When to choose ranked area integrals versus integrated gradient for explainable artificial intelligence – a comparison of algorithms
Benchmarking: An International Journal
ISSN: 1463-5771
Article publication date: 9 August 2022
Issue publication date: 1 December 2023
Abstract
Purpose
Explainable artificial intelligence (XAI) has importance in several industrial applications. The study aims to provide a comparison of two important methods used for explainable AI algorithms.
Design/methodology/approach
In this study multiple criteria has been used to compare between explainable Ranked Area Integrals (xRAI) and integrated gradient (IG) methods for the explainability of AI algorithms, based on a multimethod phase-wise analysis research design.
Findings
The theoretical part includes the comparison of frameworks of two methods. In contrast, the methods have been compared across five dimensions like functional, operational, usability, safety and validation, from a practical point of view.
Research limitations/implications
A comparison has been made by combining criteria from theoretical and practical points of view, which demonstrates tradeoffs in terms of choices for the user.
Originality/value
Our results show that the xRAI method performs better from a theoretical point of view. However, the IG method shows a good result with both model accuracy and prediction quality.
Keywords
Acknowledgements
This project is an outcome of research collaboration between BASF SE, Germany and the Indian Institute of Technology Delhi, India and, is supported financially by BASF SE for the purpose of extending knowledge in the space of advanced AI models.
Citation
Singh, V., Konovalova, I. and Kar, A.K. (2023), "When to choose ranked area integrals versus integrated gradient for explainable artificial intelligence – a comparison of algorithms", Benchmarking: An International Journal, Vol. 30 No. 9, pp. 3067-3089. https://doi.org/10.1108/BIJ-02-2022-0112
Publisher
:Emerald Publishing Limited
Copyright © 2022, Emerald Publishing Limited