Exploring the Role of Local and Global Explanations in Recommender Systems

Marissa Radensky; Doug Downey; Kyle Lo; Zoran Popovic; Daniel S. Weld

doi:10.1145/3491101.3519795

Exploring the Role of Local and Global Explanations in Recommender Systems

Marissa Radensky, Doug Downey, Kyle Lo, Zoran Popovic, Daniel S. Weld

Computer Science

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

12 Scopus citations

Abstract

Explanations are well-known to improve recommender systems' transparency. These explanations may be local, explaining individual recommendations, or global, explaining the recommender model overall. Despite their widespread use, there has been little investigation into the relative benefits of the two explanation approaches. We conducted a 30-participant exploratory study and a 30-participant controlled user study with a research-paper recommender to analyze how providing local, global, or both explanations influences user understanding of system behavior. Our results provide evidence suggesting that both are more helpful than either alone for explaining how to improve recommendations, yet both appeared less helpful than global alone for efficiently identifying false positive and negative recommendations. However, we note that the two explanation approaches may be better compared in a higher-stakes or more opaque domain.

Original language	English (US)
Title of host publication	CHI 2022 - Extended Abstracts of the 2022 CHI Conference on Human Factors in Computing Systems
Publisher	Association for Computing Machinery
ISBN (Electronic)	9781450391566
DOIs	https://doi.org/10.1145/3491101.3519795
State	Published - Apr 28 2022
Event	2022 CHI Conference on Human Factors in Computing Systems, CHI EA 2022 - New Orleans, United States Duration: Apr 30 2022 → May 5 2022

Publication series

Name	Conference on Human Factors in Computing Systems - Proceedings

Conference

Conference	2022 CHI Conference on Human Factors in Computing Systems, CHI EA 2022
Country/Territory	United States
City	New Orleans
Period	4/30/22 → 5/5/22

Funding

This research was supported by the University of Washington WRF/Cable Professorship and the Allen Institute for Artifcial Intelligence (AI2). The authors thank Matt Latzke and Cecile Nguyen for their advice regarding the project’s user-interface design and code respectively. The authors thank all the participants and pilot participants who made this work possible. Finally, the authors thank the anonymous reviewers of this work for their valuable feedback.

Keywords

explainable AI
human-AI interaction

ASJC Scopus subject areas

Software
Human-Computer Interaction
Computer Graphics and Computer-Aided Design

Access to Document

10.1145/3491101.3519795

Cite this

Radensky, M., Downey, D., Lo, K., Popovic, Z., & Weld, D. S. (2022). Exploring the Role of Local and Global Explanations in Recommender Systems. In CHI 2022 - Extended Abstracts of the 2022 CHI Conference on Human Factors in Computing Systems Article 290 (Conference on Human Factors in Computing Systems - Proceedings). Association for Computing Machinery. https://doi.org/10.1145/3491101.3519795

@inproceedings{5ddfcf3536d14b3a8507d009f2c02605,

title = "Exploring the Role of Local and Global Explanations in Recommender Systems",

abstract = "Explanations are well-known to improve recommender systems' transparency. These explanations may be local, explaining individual recommendations, or global, explaining the recommender model overall. Despite their widespread use, there has been little investigation into the relative benefits of the two explanation approaches. We conducted a 30-participant exploratory study and a 30-participant controlled user study with a research-paper recommender to analyze how providing local, global, or both explanations influences user understanding of system behavior. Our results provide evidence suggesting that both are more helpful than either alone for explaining how to improve recommendations, yet both appeared less helpful than global alone for efficiently identifying false positive and negative recommendations. However, we note that the two explanation approaches may be better compared in a higher-stakes or more opaque domain.",

keywords = "explainable AI, human-AI interaction",

author = "Marissa Radensky and Doug Downey and Kyle Lo and Zoran Popovic and Weld, {Daniel S.}",

note = "Publisher Copyright: {\textcopyright} 2022 Owner/Author.; 2022 CHI Conference on Human Factors in Computing Systems, CHI EA 2022 ; Conference date: 30-04-2022 Through 05-05-2022",

year = "2022",

month = apr,

day = "28",

doi = "10.1145/3491101.3519795",

language = "English (US)",

series = "Conference on Human Factors in Computing Systems - Proceedings",

publisher = "Association for Computing Machinery",

booktitle = "CHI 2022 - Extended Abstracts of the 2022 CHI Conference on Human Factors in Computing Systems",

}

Radensky, M, Downey, D, Lo, K, Popovic, Z & Weld, DS 2022, Exploring the Role of Local and Global Explanations in Recommender Systems. in CHI 2022 - Extended Abstracts of the 2022 CHI Conference on Human Factors in Computing Systems., 290, Conference on Human Factors in Computing Systems - Proceedings, Association for Computing Machinery, 2022 CHI Conference on Human Factors in Computing Systems, CHI EA 2022, New Orleans, United States, 4/30/22. https://doi.org/10.1145/3491101.3519795

Exploring the Role of Local and Global Explanations in Recommender Systems. / Radensky, Marissa; Downey, Doug; Lo, Kyle et al.
CHI 2022 - Extended Abstracts of the 2022 CHI Conference on Human Factors in Computing Systems. Association for Computing Machinery, 2022. 290 (Conference on Human Factors in Computing Systems - Proceedings).

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

TY - GEN

T1 - Exploring the Role of Local and Global Explanations in Recommender Systems

AU - Radensky, Marissa

AU - Downey, Doug

AU - Lo, Kyle

AU - Popovic, Zoran

AU - Weld, Daniel S.

PY - 2022/4/28

Y1 - 2022/4/28

N2 - Explanations are well-known to improve recommender systems' transparency. These explanations may be local, explaining individual recommendations, or global, explaining the recommender model overall. Despite their widespread use, there has been little investigation into the relative benefits of the two explanation approaches. We conducted a 30-participant exploratory study and a 30-participant controlled user study with a research-paper recommender to analyze how providing local, global, or both explanations influences user understanding of system behavior. Our results provide evidence suggesting that both are more helpful than either alone for explaining how to improve recommendations, yet both appeared less helpful than global alone for efficiently identifying false positive and negative recommendations. However, we note that the two explanation approaches may be better compared in a higher-stakes or more opaque domain.

AB - Explanations are well-known to improve recommender systems' transparency. These explanations may be local, explaining individual recommendations, or global, explaining the recommender model overall. Despite their widespread use, there has been little investigation into the relative benefits of the two explanation approaches. We conducted a 30-participant exploratory study and a 30-participant controlled user study with a research-paper recommender to analyze how providing local, global, or both explanations influences user understanding of system behavior. Our results provide evidence suggesting that both are more helpful than either alone for explaining how to improve recommendations, yet both appeared less helpful than global alone for efficiently identifying false positive and negative recommendations. However, we note that the two explanation approaches may be better compared in a higher-stakes or more opaque domain.

KW - explainable AI

KW - human-AI interaction

UR - http://www.scopus.com/inward/record.url?scp=85129767101&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85129767101&partnerID=8YFLogxK

U2 - 10.1145/3491101.3519795

DO - 10.1145/3491101.3519795

M3 - Conference contribution

AN - SCOPUS:85129767101

T3 - Conference on Human Factors in Computing Systems - Proceedings

BT - CHI 2022 - Extended Abstracts of the 2022 CHI Conference on Human Factors in Computing Systems

PB - Association for Computing Machinery

T2 - 2022 CHI Conference on Human Factors in Computing Systems, CHI EA 2022

Y2 - 30 April 2022 through 5 May 2022

ER -

Exploring the Role of Local and Global Explanations in Recommender Systems

Abstract

Publication series

Conference

Funding

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this