Deep Linear Networks for Matrix Completion—an Infinite Depth Limit

DOI10.1137/22m1530653zbMath1524.68313arXiv2210.12497MaRDI QIDQ6084954

Nadav Cohen, Unnamed Author, Govind K. Menon

Publication date: 2 December 2023

Published in: SIAM Journal on Applied Dynamical Systems (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2210.12497

zbMATH Keywords

matrix completiongeneralizabilityimplicit regularizationRiemannian gradient flowdeep linear network

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Applications of mathematical programming (90C90) Manifolds of metrics (especially Riemannian) (58D17) Matrix completion problems (15A83)

Cites Work

• Numerical computation of an analytic singular value decomposition of a matrix valued function
• Products of many large random matrices and gradients in deep neural networks
• Exact matrix completion via convex optimization
• The Nonlinear Geometry of Linear Programming. I Affine and Projective Scaling Trajectories
• The Nonlinear Geometry of Linear Programming. II Legendre Transform Coordinates and Central Trajectories
• Learning deep linear neural networks: Riemannian gradient flows and convergence to global minimizers
• Canonical Barriers on Convex Cones
• Unnamed Item
• Unnamed Item