Si segnala il seguente seminario a tutti gli interessati.
Giovedì 10 Ottobre 2024, ore 14:30.
Aula Seminari III Piano, Dipartimento di Matematica, Politecnico di Milano.
Speaker: Christian Hirsch, Aarhus University.
Title: Large deviations of one-hidden-layer neural networks.
Abstract: In this talk, I will present large deviations in the context of stochastic gradient descent for one-hidden-layer neural networks with quadratic loss. I will explain how to derive quenched and annealed large deviation principle for the empirical weight evolution during training when letting the number of neurons and the number of training iterations simultaneously tend to infinity. The weight evolution is treated as an interacting dynamic particle system. The distinctive aspect compared to prior work on interacting particle systems lies in the discrete particle updates, simultaneously with a growing number of particles. This talk is based on joint work with Daniel Willhalm.
Link Zoom: http://polimi-it.zoom.us/j/94501257503 Join our Cloud HD Video Meetinghttp://polimi-it.zoom.us/j/94501257503 polimi-it.zoom.ushttp://polimi-it.zoom.us/j/94501257503 [zoom.ico]http://polimi-it.zoom.us/j/94501257503
Link Seminario Polimi: https://www.mate.polimi.it/eventi/?id=2487 Eventi Dipartimento di Matematica - Politecnico di Milanohttps://www.mate.polimi.it/eventi/?id=2487 mate.polimi.ithttps://www.mate.polimi.it/eventi/?id=2487 [apple-touch-icon-152x152-precomposed.png]https://www.mate.polimi.it/eventi/?id=2487
----- Prof. Luca Scarpa, PhD Associate Professor in Probability Department of Mathematics Politecnico di Milano Via E. Bonardi 9 20133 Milano, Italy e.mail: luca.scarpa@polimi.itmailto:luca.scarpa@polimi.it url: https://sites.google.com/view/lucascarpa