Operational Research for Fairness, Privacy and Interpretability in Machine Learning

DS4DM Coffee Talk
Operational Research for Fairness, Privacy and Interpretability in Machine Learning: Leveraging ILP to Learn Optimal Fair Rule Lists
Julien Ferry, LAAS-CNRS (France)
July 7, 2022

Fairness, interpretability, and privacy are important fields for the development of responsible AI. While these three topics are often considered separately, their simultaneous application also seems desirable. The objective of my PhD is to study their interactions leveraging tools from operations research and combinatorial optimization. In a first part of my talk, I will provide an overview of my different past and current works on the frontiers of these different topics. In a second part, I will present my work on the use of Integer Linear Programming to learn interpretable, fair and optimal models.