Benefits of Machine Learning Explanations: Improved Learning in an AI-assisted Sequence Prediction Task

Yu Liang; Dennis Collaris; Martijn C. Willemsen; Jack J. van Wijk

doi:10.1145/3708359.3712155

Benefits of Machine Learning Explanations: Improved Learning in an AI-assisted Sequence Prediction Task

Yu Liang, Dennis Collaris, Martijn C. Willemsen, Jack J. van Wijk

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

1 Downloads (Pure)

Abstract

Research in Explainable AI (XAI) has shown that explanations can improve users’ understanding of AI models, improve user performance and potentially reduce overreliance on AI predictions. However, this is mostly evaluated by static rather than dynamic measures, and the role of XAI on learning over trials is rarely studied. In this study, we use a context-free sequence prediction task, in which 458 participants predict the next symbol in a fixed sequence (with some noise) over 80 trials. We compare performance with AI and XAI advice against no AI support, and subsequently we test for learning by taking away the AI support after 40 trials (i.e., a reversal study design). Our results show that users learn faster with XAI than with AI without explanations or no AI and are better able to recover in performance from the removal of AI. However, the benefits of XAI on learning are much smaller for more difficult tasks. This work demonstrates the benefits of repeated measures user studies and multilevel modeling to better understand learning processes in XAI. It also shows the potential of AI explanations to help users to learn and poses XAI design suggestions to support learning in human-AI collaboration.

Original language	English
Title of host publication	IUI '25
Subtitle of host publication	Proceedings of the 30th International Conference on Intelligent User Interfaces
Editors	Toby Li, Fabio Paternò, Kaisa Väänänen, Luis Leiva, Davide Spano, Katrien Verbert
Place of Publication	New York
Publisher	Association for Computing Machinery, Inc
Pages	231-246
Number of pages	16
ISBN (Electronic)	979-8-4007-1306-4
DOIs	https://doi.org/10.1145/3708359.3712155
Publication status	Published - 24 Mar 2025
Event	30th Annual ACM Conference on Intelligent User Interfaces 2025 - Cagliari, Italy, Cagliari, Italy Duration: 24 Mar 2025 → 27 Mar 2025 https://iui.acm.org/2025/

Conference

Conference	30th Annual ACM Conference on Intelligent User Interfaces 2025
Abbreviated title	IUI 2025
Country/Territory	Italy
City	Cagliari
Period	24/03/25 → 27/03/25
Internet address	https://iui.acm.org/2025/

Funding

This work is part of the research programme TEPAIV with project number 612.001.752, which is fnanced by the Dutch Research Council (NWO).

Funders	Funder number
Nederlandse Organisatie voor Wetenschappelijk Onderzoek

Keywords

machine learning
interpretability
explainability

Access to Document

10.1145/3708359.3712155

pdfFinal published version, 3.87 MBLicence: CC BY

Cite this

Liang, Y., Collaris, D., Willemsen, M. C., & van Wijk, J. J. (2025). Benefits of Machine Learning Explanations: Improved Learning in an AI-assisted Sequence Prediction Task. In T. Li, F. Paternò, K. Väänänen, L. Leiva, D. Spano, & K. Verbert (Eds.), IUI '25: Proceedings of the 30th International Conference on Intelligent User Interfaces (pp. 231-246). Association for Computing Machinery, Inc. https://doi.org/10.1145/3708359.3712155

Liang, Yu ; Collaris, Dennis ; Willemsen, Martijn C. et al. / Benefits of Machine Learning Explanations : Improved Learning in an AI-assisted Sequence Prediction Task. IUI '25: Proceedings of the 30th International Conference on Intelligent User Interfaces. editor / Toby Li ; Fabio Paternò ; Kaisa Väänänen ; Luis Leiva ; Davide Spano ; Katrien Verbert. New York : Association for Computing Machinery, Inc, 2025. pp. 231-246

@inproceedings{7e67d6b93f134e1b93c5d0934c516acf,

title = "Benefits of Machine Learning Explanations: Improved Learning in an AI-assisted Sequence Prediction Task",

abstract = "Research in Explainable AI (XAI) has shown that explanations can improve users{\textquoteright} understanding of AI models, improve user performance and potentially reduce overreliance on AI predictions. However, this is mostly evaluated by static rather than dynamic measures, and the role of XAI on learning over trials is rarely studied. In this study, we use a context-free sequence prediction task, in which 458 participants predict the next symbol in a fixed sequence (with some noise) over 80 trials. We compare performance with AI and XAI advice against no AI support, and subsequently we test for learning by taking away the AI support after 40 trials (i.e., a reversal study design). Our results show that users learn faster with XAI than with AI without explanations or no AI and are better able to recover in performance from the removal of AI. However, the benefits of XAI on learning are much smaller for more difficult tasks. This work demonstrates the benefits of repeated measures user studies and multilevel modeling to better understand learning processes in XAI. It also shows the potential of AI explanations to help users to learn and poses XAI design suggestions to support learning in human-AI collaboration.",

keywords = "machine learning, interpretability, explainability",

author = "Yu Liang and Dennis Collaris and Willemsen, {Martijn C.} and {van Wijk}, {Jack J.}",

year = "2025",

month = mar,

day = "24",

doi = "10.1145/3708359.3712155",

language = "English",

pages = "231--246",

editor = "Toby Li and Fabio Patern{\`o} and Kaisa V{\"a}{\"a}n{\"a}nen and Luis Leiva and Davide Spano and Katrien Verbert",

booktitle = "IUI '25",

publisher = "Association for Computing Machinery, Inc",

address = "United States",

note = "30th Annual ACM Conference on Intelligent User Interfaces 2025, IUI 2025 ; Conference date: 24-03-2025 Through 27-03-2025",

url = "https://iui.acm.org/2025/",

}

Liang, Y , Collaris, D , Willemsen, MC & van Wijk, JJ 2025, Benefits of Machine Learning Explanations: Improved Learning in an AI-assisted Sequence Prediction Task. in T Li, F Paternò, K Väänänen, L Leiva, D Spano & K Verbert (eds), IUI '25: Proceedings of the 30th International Conference on Intelligent User Interfaces. Association for Computing Machinery, Inc, New York, pp. 231-246, 30th Annual ACM Conference on Intelligent User Interfaces 2025, Cagliari, Italy, 24/03/25. https://doi.org/10.1145/3708359.3712155

Benefits of Machine Learning Explanations: Improved Learning in an AI-assisted Sequence Prediction Task. / Liang, Yu ; Collaris, Dennis ; Willemsen, Martijn C. et al.
IUI '25: Proceedings of the 30th International Conference on Intelligent User Interfaces. ed. / Toby Li; Fabio Paternò; Kaisa Väänänen; Luis Leiva; Davide Spano; Katrien Verbert. New York: Association for Computing Machinery, Inc, 2025. p. 231-246.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Benefits of Machine Learning Explanations

T2 - 30th Annual ACM Conference on Intelligent User Interfaces 2025

AU - Liang, Yu

AU - Collaris, Dennis

AU - Willemsen, Martijn C.

AU - van Wijk, Jack J.

PY - 2025/3/24

Y1 - 2025/3/24

N2 - Research in Explainable AI (XAI) has shown that explanations can improve users’ understanding of AI models, improve user performance and potentially reduce overreliance on AI predictions. However, this is mostly evaluated by static rather than dynamic measures, and the role of XAI on learning over trials is rarely studied. In this study, we use a context-free sequence prediction task, in which 458 participants predict the next symbol in a fixed sequence (with some noise) over 80 trials. We compare performance with AI and XAI advice against no AI support, and subsequently we test for learning by taking away the AI support after 40 trials (i.e., a reversal study design). Our results show that users learn faster with XAI than with AI without explanations or no AI and are better able to recover in performance from the removal of AI. However, the benefits of XAI on learning are much smaller for more difficult tasks. This work demonstrates the benefits of repeated measures user studies and multilevel modeling to better understand learning processes in XAI. It also shows the potential of AI explanations to help users to learn and poses XAI design suggestions to support learning in human-AI collaboration.

AB - Research in Explainable AI (XAI) has shown that explanations can improve users’ understanding of AI models, improve user performance and potentially reduce overreliance on AI predictions. However, this is mostly evaluated by static rather than dynamic measures, and the role of XAI on learning over trials is rarely studied. In this study, we use a context-free sequence prediction task, in which 458 participants predict the next symbol in a fixed sequence (with some noise) over 80 trials. We compare performance with AI and XAI advice against no AI support, and subsequently we test for learning by taking away the AI support after 40 trials (i.e., a reversal study design). Our results show that users learn faster with XAI than with AI without explanations or no AI and are better able to recover in performance from the removal of AI. However, the benefits of XAI on learning are much smaller for more difficult tasks. This work demonstrates the benefits of repeated measures user studies and multilevel modeling to better understand learning processes in XAI. It also shows the potential of AI explanations to help users to learn and poses XAI design suggestions to support learning in human-AI collaboration.

KW - machine learning

KW - interpretability

KW - explainability

UR - http://www.scopus.com/inward/record.url?scp=105001922664&partnerID=8YFLogxK

U2 - 10.1145/3708359.3712155

DO - 10.1145/3708359.3712155

M3 - Conference contribution

SP - 231

EP - 246

BT - IUI '25

A2 - Li, Toby

A2 - Paternò, Fabio

A2 - Väänänen, Kaisa

A2 - Leiva, Luis

A2 - Spano, Davide

A2 - Verbert, Katrien

PB - Association for Computing Machinery, Inc

CY - New York

Y2 - 24 March 2025 through 27 March 2025

ER -

Liang Y , Collaris D , Willemsen MC , van Wijk JJ. Benefits of Machine Learning Explanations: Improved Learning in an AI-assisted Sequence Prediction Task. In Li T, Paternò F, Väänänen K, Leiva L, Spano D, Verbert K, editors, IUI '25: Proceedings of the 30th International Conference on Intelligent User Interfaces. New York: Association for Computing Machinery, Inc. 2025. p. 231-246 doi: 10.1145/3708359.3712155

Benefits of Machine Learning Explanations: Improved Learning in an AI-assisted Sequence Prediction Task

Abstract

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

TEPAIV: TEPAIV

Cite this

Benefits of Machine Learning Explanations: Improved Learning in an AI-assisted Sequence Prediction Task

Abstract

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Projects

TEPAIV: TEPAIV

Cite this