Conservative SPDEs as fluctuating mean field limits of stochastic gradient descent

Authors Benjamn Gess, Rishabh S. Gvalani, Vitalii Konarovskyi

Publication date 12 July 2022

Stochastic partial differential equations (aspects of stochastic analysis) (60H15) Artificial neural networks and deep learning (68T07) Central limit and other weak theorems (60F05) Random measures (60G57) Martingales and classical analysis (60G46)

Abstract: The convergence of stochastic interacting particle systems in the mean-field limit to solutions of conservative stochastic partial differential equations is established, with optimal rate of convergence. As a second main result, a quantitative central limit theorem for such SPDEs is derived, again, with optimal rate of convergence. The results apply, in particular, to the convergence in the mean-field scaling of stochastic gradient descent dynamics in overparametrized, shallow neural networks to solutions of SPDEs. It is shown that the inclusion of fluctuations in the limiting SPDE improves the rate of convergence, and retains information about the fluctuations of stochastic gradient descent in the continuum limit.

This page was built for publication: Conservative SPDEs as fluctuating mean field limits of stochastic gradient descent

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6404754)