Intégration de connaissances dans les méthodes d’explications post-hoc

In the field of explainable artificial intelligence (XAI), post-hoc interpretability methods integrate user knowledge to improve the explanation understandability and allow for personalised explanations. In this paper, we propose to define a cost function that explicitly integrates such user knowledge into the interpretability objectives : we present a general framework for the optimization problem of post-hoc interpretability methods, and show that user knowledge can be integrated to any method by adding a compatibility term in the cost function. We instantiate the proposed formalization in the case of counterfactual explanations and propose a new interpretability method called Knowledge Integration in Counterfactual Explanation (KICE).