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

Chrome logo  Add to Chrome

Firefox logo Add to Firefox

Learning to Control Rapidly Changing Synaptic Connections: An Alternative Type of Memory in Sequence Processing Artificial Neural Networks


Nov 17, 2022
Kazuki Irie, J├╝rgen Schmidhuber

Add code

* Presented at NeurIPS 2022 Workshop on Memory in Artificial and Real Intelligence 

   Access Paper or Ask Questions

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

CTL++: Evaluating Generalization on Never-Seen Compositional Patterns of Known Functions, and Compatibility of Neural Representations


Oct 12, 2022
Róbert Csordás, Kazuki Irie, Jürgen Schmidhuber

Add code

* Accepted to EMNLP 2022 

   Access Paper or Ask Questions

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

Images as Weight Matrices: Sequential Image Generation Through Synaptic Learning Rules


Oct 07, 2022
Kazuki Irie, J├╝rgen Schmidhuber

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

Neural Differential Equations for Learning to Program Neural Nets Through Continuous Learning Rules


Jun 03, 2022
Kazuki Irie, Francesco Faccio, J├╝rgen Schmidhuber

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

Unsupervised Learning of Temporal Abstractions with Slot-based Transformers


Mar 25, 2022
Anand Gopalakrishnan, Kazuki Irie, J├╝rgen Schmidhuber, Sjoerd van Steenkiste

Add code

* 26 pages, 8 figures 

   Access Paper or Ask Questions

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

The Dual Form of Neural Networks Revisited: Connecting Test Time Predictions to Training Patterns via Spotlights of Attention


Feb 11, 2022
Kazuki Irie, Róbert Csordás, Jürgen Schmidhuber

Add code

* Two first authors 

   Access Paper or Ask Questions

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

A Modern Self-Referential Weight Matrix That Learns to Modify Itself


Feb 11, 2022
Kazuki Irie, Imanol Schlag, Róbert Csordás, Jürgen Schmidhuber

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

Improving Baselines in the Wild


Dec 31, 2021
Kazuki Irie, Imanol Schlag, Róbert Csordás, Jürgen Schmidhuber

Add code

* Presented at NeurIPS 2021 Workshop on Distribution Shifts, https://openreview.net/forum?id=9vxOrkNTs1x 

   Access Paper or Ask Questions

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

Training and Generating Neural Networks in Compressed Weight Space


Dec 31, 2021
Kazuki Irie, J├╝rgen Schmidhuber

Add code

* Presented at ICLR 2021 Workshop on Neural Compression, https://openreview.net/forum?id=qU1EUxdVd_D 

   Access Paper or Ask Questions

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

The Neural Data Router: Adaptive Control Flow in Transformers Improves Systematic Generalization


Oct 14, 2021
Róbert Csordás, Kazuki Irie, Jürgen Schmidhuber

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
>>