Model- and data-agnostic justifications with a fortiori case-based argumentation

Joeri Peters; Floris Bex; Henry Prakken

doi:10.1145/3594536.3595164

Model- and data-agnostic justifications with a fortiori case-based argumentation

Joeri Peters, Floris Bex, Henry Prakken

Sub Intelligent Systems

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

Abstract

AF-CBA is an example-based approach to XAI that draws on the case-based argumentation tradition in AI & Law. It means to explain binary classifications made by an opaque machine-learning model by presenting an argument graph to the user, which represents an argument game about the classification of a case on the basis of precedents derived from labelled data used in the training phase of the classifier. We improve the robustness of this method by modifying it to better handle inconsistent labelling and evaluate an alternative setup that does not require access to the labelled data by using earlier predictions instead.

Original language	English
Title of host publication	ICAIL '23: Proceedings of the Nineteenth International Conference on Artificial Intelligence and Law
Editors	Centro Algoritmi, Thomson Reuters
Publisher	Association for Computing Machinery
Pages	207-216
Number of pages	10
ISBN (Print)	979-8-4007-0197-9
DOIs	https://doi.org/10.1145/3594536.3595164
Publication status	Published - 7 Sept 2023

Bibliographical note

Publisher Copyright:
© ICAIL 2023. All rights reserved.

Funding

The authors would like to thank the anonymous reviewers for their feedback and suggestions.

Keywords

CBR
XAI
argumentation
precedential constraint

Access to Document

10.1145/3594536.3595164

Peters-Bex-Prakken_ModelDataAgnostic_Justifications_CBAFinal published version, 819 KBLicence: Taverne

Cite this

@inproceedings{9ee167b2b5314aa3b2e8ff4bc41dd39f,

title = "Model- and data-agnostic justifications with a fortiori case-based argumentation",

abstract = "AF-CBA is an example-based approach to XAI that draws on the case-based argumentation tradition in AI \& Law. It means to explain binary classifications made by an opaque machine-learning model by presenting an argument graph to the user, which represents an argument game about the classification of a case on the basis of precedents derived from labelled data used in the training phase of the classifier. We improve the robustness of this method by modifying it to better handle inconsistent labelling and evaluate an alternative setup that does not require access to the labelled data by using earlier predictions instead.",

keywords = "CBR, XAI, argumentation, precedential constraint",

author = "Joeri Peters and Floris Bex and Henry Prakken",

year = "2023",

month = sep,

day = "7",

doi = "10.1145/3594536.3595164",

language = "English",

isbn = "979-8-4007-0197-9",

pages = "207--216",

editor = "Algoritmi, \{Centro \} and Reuters, \{Thomson \}",

booktitle = "ICAIL '23: Proceedings of the Nineteenth International Conference on Artificial Intelligence and Law",

publisher = "Association for Computing Machinery",

}

Model- and data-agnostic justifications with a fortiori case-based argumentation. / Peters, Joeri ; Bex, Floris ; Prakken, Henry.
ICAIL '23: Proceedings of the Nineteenth International Conference on Artificial Intelligence and Law. ed. / Centro Algoritmi; Thomson Reuters. Association for Computing Machinery, 2023. p. 207-216.

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

TY - GEN

T1 - Model- and data-agnostic justifications with a fortiori case-based argumentation

AU - Peters, Joeri

AU - Bex, Floris

AU - Prakken, Henry

PY - 2023/9/7

Y1 - 2023/9/7

N2 - AF-CBA is an example-based approach to XAI that draws on the case-based argumentation tradition in AI & Law. It means to explain binary classifications made by an opaque machine-learning model by presenting an argument graph to the user, which represents an argument game about the classification of a case on the basis of precedents derived from labelled data used in the training phase of the classifier. We improve the robustness of this method by modifying it to better handle inconsistent labelling and evaluate an alternative setup that does not require access to the labelled data by using earlier predictions instead.

AB - AF-CBA is an example-based approach to XAI that draws on the case-based argumentation tradition in AI & Law. It means to explain binary classifications made by an opaque machine-learning model by presenting an argument graph to the user, which represents an argument game about the classification of a case on the basis of precedents derived from labelled data used in the training phase of the classifier. We improve the robustness of this method by modifying it to better handle inconsistent labelling and evaluate an alternative setup that does not require access to the labelled data by using earlier predictions instead.

KW - CBR

KW - XAI

KW - argumentation

KW - precedential constraint

UR - https://www.scopus.com/pages/publications/85175681663

U2 - 10.1145/3594536.3595164

DO - 10.1145/3594536.3595164

M3 - Conference contribution

SN - 979-8-4007-0197-9

SP - 207

EP - 216

BT - ICAIL '23: Proceedings of the Nineteenth International Conference on Artificial Intelligence and Law

A2 - Algoritmi, Centro

A2 - Reuters, Thomson

PB - Association for Computing Machinery

ER -

Model- and data-agnostic justifications with a fortiori case-based argumentation

Abstract

Bibliographical note

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this