Gadat, Sébastien and Gavra, Ioana (2021) Asymptotic study of stochastic adaptive algorithm in non-convex landscape. TSE Working Paper, n. 21-1175

Warning
There is a more recent version of this item available.
[thumbnail of wp_tse_1175.pdf]
Preview
Text
Download (646kB) | Preview

Abstract

This paper studies some asymptotic properties of adaptive algorithms widely used in optimization and machine learning, and among them Adagrad and Rmsprop, which are involved in most of the blackbox deep learning algorithms. Our setup is the non-convex landscape optimization point of view, we consider a one time scale parametrization and we consider the situation where these algorithms may be used or not with mini-batches. We adopt the point of view of stochastic algorithms and establish the almost sure convergence of these methods when using a decreasing step-size towards the set of critical points of the target function. With a mild extra assumption on the noise, we also obtain the convergence towards the set of minimizers of the function. Along our study, we also obtain a \convergence rate" of the methods, in the vein of the works of [GL13].

Item Type: Monograph (Working Paper)
Language: English
Date: January 2021
Uncontrolled Keywords: Stochastic optimization, Stochastic adaptive algorithm, Convergence of random variables
Subjects: B- ECONOMIE ET FINANCE
Divisions: TSE-R (Toulouse)
Institution: Université Toulouse 1 Capitole
Site: UT1
Date Deposited: 29 Jan 2021 13:14
Last Modified: 13 Sep 2022 13:56
OAI Identifier: oai:tse-fr.eu:125116
URI: https://publications.ut-capitole.fr/id/eprint/42233

Available Versions of this Item

View Item

Downloads

Downloads per month over past year