Stability for the training of deep neural networks and other classifiers

DOI10.1142/S0218202521500500MaRDI QIDQ5024400

Pierre-Emmanuel Jabin, C. Alex Safsten, Leonid Berlyand

Publication date: 31 January 2022

Published in: Mathematical Models and Methods in Applied Sciences (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2002.04122

zbMATH Keywords

stabilityentropy methodsdeep learning

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Stability and convergence of numerical methods for initial value and initial-boundary value problems involving PDEs (65M12)

Uses Software

• ImageNet
• AlexNet

Cites Work

• Nonparametric modal regression
• Subgeometric rates of convergence in Wasserstein distance for Markov chains
• Asymptotic coupling and a general form of Harris' theorem with applications to stochastic delay equations
• On the convergence of formally diverging neural net-based classifiers
• Analysis of time-frequency scattering transforms
• Subgeometric rates of convergence of Markov processes in the Wasserstein metric
• Group Invariant Scattering
• Deep Haar scattering networks
• Lipschitz properties for deep convolutional networks
• Spiking Neuron Models
• Soft and hard classification by reproducing kernel Hilbert space methods

This page was built for publication: Stability for the training of deep neural networks and other classifiers