[Archimedes Seminar Series] Fairness in machine learning: a study of the Demographic Parity constraint

Title: Fairness in machine learning: a study of the Demographic Parity constraint 

Presenter: Dr Nicolas Schreuder (Centre national de la recherche scientifique (CNRS) )

Abstract: In various domains, statistical algorithms trained on personal data take pivotal decisions which influence our lives on a daily basis. Recent studies show that a naive use of these algorithms in sensitive domains may lead to unfair and discriminating decisions, often inheriting or even amplifying biases present in data. In the first part of the talk, I will introduce and discuss the question of fairness in machine learning through concrete examples of biases coming from the data and/or from the algorithms. In a second part, I will demonstrate how statistical learning theory can help us better understand and overcome some of those biases. In particular, I will present a selection of recent results from two of my papers on the Demographic Parity constraint, a popular fairness constraint. In particular I will describe an interesting link between this constraint and optimal transport theory.
References: 
- A minimax framework for quantifying risk-fairness trade-off in regression (with E. Chzhen), Ann. Statist. 50(4): 2416-2442 (Aug. 2022). DOI: 10.1214/22-AOS2198;
- Fair learning with Wasserstein barycenters for non-decomposable performance measures (with S. Gaucher and E. Chzhen), AISTATS 2023.

Bio: Nicolas Schreuder is a researcher ("Chargé de Recherche") at CNRS (France). He is affiliated to the Laboratoire d'Informatique Gaspard Monge in Paris area. He received his PhD from Institut Polytechnique de Paris in 2021. He was supervised by Arnak Dalalyan and Victor-Emmanuel Brunel at the Center for Research in Economics and Statistics (CREST). Before joining CNRS, he did a postdoc with Lorenzo Rosasco at MaLGa (Università di Genova). Nicolas's research focuses on developing statistical learning theory in order to incorporate societal considerations such as fairness and efficiency.