Deep Integrated Explanations

Oren Barkan; Yehonathan Elisha; Jonathan Weill; Yuval Asher; Amit Eshel; Noam Koenigstein

doi:10.1145/3583780.3614836

Deep Integrated Explanations

Oren Barkan, Yehonathan Elisha, Jonathan Weill, Yuval Asher, Amit Eshel, Noam Koenigstein

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

Abstract

This paper presents Deep Integrated Explanations (DIX) - a universal method for explaining vision models. DIX generates explanation maps by integrating information from the intermediate representations of the model, coupled with their corresponding gradients. Through an extensive array of both objective and subjective evaluations spanning diverse tasks, datasets, and model configurations, we showcase the efficacy of DIX in generating faithful and accurate explanation maps, while surpassing current state-of-the-art methods. Our code is available at: https://github.com/dix-cikm23/dix.

Original language	English
Title of host publication	CIKM 2023 - Proceedings of the 32nd ACM International Conference on Information and Knowledge Management
Publisher	Association for Computing Machinery
Pages	57-67
Number of pages	11
ISBN (Electronic)	9798400701245
DOIs	https://doi.org/10.1145/3583780.3614836
State	Published - 21 Oct 2023
Externally published	Yes
Event	32nd ACM International Conference on Information and Knowledge Management, CIKM 2023 - Birmingham, United Kingdom Duration: 21 Oct 2023 → 25 Oct 2023

Publication series

Name	International Conference on Information and Knowledge Management, Proceedings

Conference

Conference	32nd ACM International Conference on Information and Knowledge Management, CIKM 2023
Country/Territory	United Kingdom
City	Birmingham
Period	21/10/23 → 25/10/23

Bibliographical note

Publisher Copyright:
© 2023 Copyright held by the owner/author(s). Publication rights licensed to ACM.

DBLP License: DBLP's bibliographic metadata records provided through http://dblp.org/ are distributed under a Creative Commons CC0 1.0 Universal Public Domain Dedication. Although the bibliographic metadata records are provided consistent with CC0 1.0 Dedication, the content described by the metadata records is not. Content may be subject to copyright, rights of privacy, rights of publicity and other restrictions.

Keywords

Computer Vision
Deep Learning
Explainable AI

Access to Document

10.1145/3583780.3614836

Cite this

Barkan, O., Elisha, Y., Weill, J., Asher, Y., Eshel, A., & Koenigstein, N. (2023). Deep Integrated Explanations. In CIKM 2023 - Proceedings of the 32nd ACM International Conference on Information and Knowledge Management (pp. 57-67). (International Conference on Information and Knowledge Management, Proceedings). Association for Computing Machinery. https://doi.org/10.1145/3583780.3614836

@inproceedings{331be1fdb57b43f1b721a2994731f937,

title = "Deep Integrated Explanations",

abstract = "This paper presents Deep Integrated Explanations (DIX) - a universal method for explaining vision models. DIX generates explanation maps by integrating information from the intermediate representations of the model, coupled with their corresponding gradients. Through an extensive array of both objective and subjective evaluations spanning diverse tasks, datasets, and model configurations, we showcase the efficacy of DIX in generating faithful and accurate explanation maps, while surpassing current state-of-the-art methods. Our code is available at: https://github.com/dix-cikm23/dix.",

keywords = "Computer Vision, Deep Learning, Explainable AI",

author = "Oren Barkan and Yehonathan Elisha and Jonathan Weill and Yuval Asher and Amit Eshel and Noam Koenigstein",

note = "Publisher Copyright: {\textcopyright} 2023 Copyright held by the owner/author(s). Publication rights licensed to ACM.; 32nd ACM International Conference on Information and Knowledge Management, CIKM 2023 ; Conference date: 21-10-2023 Through 25-10-2023",

year = "2023",

month = oct,

day = "21",

doi = "10.1145/3583780.3614836",

language = "אנגלית",

series = "International Conference on Information and Knowledge Management, Proceedings",

publisher = "Association for Computing Machinery",

pages = "57--67",

booktitle = "CIKM 2023 - Proceedings of the 32nd ACM International Conference on Information and Knowledge Management",

}

Barkan, O, Elisha, Y, Weill, J, Asher, Y, Eshel, A & Koenigstein, N 2023, Deep Integrated Explanations. in CIKM 2023 - Proceedings of the 32nd ACM International Conference on Information and Knowledge Management. International Conference on Information and Knowledge Management, Proceedings, Association for Computing Machinery, pp. 57-67, 32nd ACM International Conference on Information and Knowledge Management, CIKM 2023, Birmingham, United Kingdom, 21/10/23. https://doi.org/10.1145/3583780.3614836

Deep Integrated Explanations. / Barkan, Oren; Elisha, Yehonathan; Weill, Jonathan et al.
CIKM 2023 - Proceedings of the 32nd ACM International Conference on Information and Knowledge Management. Association for Computing Machinery, 2023. p. 57-67 (International Conference on Information and Knowledge Management, Proceedings).

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

TY - GEN

T1 - Deep Integrated Explanations

AU - Barkan, Oren

AU - Elisha, Yehonathan

AU - Weill, Jonathan

AU - Asher, Yuval

AU - Eshel, Amit

AU - Koenigstein, Noam

PY - 2023/10/21

Y1 - 2023/10/21

N2 - This paper presents Deep Integrated Explanations (DIX) - a universal method for explaining vision models. DIX generates explanation maps by integrating information from the intermediate representations of the model, coupled with their corresponding gradients. Through an extensive array of both objective and subjective evaluations spanning diverse tasks, datasets, and model configurations, we showcase the efficacy of DIX in generating faithful and accurate explanation maps, while surpassing current state-of-the-art methods. Our code is available at: https://github.com/dix-cikm23/dix.

AB - This paper presents Deep Integrated Explanations (DIX) - a universal method for explaining vision models. DIX generates explanation maps by integrating information from the intermediate representations of the model, coupled with their corresponding gradients. Through an extensive array of both objective and subjective evaluations spanning diverse tasks, datasets, and model configurations, we showcase the efficacy of DIX in generating faithful and accurate explanation maps, while surpassing current state-of-the-art methods. Our code is available at: https://github.com/dix-cikm23/dix.

KW - Computer Vision

KW - Deep Learning

KW - Explainable AI

UR - http://www.scopus.com/inward/record.url?scp=85175686880&partnerID=8YFLogxK

U2 - 10.1145/3583780.3614836

DO - 10.1145/3583780.3614836

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:85175686880

T3 - International Conference on Information and Knowledge Management, Proceedings

SP - 57

EP - 67

BT - CIKM 2023 - Proceedings of the 32nd ACM International Conference on Information and Knowledge Management

PB - Association for Computing Machinery

T2 - 32nd ACM International Conference on Information and Knowledge Management, CIKM 2023

Y2 - 21 October 2023 through 25 October 2023

ER -

Deep Integrated Explanations

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this