Decision-Focused Learning: Foundations, State of the Art, Benchmark and Future Opportunities

Decision-focused learning (DFL) is an emerging paradigm that integrates machine learning (ML) and constrained optimization to enhance decision quality by training ML models in an end-to-end system. This approach shows significant potential to revolutionize combinatorial decision-making in real-world...

Full description

Saved in:
Bibliographic Details
Published inThe Journal of artificial intelligence research Vol. 80; pp. 1623 - 1701
Main Authors Mandi, Jayanta, Kotary, James, Berden, Senne, Mulamba, Maxime, Bucarey, Victor, Guns, Tias, Fioretto, Ferdinando
Format Journal Article
LanguageEnglish
Published 31.08.2024
Online AccessGet full text

Cover

More Information
Summary:Decision-focused learning (DFL) is an emerging paradigm that integrates machine learning (ML) and constrained optimization to enhance decision quality by training ML models in an end-to-end system. This approach shows significant potential to revolutionize combinatorial decision-making in real-world applications that operate under uncertainty, where estimating unknown parameters within decision models is a major challenge. This paper presents a comprehensive review of DFL, providing an in-depth analysis of both gradient-based and gradient-free techniques used to combine ML and constrained optimization. It evaluates the strengths and limitations of these techniques and includes an extensive empirical evaluation of eleven methods across seven problems. The survey also offers insights into recent advancements and future research directions in DFL.
ISSN:1076-9757
1076-9757
DOI:10.1613/jair.1.15320