Abstract
In the past decade, Explainable Artificial Intelligence (XAI) has attracted a great interest in the research community, motivated by the need for explanations in critical AI applications. Some recent advances in XAI are based on Evolutionary Computation (EC) techniques, such as Genetic Programming. We call this trend EC for XAI. We argue that the full potential of EC methods has not been fully exploited yet in XAI, and call the community for future efforts in this field. Likewise, we find that there is a growing concern in EC regarding the explanation of population-based methods, i.e., their search process and outcomes. While some attempts have been done in this direction (although, in most cases, those are not explicitly put in the context of XAI), we believe that there are still several research opportunities and open research questions that, in principle, may promote a safer and broader adoption of EC in real-world applications. We call this trend XAI within EC. In this position paper, we briefly overview the main results in the two above trends, and suggest that the EC community may play a major role in the achievement of XAI.
DOI
10.1145/3520304.3533974
Publication Date
2022-07-09
Event
Genetic and Evolutionary Computation Conference (GECCO 2022)
Publication Title
GECCO '22: Proceedings of the Genetic and Evolutionary Computation Conference Companion
Publisher
ACM
ISBN
978-1-4503-9268-6
Embargo Period
2024-11-22
First Page
1757
Last Page
1762
Recommended Citation
Bacardit, J., Brownlee, A., Cagnoni, S., & et al. (2022) 'The intersection of Evolutionary Computation and Explainable AI', GECCO '22: Proceedings of the Genetic and Evolutionary Computation Conference Companion, , pp. 1757-1762. ACM: Available at: https://doi.org/10.1145/3520304.3533974
