Gradient descent provably escapes saddle points in the training of shallow ReLU networks

DOI10.1007/s10957-024-02513-3MaRDI QIDQ6655804

Florian Rossmannek, Patrick Cheridito, Arnulf Jentzen

Publication date: 27 December 2024

Published in: Journal of Optimization Theory and Applications (Search for Journal in Brave)

zbMATH Keywords

neural networksnonconvex optimizationgradient descentcenter-stable manifolds

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Nonconvex programming, global optimization (90C26) Numerical optimization and variational techniques (65K10)

This page was built for publication: Gradient descent provably escapes saddle points in the training of shallow ReLU networks