Fairer Machine Learning Through Multi-objective Evolutionary Learning

Qingquan ZHANG; Jialin LIU; Zeqi ZHANG; Junyi WEN; Bifei MAO; Xin YAO

doi:10.1007/978-3-030-86380-7_10

Fairer Machine Learning Through Multi-objective Evolutionary Learning

Qingquan ZHANG, Jialin LIU, Zeqi ZHANG, Junyi WEN, Bifei MAO, Xin YAO^*

^*Corresponding author for this work

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed) › Research › peer-review

11 Citations (Scopus)

Abstract

Dilemma between model accuracy and fairness in machine learning models has been shown theoretically and empirically. So far, dozens of fairness measures have been proposed, among which incompatibility and complementarity exist. However, no fairness measure has been universally accepted as the single fairest measure. No one has considered multiple fairness measures simultaneously. In this paper, we propose a multi-objective evolutionary learning framework for mitigating unfairness caused by considering a single measure only, in which a multi-objective evolutionary algorithm is used during training to balance accuracy and multiple fairness measures simultaneously. In our case study, besides the model accuracy, two fairness measures that are conflicting to each other are selected. Empirical results show that our proposed multi-objective evolutionary learning framework is able to find Pareto-front models efficiently and provide fairer machine learning models that consider multiple fairness measures.

Original language	English
Title of host publication	Artificial Neural Networks and Machine Learning – ICANN 2021 - 30th International Conference on Artificial Neural Networks, Proceedings
Editors	Igor FARKAŠ, Paolo MASULLI, Sebastian OTTE, Stefan WERMTER
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	111-123
Number of pages	13
ISBN (Electronic)	9783030863807
ISBN (Print)	9783030863791
DOIs	https://doi.org/10.1007/978-3-030-86380-7_10
Publication status	Published - 2021
Externally published	Yes
Event	30th International Conference on Artificial Neural Networks - Bratislava, Slovakia Duration: 14 Sept 2021 → 17 Sept 2021

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer
Volume	12894
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Name	Theoretical Computer Science and General Issues
Publisher	Springer
ISSN (Print)	2512-2010
ISSN (Electronic)	2512-2029

Conference

Conference	30th International Conference on Artificial Neural Networks
Abbreviated title	ICANN 2021
Country/Territory	Slovakia
City	Bratislava
Period	14/09/21 → 17/09/21

Bibliographical note

Publisher Copyright:
© 2021, Springer Nature Switzerland AG.

Funding

This work was supported by the Research Institute of Trustworthy Autonomous Systems, the Guangdong Provincial Key Laboratory (Grant No. 2020B121201001), the Program for Guangdong Introducing Innovative and Entrepreneurial Teams (Grant No. 2017ZT07X386), the Guangdong Basic and Applied Basic Research Foundation (Grant No. 2021A1515011830), the Shenzhen Science and Technology Program (Grant No. KQTD2016112514355531), the Shenzhen Fundamental Research Program (Grant Nos. JCYJ20180504165652917, JCYJ20190809121403553) and Huawei project on “Fundamental Theory and Key Technologies of Trustworthy Systems”.

Keywords

AI ethics
Discrimination in machine learning
Fairness in machine learning
Fairness measures
Multi-objective learning

Access to Document

10.1007/978-3-030-86380-7_10

Cite this

ZHANG, Q., LIU, J., ZHANG, Z., WEN, J., MAO, B., & YAO, X. (2021). Fairer Machine Learning Through Multi-objective Evolutionary Learning. In I. FARKAŠ, P. MASULLI, S. OTTE, & S. WERMTER (Eds.), Artificial Neural Networks and Machine Learning – ICANN 2021 - 30th International Conference on Artificial Neural Networks, Proceedings (pp. 111-123). (Lecture Notes in Computer Science; Vol. 12894), (Theoretical Computer Science and General Issues). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-86380-7_10

ZHANG, Qingquan ; LIU, Jialin ; ZHANG, Zeqi et al. / Fairer Machine Learning Through Multi-objective Evolutionary Learning. Artificial Neural Networks and Machine Learning – ICANN 2021 - 30th International Conference on Artificial Neural Networks, Proceedings. editor / Igor FARKAŠ ; Paolo MASULLI ; Sebastian OTTE ; Stefan WERMTER. Springer Science and Business Media Deutschland GmbH, 2021. pp. 111-123 (Lecture Notes in Computer Science). (Theoretical Computer Science and General Issues).

@inproceedings{7b00a4d96fdb4dbf8f84d321f8c97d64,

title = "Fairer Machine Learning Through Multi-objective Evolutionary Learning",

abstract = "Dilemma between model accuracy and fairness in machine learning models has been shown theoretically and empirically. So far, dozens of fairness measures have been proposed, among which incompatibility and complementarity exist. However, no fairness measure has been universally accepted as the single fairest measure. No one has considered multiple fairness measures simultaneously. In this paper, we propose a multi-objective evolutionary learning framework for mitigating unfairness caused by considering a single measure only, in which a multi-objective evolutionary algorithm is used during training to balance accuracy and multiple fairness measures simultaneously. In our case study, besides the model accuracy, two fairness measures that are conflicting to each other are selected. Empirical results show that our proposed multi-objective evolutionary learning framework is able to find Pareto-front models efficiently and provide fairer machine learning models that consider multiple fairness measures.",

keywords = "AI ethics, Discrimination in machine learning, Fairness in machine learning, Fairness measures, Multi-objective learning",

author = "Qingquan ZHANG and Jialin LIU and Zeqi ZHANG and Junyi WEN and Bifei MAO and Xin YAO",

note = "Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; 30th International Conference on Artificial Neural Networks, ICANN 2021 ; Conference date: 14-09-2021 Through 17-09-2021",

year = "2021",

doi = "10.1007/978-3-030-86380-7_10",

language = "English",

isbn = "9783030863791",

series = "Lecture Notes in Computer Science",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "111--123",

editor = "Igor FARKA{\v S} and MASULLI, {Paolo } and OTTE, {Sebastian } and Stefan WERMTER",

booktitle = "Artificial Neural Networks and Machine Learning – ICANN 2021 - 30th International Conference on Artificial Neural Networks, Proceedings",

address = "Germany",

}

ZHANG, Q, LIU, J, ZHANG, Z, WEN, J, MAO, B & YAO, X 2021, Fairer Machine Learning Through Multi-objective Evolutionary Learning. in I FARKAŠ, P MASULLI, S OTTE & S WERMTER (eds), Artificial Neural Networks and Machine Learning – ICANN 2021 - 30th International Conference on Artificial Neural Networks, Proceedings. Lecture Notes in Computer Science, vol. 12894, Theoretical Computer Science and General Issues, Springer Science and Business Media Deutschland GmbH, pp. 111-123, 30th International Conference on Artificial Neural Networks, Bratislava, Slovakia, 14/09/21. https://doi.org/10.1007/978-3-030-86380-7_10

Fairer Machine Learning Through Multi-objective Evolutionary Learning. / ZHANG, Qingquan; LIU, Jialin; ZHANG, Zeqi et al.
Artificial Neural Networks and Machine Learning – ICANN 2021 - 30th International Conference on Artificial Neural Networks, Proceedings. ed. / Igor FARKAŠ; Paolo MASULLI; Sebastian OTTE; Stefan WERMTER. Springer Science and Business Media Deutschland GmbH, 2021. p. 111-123 (Lecture Notes in Computer Science; Vol. 12894), (Theoretical Computer Science and General Issues).

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed) › Research › peer-review

TY - GEN

T1 - Fairer Machine Learning Through Multi-objective Evolutionary Learning

AU - ZHANG, Qingquan

AU - LIU, Jialin

AU - ZHANG, Zeqi

AU - WEN, Junyi

AU - MAO, Bifei

AU - YAO, Xin

PY - 2021

Y1 - 2021

N2 - Dilemma between model accuracy and fairness in machine learning models has been shown theoretically and empirically. So far, dozens of fairness measures have been proposed, among which incompatibility and complementarity exist. However, no fairness measure has been universally accepted as the single fairest measure. No one has considered multiple fairness measures simultaneously. In this paper, we propose a multi-objective evolutionary learning framework for mitigating unfairness caused by considering a single measure only, in which a multi-objective evolutionary algorithm is used during training to balance accuracy and multiple fairness measures simultaneously. In our case study, besides the model accuracy, two fairness measures that are conflicting to each other are selected. Empirical results show that our proposed multi-objective evolutionary learning framework is able to find Pareto-front models efficiently and provide fairer machine learning models that consider multiple fairness measures.

AB - Dilemma between model accuracy and fairness in machine learning models has been shown theoretically and empirically. So far, dozens of fairness measures have been proposed, among which incompatibility and complementarity exist. However, no fairness measure has been universally accepted as the single fairest measure. No one has considered multiple fairness measures simultaneously. In this paper, we propose a multi-objective evolutionary learning framework for mitigating unfairness caused by considering a single measure only, in which a multi-objective evolutionary algorithm is used during training to balance accuracy and multiple fairness measures simultaneously. In our case study, besides the model accuracy, two fairness measures that are conflicting to each other are selected. Empirical results show that our proposed multi-objective evolutionary learning framework is able to find Pareto-front models efficiently and provide fairer machine learning models that consider multiple fairness measures.

KW - AI ethics

KW - Discrimination in machine learning

KW - Fairness in machine learning

KW - Fairness measures

KW - Multi-objective learning

UR - http://www.scopus.com/inward/record.url?scp=85115673628&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-86380-7_10

DO - 10.1007/978-3-030-86380-7_10

M3 - Conference paper (refereed)

SN - 9783030863791

T3 - Lecture Notes in Computer Science

SP - 111

EP - 123

BT - Artificial Neural Networks and Machine Learning – ICANN 2021 - 30th International Conference on Artificial Neural Networks, Proceedings

A2 - FARKAŠ, Igor

A2 - MASULLI, Paolo

A2 - OTTE, Sebastian

A2 - WERMTER, Stefan

PB - Springer Science and Business Media Deutschland GmbH

T2 - 30th International Conference on Artificial Neural Networks

Y2 - 14 September 2021 through 17 September 2021

ER -

ZHANG Q, LIU J, ZHANG Z, WEN J, MAO B, YAO X. Fairer Machine Learning Through Multi-objective Evolutionary Learning. In FARKAŠ I, MASULLI P, OTTE S, WERMTER S, editors, Artificial Neural Networks and Machine Learning – ICANN 2021 - 30th International Conference on Artificial Neural Networks, Proceedings. Springer Science and Business Media Deutschland GmbH. 2021. p. 111-123. (Lecture Notes in Computer Science). (Theoretical Computer Science and General Issues). Epub 2021 Sept 7. doi: 10.1007/978-3-030-86380-7_10

Fairer Machine Learning Through Multi-objective Evolutionary Learning

Abstract

Publication series

Conference

Bibliographical note

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this