Get our free extension to see links to code for papers anywhere online!

 Add to Chrome

 Add to Firefox

CatalyzeX Code Finder - Browser extension linking code for ML papers across the web! | Product Hunt Embed
Data augmentation as stochastic optimization

Oct 21, 2020
Boris Hanin, Yi Sun

* 25 pages 

  Access Paper or Ask Questions

Finite Depth and Width Corrections to the Neural Tangent Kernel

Sep 13, 2019
Boris Hanin, Mihai Nica

* 27 pages, 2 figures, comments welcome 

  Access Paper or Ask Questions

Deep ReLU Networks Have Surprisingly Few Activation Patterns

Jun 03, 2019
Boris Hanin, David Rolnick

* 18 page, 7 figures 

  Access Paper or Ask Questions

Complexity of Linear Regions in Deep Networks

Jan 25, 2019
Boris Hanin, David Rolnick


  Access Paper or Ask Questions

Products of Many Large Random Matrices and Gradients in Deep Neural Networks

Dec 14, 2018
Boris Hanin, Mihai Nica

* v1. 26p. Comments Welcome 

  Access Paper or Ask Questions

Which Neural Net Architectures Give Rise To Exploding and Vanishing Gradients?

Oct 27, 2018
Boris Hanin

* v3. 18p. 1 fig. Accepted at NIPS 2018 

  Access Paper or Ask Questions

How to Start Training: The Effect of Initialization and Architecture

Jun 19, 2018
Boris Hanin, David Rolnick

* v2. 15p. Comments welcome 

  Access Paper or Ask Questions

Approximating Continuous Functions by ReLU Nets of Minimal Width

Mar 10, 2018
Boris Hanin, Mark Sellke

* v2. 13p. Extended main result to higher dimensional output. Comments welcome 

  Access Paper or Ask Questions

Universal Function Approximation by Deep Neural Nets with Bounded Width and ReLU Activations

Dec 20, 2017
Boris Hanin

* v3. Theorem 3 removed. Comments Welcome. 9p 

  Access Paper or Ask Questions