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

Chrome logo  Add to Chrome

Firefox logo Add to Firefox

Sparsity by Redundancy: Solving $L_1$ with a Simple Reparametrization


Oct 03, 2022
Liu Ziyin, Zihao Wang

Add code

* preprint 

   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

What shapes the loss landscape of self-supervised learning?


Oct 02, 2022
Liu Ziyin, Ekdeep Singh Lubana, Masahito Ueda, Hidenori Tanaka

Add code

* preprint 

   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

Exact Phase Transitions in Deep Learning


May 25, 2022
Liu Ziyin, Masahito Ueda

Add code

* preprint 

   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

Posterior Collapse of a Linear Latent Variable Model


May 09, 2022
Zihao Wang, Liu Ziyin

Add code

* 17 pages, 5 figures 

   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

Exact Solutions of a Deep Linear Network


Feb 14, 2022
Liu Ziyin, Botao Li, Xiangming Meng

Add code

* preprint 

   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

Stochastic Neural Networks with Infinite Width are Deterministic


Jan 30, 2022
Liu Ziyin, Hanlin Zhang, Xiangming Meng, Yuting Lu, Eric Xing, Masahito Ueda

Add code


   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

SGD May Never Escape Saddle Points


Jul 25, 2021
Liu Ziyin, Botao Li, Masahito Ueda

Add code


   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

What Data Augmentation Do We Need for Deep-Learning-Based Finance?


Jun 08, 2021
Liu Ziyin, Kentaro Minami, Kentaro Imajo

Add code

* Preprint 

   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

Logarithmic landscape and power-law escape rate of SGD


May 20, 2021
Takashi Mori, Liu Ziyin, Kangqiao Liu, Masahito Ueda

Add code

* 15+6 pages 

   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

On the Distributional Properties of Adaptive Gradients


May 15, 2021
Zhang Zhiyi, Liu Ziyin

Add code


   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email
1
2
3
>>