Deep Linear Network Training Dynamics from Random Initialization: Data, Width, Depth, and Hyperparameter Transfer

Blake Bordelon; Cengiz Pehlevan

2025 ICML ICML 2025

Deep Linear Network Training Dynamics from Random Initialization: Data, Width, Depth, and Hyperparameter Transfer

Abstract

No abstract available. No abstract was captured for this paper when its proceedings were scraped - most often an older edition published without one online. How the corpus is built.

Authors

Blake Bordelon, Cengiz Pehlevan

Topics

Machine Learning > Optimization & Theory > Neural Network Optimization Deep Learning > Architectures > Neural Networks

Keywords

hyperparameter transfer gradient descent residual network deep linear network random initialization wide

Download PDF