Explainability

Topics

• Counterfactuals
  • Counterfactuals with Constraint Solvers
• LIME and the deterministic version, DLIME
• SHAP
  • SHAP Background for cold start

Resources

• TrustyAI Explainability Toolkit11 Geada, Rob, Tommaso Teofili, Rui Vieira, Rebecca Whitworth and Daniele Zonca. “TrustyAI Explainability Toolkit.” (2021). pre-print, https://arxiv.org/abs/2104.12717
• A nice presentation on AI/ML explainability: https://explainml-tutorial.github.io/neurips20

Software

• TrustyAI22 https://github.com/trustyai-explainability ## Literature

• omnixai

• Kakogeorgiou, Ioannis, and Konstantinos Karantzalos. “Evaluating Explainable Artificial Intelligence Methods for Multi-label Deep Learning Classification Tasks in Remote Sensing.” arXiv preprint arXiv:2104.01375 (2021).