Description
Date depot: 6 décembre 2022
Titre: Defining Differential Explanations: Understanding the Dynamic of Changes in Machine Learning Models
Directeur de thèse:
Christophe MARSALA (LIP6)
Directrice de thèse:
Marie-Jeanne LESOT (LIP6)
Domaine scientifique: Sciences et technologies de l'information et de la communication
Thématique CNRS : Intelligence artificielle
Resumé: Instead of explaining a static situation (a trained model on a fixed dataset), this PhD aims to explore the novel concept of differential explainability, ie. explaining the differences or the evolution between successive versions of a ML model devoted to the same task (e.g. classification with fraud detection, regression with GDP prediction).
Doctorant.e: Rida Adam