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

Chrome logo Add to Chrome

Firefox logo Add to Firefox

Picture for Jürgen Schmidhuber

Going Beyond Linear Transformers with Recurrent Fast Weight Programmers


Jun 11, 2021
Kazuki Irie, Imanol Schlag, Róbert Csordás, Jürgen Schmidhuber


  Access Paper or Ask Questions

Is it Enough to Optimize CNN Architectures on ImageNet?


Mar 16, 2021
Lukas Tuggener, Jürgen Schmidhuber, Thilo Stadelmann


  Access Paper or Ask Questions

Spatial Dependency Networks: Neural Layers for Improved Generative Image Modeling


Mar 16, 2021
Đorđe Miladinović, Aleksandar Stanić, Stefan Bauer, Jürgen Schmidhuber, Joachim M. Buhmann

* International Conference on Learning Representations (2021); 

  Access Paper or Ask Questions

Linear Transformers Are Secretly Fast Weight Memory Systems


Feb 23, 2021
Imanol Schlag, Kazuki Irie, Jürgen Schmidhuber


  Access Paper or Ask Questions

Meta Learning Backpropagation And Improving It


Dec 29, 2020
Louis Kirsch, Jürgen Schmidhuber


  Access Paper or Ask Questions

On the Binding Problem in Artificial Neural Networks


Dec 09, 2020
Klaus Greff, Sjoerd van Steenkiste, Jürgen Schmidhuber


  Access Paper or Ask Questions

Unsupervised Object Keypoint Learning using Local Spatial Predictability


Nov 25, 2020
Anand Gopalakrishnan, Sjoerd van Steenkiste, Jürgen Schmidhuber

* pre-print. under review. 19 pages, 10 figures 

  Access Paper or Ask Questions

Learning Associative Inference Using Fast Weight Memory


Nov 16, 2020
Imanol Schlag, Tsendsuren Munkhdalai, Jürgen Schmidhuber


  Access Paper or Ask Questions

Hierarchical Relational Inference


Oct 07, 2020
Aleksandar Stanić, Sjoerd van Steenkiste, Jürgen Schmidhuber


  Access Paper or Ask Questions

Are Neural Nets Modular? Inspecting Functional Modularity Through Differentiable Weight Masks


Oct 05, 2020
Róbert Csordás, Sjoerd van Steenkiste, Jürgen Schmidhuber


  Access Paper or Ask Questions

Recurrent Neural-Linear Posterior Sampling for Non-Stationary Contextual Bandits


Jul 09, 2020
Aditya Ramesh, Paulo Rauber, Jürgen Schmidhuber


  Access Paper or Ask Questions

Parameter-based Value Functions


Jun 16, 2020
Francesco Faccio, Jürgen Schmidhuber

* 24 pages, 67 references 

  Access Paper or Ask Questions

Training Agents using Upside-Down Reinforcement Learning


Dec 05, 2019
Rupesh Kumar Srivastava, Pranav Shyam, Filipe Mutz, Wojciech Jaśkowski, Jürgen Schmidhuber

* NNAISENSE Technical Report. 17 pages, 6 figures 

  Access Paper or Ask Questions

Enhancing the Transformer with Explicit Relational Encoding for Math Problem Solving


Oct 15, 2019
Imanol Schlag, Paul Smolensky, Roland Fernandez, Nebojsa Jojic, Jürgen Schmidhuber, Jianfeng Gao


  Access Paper or Ask Questions

R-SQAIR: Relational Sequential Attend, Infer, Repeat


Oct 11, 2019
Aleksandar Stanić, Jürgen Schmidhuber

* 4 page workshop paper accepted at the NeurIPS 2019 Workshop on Perception as Generative Reasoning: Structure, Causality, Probability 

  Access Paper or Ask Questions

Improving Generalization in Meta Reinforcement Learning using Learned Objectives


Oct 09, 2019
Louis Kirsch, Sjoerd van Steenkiste, Jürgen Schmidhuber


  Access Paper or Ask Questions

Recurrent Neural Processes


Jun 13, 2019
Timon Willi, Jonathan Masci, Jürgen Schmidhuber, Christian Osendorfer


  Access Paper or Ask Questions

A Perspective on Objects and Systematic Generalization in Model-Based RL


Jun 03, 2019
Sjoerd van Steenkiste, Klaus Greff, Jürgen Schmidhuber

* Accepted to the ICML 2019 workshop on Workshop on Generative Modeling and Model-Based Reasoning for Robotics and AI 

  Access Paper or Ask Questions

Are Disentangled Representations Helpful for Abstract Visual Reasoning?


May 29, 2019
Sjoerd van Steenkiste, Francesco Locatello, Jürgen Schmidhuber, Olivier Bachem

* This is a preliminary pre-print 

  Access Paper or Ask Questions

Improving Differentiable Neural Computers Through Memory Masking, De-allocation, and Link Distribution Sharpness Control


Apr 23, 2019
Róbert Csordás, Jürgen Schmidhuber


  Access Paper or Ask Questions

Learning to Reason with Third-Order Tensor Products


Nov 29, 2018
Imanol Schlag, Jürgen Schmidhuber


  Access Paper or Ask Questions

Recurrent World Models Facilitate Policy Evolution


Sep 04, 2018
David Ha, Jürgen Schmidhuber

* To appear at NIPS 2018, selected for an oral presentation. arXiv admin note: substantial text overlap with arXiv:1803.10122 

  Access Paper or Ask Questions

DeepScores -- A Dataset for Segmentation, Detection and Classification of Tiny Objects


May 26, 2018
Lukas Tuggener, Ismail Elezi, Jürgen Schmidhuber, Marcello Pelillo, Thilo Stadelmann

* 6 pages, accepted on IEEE International Conference on Pattern Recognition 2018 

  Access Paper or Ask Questions

World Models


May 09, 2018
David Ha, Jürgen Schmidhuber


  Access Paper or Ask Questions

Relational Neural Expectation Maximization: Unsupervised Discovery of Objects and their Interactions


Feb 28, 2018
Sjoerd van Steenkiste, Michael Chang, Klaus Greff, Jürgen Schmidhuber

* Accepted to ICLR 2018 

  Access Paper or Ask Questions

Neural Expectation Maximization


Nov 04, 2017
Klaus Greff, Sjoerd van Steenkiste, Jürgen Schmidhuber

* Accepted to NIPS 2017 

  Access Paper or Ask Questions

LSTM: A Search Space Odyssey


Oct 04, 2017
Klaus Greff, Rupesh Kumar Srivastava, Jan Koutník, Bas R. Steunebrink, Jürgen Schmidhuber

* IEEE Transactions on Neural Networks and Learning Systems ( Volume: 28, Issue: 10, Oct. 2017 ) Pages: 2222 - 2232 
* 12 pages, 6 figures 

  Access Paper or Ask Questions

Recurrent Highway Networks


Jul 04, 2017
Julian Georg Zilly, Rupesh Kumar Srivastava, Jan Koutník, Jürgen Schmidhuber

* 12 pages, 6 figures, 3 tables 

  Access Paper or Ask Questions

Highway and Residual Networks learn Unrolled Iterative Estimation


Mar 14, 2017
Klaus Greff, Rupesh K. Srivastava, Jürgen Schmidhuber

* 10 + 4 pages, accepted for ICLR 2017 

  Access Paper or Ask Questions