Picture for Shun-ichi Amari

Shun-ichi Amari

Information Geometry of Wasserstein Statistics on Shapes and Affine Deformations

Add code
Jul 24, 2023
Viaarxiv icon

Deep Learning in Random Neural Fields: Numerical Experiments via Neural Tangent Kernel

Add code
Feb 10, 2022
Figure 1 for Deep Learning in Random Neural Fields: Numerical Experiments via Neural Tangent Kernel
Figure 2 for Deep Learning in Random Neural Fields: Numerical Experiments via Neural Tangent Kernel
Figure 3 for Deep Learning in Random Neural Fields: Numerical Experiments via Neural Tangent Kernel
Figure 4 for Deep Learning in Random Neural Fields: Numerical Experiments via Neural Tangent Kernel
Viaarxiv icon

When Does Preconditioning Help or Hurt Generalization?

Add code
Jul 02, 2020
Figure 1 for When Does Preconditioning Help or Hurt Generalization?
Figure 2 for When Does Preconditioning Help or Hurt Generalization?
Figure 3 for When Does Preconditioning Help or Hurt Generalization?
Figure 4 for When Does Preconditioning Help or Hurt Generalization?
Viaarxiv icon

Any Target Function Exists in a Neighborhood of Any Sufficiently Wide Random Network: A Geometrical Perspective

Add code
Jan 20, 2020
Figure 1 for Any Target Function Exists in a Neighborhood of Any Sufficiently Wide Random Network: A Geometrical Perspective
Figure 2 for Any Target Function Exists in a Neighborhood of Any Sufficiently Wide Random Network: A Geometrical Perspective
Figure 3 for Any Target Function Exists in a Neighborhood of Any Sufficiently Wide Random Network: A Geometrical Perspective
Viaarxiv icon

Pathological spectra of the Fisher information metric and its variants in deep neural networks

Add code
Oct 14, 2019
Figure 1 for Pathological spectra of the Fisher information metric and its variants in deep neural networks
Figure 2 for Pathological spectra of the Fisher information metric and its variants in deep neural networks
Figure 3 for Pathological spectra of the Fisher information metric and its variants in deep neural networks
Figure 4 for Pathological spectra of the Fisher information metric and its variants in deep neural networks
Viaarxiv icon

The Normalization Method for Alleviating Pathological Sharpness in Wide Neural Networks

Add code
Jun 07, 2019
Figure 1 for The Normalization Method for Alleviating Pathological Sharpness in Wide Neural Networks
Viaarxiv icon

Fisher Information and Natural Gradient Learning of Random Deep Networks

Add code
Aug 22, 2018
Figure 1 for Fisher Information and Natural Gradient Learning of Random Deep Networks
Figure 2 for Fisher Information and Natural Gradient Learning of Random Deep Networks
Viaarxiv icon

Statistical Neurodynamics of Deep Networks: Geometry of Signal Spaces

Add code
Aug 22, 2018
Figure 1 for Statistical Neurodynamics of Deep Networks: Geometry of Signal Spaces
Figure 2 for Statistical Neurodynamics of Deep Networks: Geometry of Signal Spaces
Figure 3 for Statistical Neurodynamics of Deep Networks: Geometry of Signal Spaces
Figure 4 for Statistical Neurodynamics of Deep Networks: Geometry of Signal Spaces
Viaarxiv icon

Universal Statistics of Fisher Information in Deep Neural Networks: Mean Field Approach

Add code
Jun 04, 2018
Figure 1 for Universal Statistics of Fisher Information in Deep Neural Networks: Mean Field Approach
Figure 2 for Universal Statistics of Fisher Information in Deep Neural Networks: Mean Field Approach
Viaarxiv icon

Bayesian Robust Tensor Factorization for Incomplete Multiway Data

Add code
Apr 16, 2015
Figure 1 for Bayesian Robust Tensor Factorization for Incomplete Multiway Data
Figure 2 for Bayesian Robust Tensor Factorization for Incomplete Multiway Data
Figure 3 for Bayesian Robust Tensor Factorization for Incomplete Multiway Data
Figure 4 for Bayesian Robust Tensor Factorization for Incomplete Multiway Data
Viaarxiv icon