Retrain AI Systems Responsibly! Use Sustainable Concept Drift Adaptation Techniques

Lorena Poenaru-Olaru; June Sallou; Luis Cruz; Jan S. Rellermeyer; Arie Van Deursen

doi:10.1109/GREENS59328.2023.00009

Retrain AI Systems Responsibly! Use Sustainable Concept Drift Adaptation Techniques

Lorena Poenaru-Olaru^*, June Sallou, Luis Cruz, Jan S. Rellermeyer, Arie Van Deursen

^*Corresponding author for this work

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

6 Downloads (Pure)

Abstract

Deployed machine learning systems often suffer from accuracy degradation over time generated by constant data shifts, also known as concept drift. Therefore, these systems require regular maintenance, in which the machine learning model needs to be adapted to concept drift. The literature presents plenty of model adaptation techniques. The most common technique is periodically executing the whole training pipeline with all the data gathered until a particular point in time, yielding a massive energy footprint. In this paper, we propose a research path that uses concept drift detection and adaptation to enable sustainable AI systems.

Original language	English
Title of host publication	Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	17-18
Number of pages	2
ISBN (Electronic)	9798350312386
DOIs	https://doi.org/10.1109/GREENS59328.2023.00009
Publication status	Published - 2023
Event	7th IEEE/ACM International Workshop on Green And Sustainable Software, GREENS 2023 - Melbourne, Australia Duration: 14 May 2023 → …

Publication series

Name	Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023

Conference

Conference	7th IEEE/ACM International Workshop on Green And Sustainable Software, GREENS 2023
Country/Territory	Australia
City	Melbourne
Period	14/05/23 → …

Bibliographical note

Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Keywords

concept drift adaptation
sustainable model maintenance
sustainable model retraining

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1109/GREENS59328.2023.00009

Retrain_AI_Systems_Responsibly_Use_Sustainable_Concept_Drift_Adaptation_TechniquesFinal published version, 183 KB

Cite this

Poenaru-Olaru, L., Sallou, J., Cruz, L., Rellermeyer, J. S., & Van Deursen, A. (2023). Retrain AI Systems Responsibly! Use Sustainable Concept Drift Adaptation Techniques. In Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023 (pp. 17-18). (Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023). Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/GREENS59328.2023.00009

Poenaru-Olaru, Lorena ; Sallou, June ; Cruz, Luis et al. / Retrain AI Systems Responsibly! Use Sustainable Concept Drift Adaptation Techniques. Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023. Institute of Electrical and Electronics Engineers (IEEE), 2023. pp. 17-18 (Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023).

@inproceedings{d115c5baf67147139513a0c7c5abf7f1,

title = "Retrain AI Systems Responsibly! Use Sustainable Concept Drift Adaptation Techniques",

abstract = "Deployed machine learning systems often suffer from accuracy degradation over time generated by constant data shifts, also known as concept drift. Therefore, these systems require regular maintenance, in which the machine learning model needs to be adapted to concept drift. The literature presents plenty of model adaptation techniques. The most common technique is periodically executing the whole training pipeline with all the data gathered until a particular point in time, yielding a massive energy footprint. In this paper, we propose a research path that uses concept drift detection and adaptation to enable sustainable AI systems.",

keywords = "concept drift adaptation, sustainable model maintenance, sustainable model retraining",

author = "Lorena Poenaru-Olaru and June Sallou and Luis Cruz and Rellermeyer, {Jan S.} and {Van Deursen}, Arie",

note = "Green Open Access added to TU Delft Institutional Repository {\textquoteleft}You share, we take care!{\textquoteright} – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public. ; 7th IEEE/ACM International Workshop on Green And Sustainable Software, GREENS 2023 ; Conference date: 14-05-2023",

year = "2023",

doi = "10.1109/GREENS59328.2023.00009",

language = "English",

series = "Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

pages = "17--18",

booktitle = "Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023",

address = "United States",

}

Poenaru-Olaru, L , Sallou, J , Cruz, L , Rellermeyer, JS & Van Deursen, A 2023, Retrain AI Systems Responsibly! Use Sustainable Concept Drift Adaptation Techniques. in Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023. Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023, Institute of Electrical and Electronics Engineers (IEEE), pp. 17-18, 7th IEEE/ACM International Workshop on Green And Sustainable Software, GREENS 2023, Melbourne, Australia, 14/05/23. https://doi.org/10.1109/GREENS59328.2023.00009

Retrain AI Systems Responsibly! Use Sustainable Concept Drift Adaptation Techniques. / Poenaru-Olaru, Lorena ; Sallou, June ; Cruz, Luis et al.
Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023. Institute of Electrical and Electronics Engineers (IEEE), 2023. p. 17-18 (Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023).

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Retrain AI Systems Responsibly! Use Sustainable Concept Drift Adaptation Techniques

AU - Poenaru-Olaru, Lorena

AU - Sallou, June

AU - Cruz, Luis

AU - Rellermeyer, Jan S.

AU - Van Deursen, Arie

N1 - Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

PY - 2023

Y1 - 2023

N2 - Deployed machine learning systems often suffer from accuracy degradation over time generated by constant data shifts, also known as concept drift. Therefore, these systems require regular maintenance, in which the machine learning model needs to be adapted to concept drift. The literature presents plenty of model adaptation techniques. The most common technique is periodically executing the whole training pipeline with all the data gathered until a particular point in time, yielding a massive energy footprint. In this paper, we propose a research path that uses concept drift detection and adaptation to enable sustainable AI systems.

AB - Deployed machine learning systems often suffer from accuracy degradation over time generated by constant data shifts, also known as concept drift. Therefore, these systems require regular maintenance, in which the machine learning model needs to be adapted to concept drift. The literature presents plenty of model adaptation techniques. The most common technique is periodically executing the whole training pipeline with all the data gathered until a particular point in time, yielding a massive energy footprint. In this paper, we propose a research path that uses concept drift detection and adaptation to enable sustainable AI systems.

KW - concept drift adaptation

KW - sustainable model maintenance

KW - sustainable model retraining

UR - http://www.scopus.com/inward/record.url?scp=85168082338&partnerID=8YFLogxK

U2 - 10.1109/GREENS59328.2023.00009

DO - 10.1109/GREENS59328.2023.00009

M3 - Conference contribution

AN - SCOPUS:85168082338

T3 - Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023

SP - 17

EP - 18

BT - Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023

PB - Institute of Electrical and Electronics Engineers (IEEE)

T2 - 7th IEEE/ACM International Workshop on Green And Sustainable Software, GREENS 2023

Y2 - 14 May 2023

ER -

Poenaru-Olaru L , Sallou J , Cruz L , Rellermeyer JS , Van Deursen A. Retrain AI Systems Responsibly! Use Sustainable Concept Drift Adaptation Techniques. In Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023. Institute of Electrical and Electronics Engineers (IEEE). 2023. p. 17-18. (Proceedings - 2023 IEEE/ACM 7th International Workshop on Green And Sustainable Software, GREENS 2023). doi: 10.1109/GREENS59328.2023.00009

Retrain AI Systems Responsibly! Use Sustainable Concept Drift Adaptation Techniques

Abstract

Publication series

Conference

Bibliographical note

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this