Explainable AI (open)

Introduction

Explainable Artificial Intelligence (XAI) is a research field whose goal is to provide explanations about black-box models to help the different stakeholders (end-users, data analysts, and model developers) understand the behavior and decision-making mechanism of the algorithms. XAI techniques have been rapidly adopted by multiple research fields such as medical [1], computer vision [2], and the recommendation [3].

Goal

The independent study aims to collect and analyze the highest number of papers from different domains to generate a clear overview of the XAI space with a special focus on what visual approaches have been used to explain the generated explanation.

Contact

The independent study will be supervised by Ibrahim Al Hazwani and Prof. Jürgen Bernad. If you have any question feel free to contact one of the two mention above.

References

[1] E. Tjoa and C. Guan, "A Survey on Explainable Artificial Intelligence (XAI): Toward Medical XAI," in IEEE Transactions on Neural Networks and Learning Systems, vol. 32, no. 11, pp. 4793-4813, Nov. 2021, DOI: 10.1109/TNNLS.2020.3027314.

[2] Buhrmester, Vanessa, David Münch, and Michael Arens. "Analysis of explainers of black box deep neural networks for computer vision: A survey." Machine Learning and Knowledge Extraction 3.4 (2021): 966-989.

[3] Afchar, Darius, et al. "Explainability in Music Recommender Systems." arXiv preprint arXiv:2201.10528 (2022).

Illustration by Pablo Stanley