Decision-Focused Learning: Through the Lens of Learning to Rank

Jayanta Mandi, Victor Bucarey Lopez, Maxime Mulamba, Tias Guns

Research output: Chapter in Book/Report/Conference proceedingConference paper

31 Citations (Scopus)

Abstract

In the last years, decision-focused learning, also known as predict-and-optimize, has received increasing attention. In this setting, the predictions of a machine learning model are used as estimated cost coefficients in the objective function of a discrete combinatorial optimization problem for decision making. Decision-focused learning proposes to train the ML models, often neural network models, by directly optimizing the quality of decisions made by the optimization solvers. Based on a recent work that proposed a noise contrastive estimation loss over a subset of the solution space, we observe that decision-focused learning can more generally be seen as a learning-to-rank problem, where the goal is to learn an objective function that ranks the feasible points correctly. This observation is independent of the optimization method used and of the form of the objective function. We develop pointwise, pairwise and listwise ranking loss functions, which can be differentiated in closed form given a subset of solutions. We empirically investigate the quality of our generic methods compared to existing decision-focused learning approaches with competitive results. Furthermore, controlling the subset of solutions allows controlling the runtime considerably, with limited effect on regret.

Original languageEnglish
Title of host publicationDecision-Focused Learning: Through the Lens of Learning to Rank
Place of PublicationBaltimore, Maryland, USA
PublisherProceedings of Machine Learning Research
Pages14935-14947
Number of pages13
Volume162
Publication statusPublished - 17 Jul 2022
EventThirty-eighth International Conference on Machine Learning - Baltimore, United States
Duration: 17 Jul 202223 Jul 2022
Conference number: 2022
https://icml.cc/virtual/2022/index.html

Publication series

NameProceedings of Machine Learning Research

Conference

ConferenceThirty-eighth International Conference on Machine Learning
Abbreviated titleICML
Country/TerritoryUnited States
CityBaltimore
Period17/07/2223/07/22
Internet address

Bibliographical note

Funding Information:
This research was partially funded by the FWO Flanders project Data-driven logistics (FWO-S007318N), the ANID-Fondecyt Iniciacion grant no 11220864, the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant No. 101002802, CHAT-Opt) and the Flemish Government under the “Onderzoeksprogramma Artificiele Intelligentie (AI) Vlaanderen” programme. We are also thankful to the reviewers for their insightful comments.

Funding Information:
This research was partially funded by the FWO Flanders project Data-driven logistics (FWO-S007318N), the ANIDFondecyt Iniciacion grant no 11220864, the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (Grant No. 101002802, CHAT-Opt) and the Flemish Government under the “Onderzoeksprogramma Artificiele Intelligentie (AI) Vlaanderen” programme. We are also thankful to the reviewers for their insightful comments.

Publisher Copyright:
Copyright © 2022 by the author(s)

Copyright:
Copyright 2023 Elsevier B.V., All rights reserved.

Fingerprint

Dive into the research topics of 'Decision-Focused Learning: Through the Lens of Learning to Rank'. Together they form a unique fingerprint.

Cite this