools, workload/budget control, fairness, interpretability, robustness, and uncertainty handling, concluding with open challenges for reliable human–AI decision systems.<br><br><strong>Note:</strong> This is the preprint version of a manuscript currently under review. It has not undergone peer review.</p>
title Learning to Defer: A Survey
topic Machine Learning
Artificial Intelligence
Human-AI Collaboration
Learning to Defer
Survey
url https://doi.org/10.5281/zenodo.17843044