Non-asymptotic estimates for TUSLA algorithm for non-convex learning with applications to neural networks with ReLU activation function

Non-asymptotic estimates for TUSLA algorithm for non-convex learning with applications to neural networks with ReLU activation function

We consider nonconvex stochastic optimization problems where the objective functions have super-linearly growing and discontinuous stochastic gradients. In such a setting, we provide a nonasymptotic analysis for the tamed unadjusted stochastic Langevin algorithm (TUSLA) introduced in Lovas et al. (2...

Full description

Bibliographic Details
Main Authors: Lim, Dong-Young, Neufeld, Ariel, Sabanis, Sotirios, Zhang, Ying
Other Authors: School of Physical and Mathematical Sciences
Format: Journal Article
Language:English
Published: 2024
Subjects:
Mathematical Sciences
Nonconvex optimization
Nonasymptotic estimates
Online Access:https://hdl.handle.net/10356/179411

Internet

https://hdl.handle.net/10356/179411

Similar Items