Some Thoughts on the Use of Causal Modelling in Algorithmic Fairness

Using data-driven and automated approaches to aid decision making has the merits of scaling services up and allowing for standardized policies to treat people on a seemingly objective way. However, it is known that algorithms are often not capable to be fair, according to a variety of value judgements, due to reasons such as strong biases in the datasets used to build such algorithms. This means some demographic groups are put at disadvantage, now at an automated scale that can bring new potentially harmful consequences. The notion of algorithmic fairness is however not straightforward, as there are several roles algorithms play (from passive information retrieval to high-stakes resource allocation problems) and several aspects of what fairness means. In this talk, I will discuss how causal reasoning and inference can help some aspects of algorithmic fairness. Causality plays a role via the formulation of what-if questions that illuminate how an individual’s history could have been different given alternative exposures, and how hypothetical policies not previously considered in the data could change the balance of outcomes across different demographic groups.

Please join the event.

About Ricardo Silva

Ricardo Silva is a Professor of Statistical Machine Learning and Data Science at the Department of Statistical Science, UCL, a Faculty Fellow at the Alan Turing Institute, and a recipient of a EPSRC Open Fellowship (2023-2027). Ricardo obtained a PhD in Machine Learning from Carnegie Mellon University, 2005, followed by postdoctoral positions at the Gatsby Computational Neuroscience Unit (UCL) and at the Statistical Laboratory (University of Cambridge). His main interests are on causal inference, latent variable models, and probabilistic machine learning. His work has received funding from organisations such as the UK Engineering and Physical Sciences Research Council, Innovate UK, the Office of Naval Research, Winton Research and Adobe Research, among others. He is currently Deputy Head of Department and a co-investigator in the UK AI Hub on Causality in Healthcare AI with Real Data.