LibMOON: A Gradient-based MultiObjective OptimizatioN Library in PyTorch

Xiaoyuan Zhang; Liang Zhao; Yingying Yu; Xi Lin; Yifan Chen; Han Zhao; Qingfu Zhang

LibMOON: A Gradient-based MultiObjective OptimizatioN Library in PyTorch

Xiaoyuan Zhang
, Liang Zhao
, Yingying Yu
, Xi Lin
, Yifan Chen
, Han Zhao
, Qingfu Zhang^*

^*Corresponding author for this work

Department of Computer Science

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

5 Citations (Scopus)

Abstract

Multiobjective optimization problems (MOPs) are prevalent in machine learning, with applications in multi-task learning, fairness, robustness, and more. Unlike single-objective optimization, which aggregates objectives into a scalar through weighted sums, MOPs focus on generating specific or diverse Pareto solutions and learning the entire Pareto set directly. Existing MOP benchmarks primarily focus on evolutionary algorithms, which are zeroth-order or meta-heuristic methods that fail to leverage higher-order objective information and cannot scale to large models. To address these challenges, we introduce LibMOON, the first multiobjective optimization library supporting state-of-the-art gradient-based methods, offering a fair and comprehensive benchmark, and open-sourced for the community. © 2024 Neural information processing systems foundation. All rights reserved.

Original language	English
Title of host publication	NeurIPS Proceedings
Subtitle of host publication	Advances in Neural Information Processing Systems 37 (NeurIPS 2024)
Editors	A. Globerson, L. Mackey , D. Belgrave , A. Fan, U. Paquet , J. Tomczak , C. Zhang
Publisher	Neural Information Processing Systems (NeurIPS)
ISBN (Print)	9798331314385
Publication status	Published - Dec 2024
Event	38th Annual Conference on Neural Information Processing Systems (NeurIPS 2024) - Vancouver Convention Center, Vancouver, Canada Duration: 10 Dec 2024 → 15 Dec 2024 https://neurips.cc/ https://proceedings.neurips.cc/

Conference

Conference	38th Annual Conference on Neural Information Processing Systems (NeurIPS 2024)
Abbreviated title	NeurIPS 2024
Place	Canada
City	Vancouver
Period	10/12/24 → 15/12/24
Internet address	https://neurips.cc/ https://proceedings.neurips.cc/

Bibliographical note

Full text of this publication does not contain sufficient affiliation information. With consent from the author(s) concerned, the Research Unit(s) information for this record is based on the existing academic department affiliation of the author(s).

Funding

The work described in this paper was supported by the Research Grants Council of the Hong Kong Special Administrative Region, China [GRF Project No. CityU 11215622].

RGC Funding Information

RGC-funded

Access Output

https://proceedings.neurips.cc/paper_files/paper/2024/hash/03ce71224741e6596059b8680a62978b-Abstract-Datasets_and_Benchmarks_Track.html

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

Zhang, X., Zhao, L., Yu, Y., Lin, X., Chen, Y., Zhao, H., & Zhang, Q. (2024). LibMOON: A Gradient-based MultiObjective OptimizatioN Library in PyTorch. In A. Globerson, L. Mackey , D. Belgrave , A. Fan, U. Paquet , J. Tomczak , & C. Zhang (Eds.), NeurIPS Proceedings: Advances in Neural Information Processing Systems 37 (NeurIPS 2024) Neural Information Processing Systems (NeurIPS). https://proceedings.neurips.cc/paper_files/paper/2024/hash/03ce71224741e6596059b8680a62978b-Abstract-Datasets_and_Benchmarks_Track.html

@inproceedings{d4f55fb15bd445d3990e6bcc59ddd1fe,

title = "LibMOON: A Gradient-based MultiObjective OptimizatioN Library in PyTorch",

abstract = "Multiobjective optimization problems (MOPs) are prevalent in machine learning, with applications in multi-task learning, fairness, robustness, and more. Unlike single-objective optimization, which aggregates objectives into a scalar through weighted sums, MOPs focus on generating specific or diverse Pareto solutions and learning the entire Pareto set directly. Existing MOP benchmarks primarily focus on evolutionary algorithms, which are zeroth-order or meta-heuristic methods that fail to leverage higher-order objective information and cannot scale to large models. To address these challenges, we introduce LibMOON, the first multiobjective optimization library supporting state-of-the-art gradient-based methods, offering a fair and comprehensive benchmark, and open-sourced for the community. {\textcopyright} 2024 Neural information processing systems foundation. All rights reserved.",

author = "Xiaoyuan Zhang and Liang Zhao and Yingying Yu and Xi Lin and Yifan Chen and Han Zhao and Qingfu Zhang",

note = "Full text of this publication does not contain sufficient affiliation information. With consent from the author(s) concerned, the Research Unit(s) information for this record is based on the existing academic department affiliation of the author(s).; 38th Annual Conference on Neural Information Processing Systems (NeurIPS 2024), NeurIPS 2024 ; Conference date: 10-12-2024 Through 15-12-2024",

year = "2024",

month = dec,

language = "English",

isbn = "9798331314385",

editor = "A. Globerson and \{Mackey \}, L. and \{Belgrave \}, D. and A. Fan and \{Paquet \}, U. and \{Tomczak \}, J. and C. Zhang",

booktitle = "NeurIPS Proceedings",

publisher = "Neural Information Processing Systems (NeurIPS)",

address = "United States",

url = "https://neurips.cc/, https://proceedings.neurips.cc/",

}

Zhang, X , Zhao, L , Yu, Y , Lin, X, Chen, Y, Zhao, H & Zhang, Q 2024, LibMOON: A Gradient-based MultiObjective OptimizatioN Library in PyTorch. in A Globerson, L Mackey , D Belgrave , A Fan, U Paquet , J Tomczak & C Zhang (eds), NeurIPS Proceedings: Advances in Neural Information Processing Systems 37 (NeurIPS 2024). Neural Information Processing Systems (NeurIPS), 38th Annual Conference on Neural Information Processing Systems (NeurIPS 2024), Vancouver, Canada, 10/12/24. <https://proceedings.neurips.cc/paper_files/paper/2024/hash/03ce71224741e6596059b8680a62978b-Abstract-Datasets_and_Benchmarks_Track.html>

LibMOON: A Gradient-based MultiObjective OptimizatioN Library in PyTorch. / Zhang, Xiaoyuan ; Zhao, Liang ; Yu, Yingying et al.
NeurIPS Proceedings: Advances in Neural Information Processing Systems 37 (NeurIPS 2024). ed. / A. Globerson; L. Mackey ; D. Belgrave ; A. Fan; U. Paquet ; J. Tomczak ; C. Zhang. Neural Information Processing Systems (NeurIPS), 2024.

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - LibMOON

T2 - 38th Annual Conference on Neural Information Processing Systems (NeurIPS 2024)

AU - Zhang, Xiaoyuan

AU - Zhao, Liang

AU - Yu, Yingying

AU - Lin, Xi

AU - Chen, Yifan

AU - Zhao, Han

AU - Zhang, Qingfu

N1 - Full text of this publication does not contain sufficient affiliation information. With consent from the author(s) concerned, the Research Unit(s) information for this record is based on the existing academic department affiliation of the author(s).

PY - 2024/12

Y1 - 2024/12

N2 - Multiobjective optimization problems (MOPs) are prevalent in machine learning, with applications in multi-task learning, fairness, robustness, and more. Unlike single-objective optimization, which aggregates objectives into a scalar through weighted sums, MOPs focus on generating specific or diverse Pareto solutions and learning the entire Pareto set directly. Existing MOP benchmarks primarily focus on evolutionary algorithms, which are zeroth-order or meta-heuristic methods that fail to leverage higher-order objective information and cannot scale to large models. To address these challenges, we introduce LibMOON, the first multiobjective optimization library supporting state-of-the-art gradient-based methods, offering a fair and comprehensive benchmark, and open-sourced for the community. © 2024 Neural information processing systems foundation. All rights reserved.

AB - Multiobjective optimization problems (MOPs) are prevalent in machine learning, with applications in multi-task learning, fairness, robustness, and more. Unlike single-objective optimization, which aggregates objectives into a scalar through weighted sums, MOPs focus on generating specific or diverse Pareto solutions and learning the entire Pareto set directly. Existing MOP benchmarks primarily focus on evolutionary algorithms, which are zeroth-order or meta-heuristic methods that fail to leverage higher-order objective information and cannot scale to large models. To address these challenges, we introduce LibMOON, the first multiobjective optimization library supporting state-of-the-art gradient-based methods, offering a fair and comprehensive benchmark, and open-sourced for the community. © 2024 Neural information processing systems foundation. All rights reserved.

UR - https://www.scopus.com/pages/publications/105000515381

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-105000515381&origin=recordpage

M3 - RGC 32 - Refereed conference paper (with host publication)

SN - 9798331314385

BT - NeurIPS Proceedings

A2 - Globerson, A.

A2 - Mackey , L.

A2 - Belgrave , D.

A2 - Fan, A.

A2 - Paquet , U.

A2 - Tomczak , J.

A2 - Zhang, C.

PB - Neural Information Processing Systems (NeurIPS)

Y2 - 10 December 2024 through 15 December 2024

ER -

LibMOON: A Gradient-based MultiObjective OptimizatioN Library in PyTorch

Abstract

Conference

Bibliographical note

Funding

RGC Funding Information

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

GRF: Few for Many: A Non-Pareto Approach for Many Objective Optimization

Cite this

LibMOON: A Gradient-based MultiObjective OptimizatioN Library in PyTorch

Abstract

Conference

Bibliographical note

Funding

RGC Funding Information

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Projects

GRF: Few for Many: A Non-Pareto Approach for Many Objective Optimization

Cite this