This introductory school, as well as the associated workshop that will be held in 2022, is organized within the framework of the Dipartimento di Eccellenza "Classe di Scienze” of the Scuola Normale Superiore.
The spirit of these two activities is related to that of the school: Mathematical and Computational Aspects of Machine Learning organized at Centro De Giorgi in October 2019.
The plan of both the school and the workshop is to bring together Machine Learning researchers working on theoretical questions at the interface between Mathematics and Computer science.
The speakers will introduce the mathematical foundations of Machine Learning, discuss the current state of the art, but also highlight future research directions and open problems.
The lectures will target a broad audience of early stage mathematicians (Ph.D and advanced Master level).
In particular, the introductory week will focus on master level courses introducing the mathematical basis of machine learning.
The prerequisites will be master level courses in probability, calculus and linear algebra.
The boot-camp week, as well as the workshop, is to focus on the following topics:
1) Statistical Learning Theory and Regularization (L. Rosasco)
2) Mathematics of Neural Networks (L. Chizat and A. Montanari)
3) Optimization for Machine Learning (S. Villa)
4) Optimal Transport and Machine Learning (S. Di Marino, A. Gerolin and P. Rigollet)
5) Applications in Medical Sciences (Jean-Philippe Vert)
In principle, the School will take place at the Centro De Giorgi and the number of participants will be constrained to the capacity limit for this type of event.
In any case, the organizers will evaluate the necessity to move it to a virtual format in April.
The deadline for application is 10th May 2021. For the application procedure, please go to Registration
The confirmation of admission for the school as well as the format (presential/virtual) will be communicated to registered participants in the first half of May.
The information will also be updated on this website.