Variational Analysis and Optimisation Webinars, http://www.mocao.org/va-webinar/

Title: Adaptive gradient descent without descent

Speaker: Yura Malitsky (Linköping University)

Abstract: In this talk I will present some recent results for the most classical optimization method — gradient descent. We will show that a simple zero cost rule is sufficient to completely automate gradient descent. The method adapts to the local geometry, with convergence guarantees depending only on the smoothness in a neighborhood of a solution. The presentation is based on a joint work with K. Mishchenko, see arxiv.org/abs/1910.09529.

Author:	Vera Roshchina (via theBox)
Created:	19/05/2021
Updated:	19/05/2021
Theme:	Default
Copyright:	Creative Commons

Info