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 Taiji Suzuki

On Learnability via Gradient Method for Two-Layer ReLU Neural Networks in Teacher-Student Setting


Jun 29, 2021
Shunta Akiyama, Taiji Suzuki

* 47 pages, 3 figures 

  Access Paper or Ask Questions

Deep Two-Way Matrix Reordering for Relational Data Analysis


Apr 09, 2021
Chihiro Watanabe, Taiji Suzuki


  Access Paper or Ask Questions

Deep Two-way Matrix Reordering for Relational Data Analysis


Mar 26, 2021
Chihiro Watanabe, Taiji Suzuki


  Access Paper or Ask Questions

Goodness-of-fit Test on the Number of Biclusters in Relational Data Matrix


Feb 23, 2021
Chihiro Watanabe, Taiji Suzuki


  Access Paper or Ask Questions

Bias-Variance Reduced Local SGD for Less Heterogeneous Federated Learning


Feb 05, 2021
Tomoya Murata, Taiji Suzuki

* 19 pages 

  Access Paper or Ask Questions

Particle Dual Averaging: Optimization of Mean Field Neural Networks with Global Convergence Rate Analysis


Dec 31, 2020
Atsushi Nitanda, Denny Wu, Taiji Suzuki

* 32 pages 

  Access Paper or Ask Questions

Benefit of deep learning with non-convex noisy gradient descent: Provable excess risk bound and superiority to kernel methods


Dec 06, 2020
Taiji Suzuki, Shunta Akiyama

* 21 pages 

  Access Paper or Ask Questions

Estimation error analysis of deep learning on the regression problem on the variable exponent Besov space


Sep 27, 2020
Kazuma Tsuji, Taiji Suzuki


  Access Paper or Ask Questions

Neural Architecture Search Using Stable Rank of Convolutional Layers


Sep 19, 2020
Kengo Machida, Kuniaki Uto, Koichi Shinoda, Taiji Suzuki


  Access Paper or Ask Questions

Generalization bound of globally optimal non-convex neural network training: Transportation map estimation by infinite dimensional Langevin dynamics


Jul 11, 2020
Taiji Suzuki


  Access Paper or Ask Questions

When Does Preconditioning Help or Hurt Generalization?


Jul 02, 2020
Shun-ichi Amari, Jimmy Ba, Roger Grosse, Xuechen Li, Atsushi Nitanda, Taiji Suzuki, Denny Wu, Ji Xu

* 38 pages 

  Access Paper or Ask Questions

Optimal Rates for Averaged Stochastic Gradient Descent under Neural Tangent Kernel Regime


Jun 22, 2020
Atsushi Nitanda, Taiji Suzuki

* 36 pages 

  Access Paper or Ask Questions

Gradient Descent in RKHS with Importance Labeling


Jun 19, 2020
Tomoya Murata, Taiji Suzuki

* 18 pages, 12 figures 

  Access Paper or Ask Questions

Optimization and Generalization Analysis of Transduction through Gradient Boosting and Application to Multi-scale Graph Neural Networks


Jun 15, 2020
Kenta Oono, Taiji Suzuki

* 9 pages, Reference 5 pages, Supplemental material 18 pages 

  Access Paper or Ask Questions

Selective Inference for Latent Block Models


May 27, 2020
Chihiro Watanabe, Taiji Suzuki


  Access Paper or Ask Questions

Dimension-free convergence rates for gradient Langevin dynamics in RKHS


Mar 26, 2020
Boris Muzellec, Kanji Sato, Mathurin Massias, Taiji Suzuki


  Access Paper or Ask Questions

Meta Cyclical Annealing Schedule: A Simple Approach to Avoiding Meta-Amortization Error


Mar 04, 2020
Yusuke Hayashi, Taiji Suzuki

* 10 pages, 4 figures, 2 tables 

  Access Paper or Ask Questions

Understanding Generalization in Deep Learning via Tensor Methods


Jan 14, 2020
Jingling Li, Yanchao Sun, Jiahao Su, Taiji Suzuki, Furong Huang

* 9 pages (main paper), 42 pages (full version) 

  Access Paper or Ask Questions

Domain Adaptation Regularization for Spectral Pruning


Dec 26, 2019
Laurent Dillard, Yosuke Shinya, Taiji Suzuki


  Access Paper or Ask Questions

Exponential Convergence Rates of Classification Errors on Learning with SGD and Random Features


Nov 13, 2019
Shingo Yashima, Atsushi Nitanda, Taiji Suzuki


  Access Paper or Ask Questions

Scalable Deep Neural Networks via Low-Rank Matrix Factorization


Oct 29, 2019
Atsushi Yaguchi, Taiji Suzuki, Shuhei Nitta, Yukinobu Sakata, Akiyuki Tanizawa


  Access Paper or Ask Questions

Deep learning is adaptive to intrinsic dimensionality of model smoothness in anisotropic Besov space


Oct 28, 2019
Taiji Suzuki, Atsushi Nitanda


  Access Paper or Ask Questions

Compression based bound for non-compressed network: unified generalization error analysis of large compressible deep neural network


Sep 26, 2019
Taiji Suzuki


  Access Paper or Ask Questions

Understanding the Effects of Pre-Training for Object Detectors via Eigenspectrum


Sep 09, 2019
Yosuke Shinya, Edgar Simo-Serra, Taiji Suzuki

* ICCV 2019 Workshop on Neural Architects (Oral) 

  Access Paper or Ask Questions

Goodness-of-fit Test for Latent Block Models


Jul 09, 2019
Chihiro Watanabe, Taiji Suzuki


  Access Paper or Ask Questions