Explaining Memristive Reservoir Computing Through Evolving Feature Attribution

Xinming Shi; Zilu Wang; Leandro Minku; Xin Yao

doi:10.1145/3583133.3590619

Explaining Memristive Reservoir Computing Through Evolving Feature Attribution

Xinming Shi, Zilu Wang, Leandro Minku, Xin Yao

Computer Science

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

Abstract

Memristive Reservoir Computing (MRC) is a promising computing architecture for time series tasks, but lacks explainability, leading to unreliable predictions. To address this issue, we propose an evolutionary framework to explain the time series predictions of MRC systems. Our proposed approach attributes the feature importance of the time series via an evolutionary approach to explain the predictions. Our experiments show that our approach successfully identified the most influential factors, demonstrating the effectiveness of our design and its superiority in terms of explanation compared to state-of-the-art methods.

Original language	English
Title of host publication	GECCO '23 Companion
Subtitle of host publication	Proceedings of the Companion Conference on Genetic and Evolutionary Computation
Publisher	Association for Computing Machinery (ACM)
Pages	683–686
Number of pages	4
ISBN (Electronic)	9798400701207
DOIs	https://doi.org/10.1145/3583133.3590619
Publication status	Published - 24 Jul 2023

Publication series

Name	GECCO: Genetic and Evolutionary Computation Conference

Bibliographical note

Acknowledgments:
This work was supported by NSFC (Grant No. 62250710682), the Research Institute of Trustworthy Autonomous Systems (RITAS), the Guangdong Provincial Key Laboratory (Grant No. 2020B121201001), the Program for Guangdong Introducing Innovative and Enterpreneurial Teams (Grant No. 2017ZT07X386), the Shenzhen Science and Technology Program (Grant No. KQTD2016112514355531).

Keywords

Explainability
evolutionary algorithm
reservoir computing
memristor

Access to Document

10.1145/3583133.3590619

Cite this

Shi, X., Wang, Z., Minku, L., & Yao, X. (2023). Explaining Memristive Reservoir Computing Through Evolving Feature Attribution. In GECCO '23 Companion: Proceedings of the Companion Conference on Genetic and Evolutionary Computation (pp. 683–686). (GECCO: Genetic and Evolutionary Computation Conference). Association for Computing Machinery (ACM). https://doi.org/10.1145/3583133.3590619

@inproceedings{57313edca59449cda1d053892df20e3f,

title = "Explaining Memristive Reservoir Computing Through Evolving Feature Attribution",

abstract = "Memristive Reservoir Computing (MRC) is a promising computing architecture for time series tasks, but lacks explainability, leading to unreliable predictions. To address this issue, we propose an evolutionary framework to explain the time series predictions of MRC systems. Our proposed approach attributes the feature importance of the time series via an evolutionary approach to explain the predictions. Our experiments show that our approach successfully identified the most influential factors, demonstrating the effectiveness of our design and its superiority in terms of explanation compared to state-of-the-art methods.",

keywords = "Explainability, evolutionary algorithm, reservoir computing, memristor",

author = "Xinming Shi and Zilu Wang and Leandro Minku and Xin Yao",

note = "Acknowledgments: This work was supported by NSFC (Grant No. 62250710682), the Research Institute of Trustworthy Autonomous Systems (RITAS), the Guangdong Provincial Key Laboratory (Grant No. 2020B121201001), the Program for Guangdong Introducing Innovative and Enterpreneurial Teams (Grant No. 2017ZT07X386), the Shenzhen Science and Technology Program (Grant No. KQTD2016112514355531). ",

year = "2023",

month = jul,

day = "24",

doi = "10.1145/3583133.3590619",

language = "English",

series = "GECCO: Genetic and Evolutionary Computation Conference",

publisher = "Association for Computing Machinery (ACM)",

pages = "683–686",

booktitle = "GECCO '23 Companion",

address = "United States",

}

Shi, X, Wang, Z, Minku, L & Yao, X 2023, Explaining Memristive Reservoir Computing Through Evolving Feature Attribution. in GECCO '23 Companion: Proceedings of the Companion Conference on Genetic and Evolutionary Computation. GECCO: Genetic and Evolutionary Computation Conference, Association for Computing Machinery (ACM), pp. 683–686. https://doi.org/10.1145/3583133.3590619

Explaining Memristive Reservoir Computing Through Evolving Feature Attribution. / Shi, Xinming; Wang, Zilu; Minku, Leandro et al.
GECCO '23 Companion: Proceedings of the Companion Conference on Genetic and Evolutionary Computation. Association for Computing Machinery (ACM), 2023. p. 683–686 (GECCO: Genetic and Evolutionary Computation Conference).

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

TY - GEN

T1 - Explaining Memristive Reservoir Computing Through Evolving Feature Attribution

AU - Shi, Xinming

AU - Wang, Zilu

AU - Minku, Leandro

AU - Yao, Xin

N1 - Acknowledgments: This work was supported by NSFC (Grant No. 62250710682), the Research Institute of Trustworthy Autonomous Systems (RITAS), the Guangdong Provincial Key Laboratory (Grant No. 2020B121201001), the Program for Guangdong Introducing Innovative and Enterpreneurial Teams (Grant No. 2017ZT07X386), the Shenzhen Science and Technology Program (Grant No. KQTD2016112514355531).

PY - 2023/7/24

Y1 - 2023/7/24

N2 - Memristive Reservoir Computing (MRC) is a promising computing architecture for time series tasks, but lacks explainability, leading to unreliable predictions. To address this issue, we propose an evolutionary framework to explain the time series predictions of MRC systems. Our proposed approach attributes the feature importance of the time series via an evolutionary approach to explain the predictions. Our experiments show that our approach successfully identified the most influential factors, demonstrating the effectiveness of our design and its superiority in terms of explanation compared to state-of-the-art methods.

AB - Memristive Reservoir Computing (MRC) is a promising computing architecture for time series tasks, but lacks explainability, leading to unreliable predictions. To address this issue, we propose an evolutionary framework to explain the time series predictions of MRC systems. Our proposed approach attributes the feature importance of the time series via an evolutionary approach to explain the predictions. Our experiments show that our approach successfully identified the most influential factors, demonstrating the effectiveness of our design and its superiority in terms of explanation compared to state-of-the-art methods.

KW - Explainability

KW - evolutionary algorithm

KW - reservoir computing

KW - memristor

UR - https://dl.acm.org/conference/gecco

U2 - 10.1145/3583133.3590619

DO - 10.1145/3583133.3590619

M3 - Conference contribution

T3 - GECCO: Genetic and Evolutionary Computation Conference

SP - 683

EP - 686

BT - GECCO '23 Companion

PB - Association for Computing Machinery (ACM)

ER -

Explaining Memristive Reservoir Computing Through Evolving Feature Attribution

Abstract

Publication series

Bibliographical note

Keywords

Access to Document

Fingerprint

Cite this