Efficient approximations of the fisher matrix in neural networks using kronecker product singular value decomposition

DOI10.1051/proc/202373218arXiv2201.10285OpenAlexW4281255645MaRDI QIDQ6127062

Unnamed Author, Unnamed Author, Ibtihel Ben Gharbia, Mounir Haddou, Valérie Garès, Quang Huy Tran

Publication date: 10 April 2024

Published in: ESAIM: Proceedings and Surveys (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2201.10285

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Numerical computation of eigenvalues and eigenvectors of matrices (65F15)

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• On the limited memory BFGS method for large scale optimization
• The ubiquitous Kronecker product
• Numerical Methods for Large Eigenvalue Problems
• Reducing the Dimensionality of Data with Neural Networks
• Fast Curvature Matrix-Vector Products for Second-Order Gradient Descent
• Riemannian metrics for neural networks I: feedforward networks
• Calculating the Singular Values and Pseudo-Inverse of a Matrix
• A Family of Variable-Metric Methods Derived by Variational Means
• A new approach to variable metric algorithms
• The Convergence of a Class of Double-rank Minimization Algorithms 1. General Considerations
• Conditioning of Quasi-Newton Methods for Function Minimization
• A Stochastic Approximation Method

This page was built for publication: Efficient approximations of the fisher matrix in neural networks using kronecker product singular value decomposition