From Fitness Landscapes to Explainable AI and Back

Thomson, Sarah L.; Adair, Jason; Brownlee, Alexander E. I.; van den Berg, Daan

Abstract

We consider and discuss the ways in which search landscapes might contribute to the future of explainable artificial intelligence (XAI), and vice versa. Landscapes are typically used to gain insight into algorithm search dynamics on optimisation problems; as such, it could be said that they explain algorithms and that they are a natural bridge between XAI and evolutionary computation. Despite this, there is very little existing literature which utilises landscapes for XAI, or which applies XAI techniques to landscape analysis. This position paper reviews the existing works, discusses possible future avenues, and advocates for increased research effort in this area.

• Presentation Conference Type: Conference Paper (Published)
• Conference Name: GECCO '23
• Start Date: Jul 15, 2023
• End Date: Jul 19, 2023
• Acceptance Date: May 1, 2023
• Online Publication Date: Jul 24, 2023
• Publication Date: 2023-07
• Deposit Date: Aug 16, 2023
• Publicly Available Date: Aug 17, 2023
• Publisher: Association for Computing Machinery (ACM)
• Pages: 1663-1667
• Book Title: GECCO '23 Companion: Proceedings of the Companion Conference on Genetic and Evolutionary Computation
• ISBN: 9798400701207
• DOI: https://doi.org/10.1145/3583133.3596395
• Keywords: Fitness Landscapes, Search Landscapes, Neural Networks, Explainable AI, XAI