Function space and critical points of linear convolutional networks

DOI10.1137/23M1565504MaRDI QIDQ6562372zbMATH OpenFDO

Authors Kathlén Kohn, Guido Montúfar, Vahid Shahverdi, Matthew Trager

Publication date 26 June 2024

Published in SIAM Journal on Applied Algebra and Geometry (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1137/23M1565504

zbMATH Keywords

critical pointsneural networksemialgebraic set

Mathematics Subject Classification ID

Singularities in algebraic geometry (14B05) Birational geometry (14E99) Surfaces and higher-dimensional varieties (14J99) Projective techniques in algebraic geometry (14N05) Semialgebraic sets and related spaces (14P10) Artificial neural networks and deep learning (68T07) Polynomial optimization (90C23)

Recommendations

• Geometry of linear convolutional networks
• Optimization Landscape of Neural Networks
• Approximation Analysis of Convolutional Neural Networks
• Theory of deep convolutional neural networks. II: Spherical analysis
• Convergence of deep convolutional neural networks

Cites work

• Basic Algebraic Geometry 2
• Geometry of linear convolutional networks
• Learning deep linear neural networks: Riemannian gradient flows and convergence to global minimizers
• Secants, bitangents, and their congruences
• The global optimization geometry of shallow linear neural networks

Cited in

(4)

• Activation degree thresholds and expressiveness of polynomial neural networks
• Geometry of linear neural networks: equivariance and invariance under permutation groups
• The geometry of the deep linear network
• Geometry and optimization of shallow polynomial networks

This page was built for publication: Function space and critical points of linear convolutional networks

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6562372)