The LMS and ICMS jointly hosted the Symposium: Analytic and Geometric Approaches to Machine Learning workshop. 

At the time of this workshop, machine learning has been remarkably successful in it’s applications (e.g. classification/clustering, regression, data mining and prediction) but our theoretical understanding of many machine learning algorithms is still missing. This has led to an increasing appetite for the mathematical analysis of machine learning algorithms. Particularly exciting is the potential for methods from applied mathematics, probability theory, and statistics to contribute to machine learning theory.
The aim of the workshop was to bring together researchers that apply mathematical methodology to machine learning. There was a particular emphasis on how mathematical theory can inform applications and vice versa.

This virtual workshop was the first of two workshops on this topic. The second was an in-person workshop held at the University of Bath, in December 2021. In this first workshop, invited speakers were encouraged to present open problems and explore interesting directions for potential research as part of their talk. The schedule allowed participants time to initiate conversations and collaborations that could be developed at the winter workshop.