Deep regularization and direct training of the inner layers of neural networks with kernel flows

DOI10.1016/j.physd.2021.132952OpenAlexW3007786121MaRDI QIDQ2077570

Gene Ryan Yoo, Houman Owhadi

Publication date: 21 February 2022

Published in: Physica D (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2002.08335

zbMATH Keywords

artificial neural networksmachine learningimage classificationGaussian process regressioninner layer trainingkernel flows

Mathematics Subject Classification ID

Artificial intelligence (68Txx)

Related Items (4)

Computational graph completion ⋮ Learning dynamical systems from data: a simple cross-validation perspective. III: Irregularly-sampled time series ⋮ Learning ``best kernels from data in Gaussian process regression. With application to aerodynamics ⋮ Do ideas have shape? Idea registration as the continuous limit of artificial neural networks

Uses Software

• TensorFlow
• GitHub
• BranchyNet

Cites Work

• Unnamed Item
• A comparison of generalized cross validation and modified maximum likelihood for estimating the parameters of a stochastic process
• Cross validation and maximum likelihood estimations of hyper-parameters of Gaussian processes with model misspecification
• Kernel flows: from learning kernels from data into the abyss
• Bayesian Inference for Non-Stationary Spatial Covariance Structure via Spatial Deformations
• Can a finite element method perform arbitrarily badly?
• Operator-Adapted Wavelets, Fast Solvers, and Numerical Homogenization
• Prevalence of neural collapse during the terminal phase of deep learning training
• Statistical Numerical Approximation
• Bayesian Numerical Homogenization

This page was built for publication: Deep regularization and direct training of the inner layers of neural networks with kernel flows