Speaker: Marco Scutari (https://www.bnlearn.com/about)

Title: Achieving fairness with a simple ridge penalty

Abstract: The adoption of machine learning in applications where it is crucial to ensure fairness and accountability has led to a large number of model proposals in the literature, largely formulated as optimisation problems with constraints reducing or eliminating the effect of sensitive attributes on the response. While this approach is very flexible from a theoretical perspective, the resulting models are somewhat black-box in nature: very little can be said about their statistical properties, what are the best practices in their applied use, and how they can be extended to problems other than those they were originally designed for. Furthermore, the estimation of each model requires a bespoke implementation involving an appropriate solver which is less than desirable from a software engineering perspective. In this talk, I will take the opposite view that classical statistical models can be adapted to enforce fairness while preserving their well-known properties and the associated best practices in their applied use. I discuss how combining generalised linear models (GLMs) with penalised regression works into a fair (generalised) ridge regression model works very well for this purpose. Results from real-world data show this approach has competitive predictive accuracy, simple to implement, and suitable to a wide variety of applications.

Bio: Marco Scutari is a Senior Researcher at Istituto Dalle Molle di Studi sull'Intelligenza Artificiale (IDSIA), Switzerland. He has held positions in Statistics, Statistical Genetics and Machine Learning at UCL, University of Oxford and IDSIA since completing his Ph.D. in Statistics in 2011. His research focuses on the theory of Bayesian networks and their applications to biological and clinical data, as well as statistical computing and software engineering. Recently he has worked on introducing fairness in financial applications in a 3-years-long project with UBS Switzerland.