Dipartimento di Matematica

Seminario / Workshop
Image
globo con formule matematiche

Machine Learning through the lens of sub-Riemannian Geometry

14 Ottobre 2025 , ore 15:00 - 16:00
PovoZero, Via Sommarive 14, Povo (Trento)
Aula Seminari "1"
Ingresso libero
Organizzato da: Dipartimento di Matematica
Destinatari: Comunità universitaria
Referente: Andrea Pinamonti e Gian Paolo Leonardi
Contatti: 
Staff del Dipartimento di Matematica
Image
globo con formule matematiche

Ciclo di seminari del Dipartimento di Matematica organizzato da Gian Paolo Leonardi in collaborazione con: Claudio Agostinelli, Fabio Bagagiolo, Luigi Amedeo Bianchi, Stefano Bonaccorsi, Michele Coghi, Alessandro Oneto, Riccardo Ghiloni, Veronica Vinciotti.

Speaker: Alessandro Scagliotti (Technische Universität München)

In this talk we illustrate the link between Deep Neural Networks and flows induced by control systems (Neural ODEs), and we relate the "expressivity'' of a Residual Neural Network (ResNets) to the controllability properties of the corresponding Neural ODE in the space of diffeomorphisms. In case of control-linear Neural ODEs, a sub-Riemannian structure emerges. We show how the Lie Algebra Strong Approximating Property (see [Agrachev & Sarychev 2020,2022]) guarantees that, given two M-tuples of pairwise distinct points (M>1), we can steer one to the other. Moreover, this condition implies that we can approximate on compact sets any diffeomorphism isotopic to the identity using flows induced by the controlled dynamics.
We then formulate the (ensemble) optimal control problem related to the diffeomorphism approximation task, and we study its limiting behaviour when the size of the data-set tends to infinity. Finally, we show how this machinery can be used for the numerical construction of the optimal transport map.