The Shapley Value in Machine Learning

Benedek Rozemberczki; Lauren Watson; Péter Bayer; Hao-Tsung Yang; Olivér Kiss; Sebastian Nilsson; Rik Sarkar

doi:10.24963/ijcai.2022/778

The Shapley Value in Machine Learning

Benedek Rozemberczki, Lauren Watson, Péter Bayer, Hao-Tsung Yang, Olivér Kiss, Sebastian Nilsson, Rik Sarkar

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract / Description of output

Over the last few years, the Shapley value, a solution concept from cooperative game theory, has found numerous applications in machine learning. In this paper, we first discuss fundamental concepts of cooperative game theory and axiomatic properties of the Shapley value. Then we give an overview of the most important applications of the Shapley value in machine learning: feature selection, explainability, multi-agent reinforcement learning, ensemble pruning, and data valuation. We examine the most crucial limitations of the Shapley value and point out directions for future research.

Original language	English
Title of host publication	Proceedings of the 31st International Joint Conference on Artifical Intelligence, IJCAI-ECAI 2022
Editors	Luc De Raedt
Publisher	International Joint Conferences on Artificial Intelligence Organization
Pages	5572-5579
Number of pages	8
ISBN (Print)	978-1-956792-00-3
DOIs	https://doi.org/10.24963/ijcai.2022/778
Publication status	Published - 23 Jul 2022
Event	The 31st International Joint Conference on Artificial Intelligence and the 25th European Conference on Artificial Intelligence - Vienna, Austria Duration: 23 Jul 2022 → 29 Jul 2022 https://ijcai-22.org/

Conference

Conference	The 31st International Joint Conference on Artificial Intelligence and the 25th European Conference on Artificial Intelligence
Abbreviated title	IJCAI-ECAI 2022
Country/Territory	Austria
City	Vienna
Period	23/07/22 → 29/07/22
Internet address	https://ijcai-22.org/

Access to Document

10.24963/ijcai.2022/778Licence: All Rights Reserved

2202.05594v2Accepted author manuscript, 1 MBLicence: Creative Commons: Attribution (CC-BY)

https://www.ijcai.org/proceedings/2022/778Licence: All Rights Reserved

REPHRAIN: Research centre on Privacy, Harm Reduction and Adversarial Influence online
Elahi, T., Nissen, B. & Vaniea, K.
EPSRC
1/10/20 → 30/09/22
Project: Research

Cite this

Rozemberczki, B., Watson, L., Bayer, P., Yang, H-T., Kiss, O., Nilsson, S., & Sarkar, R. (2022). The Shapley Value in Machine Learning. In L. De Raedt (Ed.), Proceedings of the 31st International Joint Conference on Artifical Intelligence, IJCAI-ECAI 2022 (pp. 5572-5579). International Joint Conferences on Artificial Intelligence Organization. https://doi.org/10.24963/ijcai.2022/778

@inproceedings{c86b81c9231b4414be4b3df12fe2f0b7,

title = "The Shapley Value in Machine Learning",

abstract = "Over the last few years, the Shapley value, a solution concept from cooperative game theory, has found numerous applications in machine learning. In this paper, we first discuss fundamental concepts of cooperative game theory and axiomatic properties of the Shapley value. Then we give an overview of the most important applications of the Shapley value in machine learning: feature selection, explainability, multi-agent reinforcement learning, ensemble pruning, and data valuation. We examine the most crucial limitations of the Shapley value and point out directions for future research.",

author = "Benedek Rozemberczki and Lauren Watson and P{\'e}ter Bayer and Hao-Tsung Yang and Oliv{\'e}r Kiss and Sebastian Nilsson and Rik Sarkar",

year = "2022",

month = jul,

day = "23",

doi = "10.24963/ijcai.2022/778",

language = "English",

isbn = "978-1-956792-00-3",

pages = "5572--5579",

editor = "{De Raedt}, Luc",

booktitle = "Proceedings of the 31st International Joint Conference on Artifical Intelligence, IJCAI-ECAI 2022",

publisher = "International Joint Conferences on Artificial Intelligence Organization",

note = "The 31st International Joint Conference on Artificial Intelligence and the 25th European Conference on Artificial Intelligence, IJCAI-ECAI 2022 ; Conference date: 23-07-2022 Through 29-07-2022",

url = "https://ijcai-22.org/",

}

Rozemberczki, B, Watson, L, Bayer, P, Yang, H-T, Kiss, O, Nilsson, S & Sarkar, R 2022, The Shapley Value in Machine Learning. in L De Raedt (ed.), Proceedings of the 31st International Joint Conference on Artifical Intelligence, IJCAI-ECAI 2022. International Joint Conferences on Artificial Intelligence Organization, pp. 5572-5579, The 31st International Joint Conference on Artificial Intelligence and the 25th European Conference on Artificial Intelligence, Vienna, Austria, 23/07/22. https://doi.org/10.24963/ijcai.2022/778

The Shapley Value in Machine Learning. / Rozemberczki, Benedek; Watson, Lauren; Bayer, Péter et al.
Proceedings of the 31st International Joint Conference on Artifical Intelligence, IJCAI-ECAI 2022. ed. / Luc De Raedt. International Joint Conferences on Artificial Intelligence Organization, 2022. p. 5572-5579.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - The Shapley Value in Machine Learning

AU - Rozemberczki, Benedek

AU - Watson, Lauren

AU - Bayer, Péter

AU - Yang, Hao-Tsung

AU - Kiss, Olivér

AU - Nilsson, Sebastian

AU - Sarkar, Rik

PY - 2022/7/23

Y1 - 2022/7/23

N2 - Over the last few years, the Shapley value, a solution concept from cooperative game theory, has found numerous applications in machine learning. In this paper, we first discuss fundamental concepts of cooperative game theory and axiomatic properties of the Shapley value. Then we give an overview of the most important applications of the Shapley value in machine learning: feature selection, explainability, multi-agent reinforcement learning, ensemble pruning, and data valuation. We examine the most crucial limitations of the Shapley value and point out directions for future research.

AB - Over the last few years, the Shapley value, a solution concept from cooperative game theory, has found numerous applications in machine learning. In this paper, we first discuss fundamental concepts of cooperative game theory and axiomatic properties of the Shapley value. Then we give an overview of the most important applications of the Shapley value in machine learning: feature selection, explainability, multi-agent reinforcement learning, ensemble pruning, and data valuation. We examine the most crucial limitations of the Shapley value and point out directions for future research.

U2 - 10.24963/ijcai.2022/778

DO - 10.24963/ijcai.2022/778

M3 - Conference contribution

SN - 978-1-956792-00-3

SP - 5572

EP - 5579

BT - Proceedings of the 31st International Joint Conference on Artifical Intelligence, IJCAI-ECAI 2022

A2 - De Raedt, Luc

PB - International Joint Conferences on Artificial Intelligence Organization

T2 - The 31st International Joint Conference on Artificial Intelligence and the 25th European Conference on Artificial Intelligence

Y2 - 23 July 2022 through 29 July 2022

ER -

The Shapley Value in Machine Learning

Abstract / Description of output

Conference

Access to Document

Fingerprint

Projects

REPHRAIN: Research centre on Privacy, Harm Reduction and Adversarial Influence online

Cite this