Archive / INF Seminars / INF_2024_03_13_Scutari_Marco
USI - Email

Causality: Beyond Predictive Modelling


Host: Prof. Ernst-Jan Camiel Wit




USI Campus Est, room D1.14, Sector D
16:15 - 17:00

Marco Scutari
Istituto Dalle Molle di Studi sull’Intelligenza Artificiale (IDSIA), Switzerland
Machine learning is changing science and our society thanks to our ability to learn models that can capture information effectively. However, scientific questions are inherently causal. Causation is central to how we think as human beings.
Machine learning focuses on predictive modelling, which is not equipped to understand the causality that drives reality. On the other hand, network models are ideal for this task because they can rigorously express causal relationships. In this talk, I will provide an overview of the fundamental problems we face in the statistical learning of causal network models from complex data, their successful applications and the challenges we are overcoming to use them to better understand the world we live in.

Marco Scutari is a Senior Researcher at Istituto Dalle Molle di Studi sull'Intelligenza Artificiale (IDSIA), Switzerland. He completed his PhD in Statistics from the University of Padova and has held positions in statistics, statistical genetics and machine learning at University College London and the University of Oxford. He was awarded the Excellence Recognition Scheme by the University of Oxford in 2017 and has been ranked the "World's Top 2% Scientist" by Stanford University since 2022. His research focuses on the theory of Bayesian and causal networks and their applications to biological and clinical data, for which he is also a long-term consultant at Pfizer and LaRoche-Posay, as well as statistical computing and software engineering. He is the author of the bnlearn package, the most widely used free software package for Bayesian networks, and is the main author of the best-selling book "Bayesian Networks: with Examples in R."

Host: Prof. Ernst-Jan Camiel Wit