Â鶹Éçmadou

About the lecture

Machine learning methods perform well in prediction tasks within the range of the training data.

These methods often fail when interest is in extrapolation, that is, (1) prediction in areas of the predictor space with few or no training observations; or (2) prediction of quantiles of the response that go beyond the observed records. 

Extreme value theory provides the mathematical foundation for extrapolation beyond the range of the training data, both in the dimension of the predictor space and the response variable. In this talk we present recent methodology that combines this extrapolation theory with flexible machine learning methods to tackle the out-of-distribution generalization problem (1) and the extreme quantile regression problem (2). We highlight the extrapolation problem using AI-driven weather forecasting models such as GraphCast and Pangu-Weather. While these models outperform classical physical models in terms of overall mean squared error, they fall short when it comes to accurately forecasting extreme events.

About the speaker

Sebastian Engelke is Associate Professor of statistics at the University of Geneva, where is holding an SNSF Eccellenza grant. He was a visiting professor at University of Toronto from 2018 – 2019. Previously he was an Ambizione fellow at EPF Lausanne with Anthony Davison. Sebastian did his studies in Mathematics at University of Göttingen and UC Berkeley, and he finished his PhD in 2013 at the University of Göttingen with Martin Schlather. His research interests are in extreme value theory, graphical models, spatial statistics and machine learning. Sebastian is Associate Editor of JRSSB and the journal Extremes.

Schedule

2-3pm: Presentation by A/Prof. Sebastian Engelke in Room 4082, Anita B. Lawrence Centre, Â鶹Éçmadou.
3-4pm: Light refreshments.

Venue

The lecture will be presented in Room 4082/3 on level 4 of the Anita B. Lawrence Centre at Â鶹Éçmadou (ref: ). Light refreshments will be served in Room 3082 following the talk.

Frontiers in Fundamental Mathematics Research Nexus

This event is presented by the Â鶹Éçmadou School of Mathematics and Statistics and is part of our Frontiers in Fundamental Mathematics Research Nexus series, which aims to highlight fundamental research in the mathematical sciences, with an emphasis on the significance and impact of fundamental mathematics to a diverse range of areas within mathematics and beyond.
Learn more about the Nexus Program.