Geometry and generalization: eigenvalues as predictors of where a network will fail to generalize

DOI10.3934/FODS.2022004zbMath1496.53107arXiv2107.06386OpenAlexW3181780919MaRDI QIDQ2148963

Ben K. Dees, Corey Lowman, Susama Agarwala, Andrew Gearheart

Publication date: 24 June 2022

Published in: Foundations of Data Science (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2107.06386

zbMATH Keywords

neural networksdifferential geometrydata representation theoryeigenvalues of local Jacobians

Mathematics Subject Classification ID

Theory of computing (68Q99) Differential geometric aspects of statistical manifolds and information geometry (53B12) Applications of differential geometry to data and computer science (53Z50)

Uses Software

• GitHub
• UMAP

Cites Work

• Overparameterized neural networks implement associative memory
• An Introduction to Variational Autoencoders
• Nonlinear Dimensionality Reduction
• Matrix inequalities

This page was built for publication: Geometry and generalization: eigenvalues as predictors of where a network will fail to generalize