R
Ryota Tomioka
Researcher at Microsoft
Publications - 115
Citations - 8789
Ryota Tomioka is an academic researcher from Microsoft. The author has contributed to research in topics: Artificial neural network & Rank (linear algebra). The author has an hindex of 38, co-authored 110 publications receiving 7514 citations. Previous affiliations of Ryota Tomioka include University of Illinois at Chicago & Tokyo Institute of Technology.
Papers
More filters
Journal ArticleDOI
Optimizing Spatial filters for Robust EEG Single-Trial Analysis
Benjamin Blankertz,Ryota Tomioka,S. Lemm,Motoaki Kawanabe,Klaus-Robert Müller,Klaus-Robert Müller +5 more
TL;DR: The theoretical background of the common spatial pattern (CSP) algorithm, a popular method in brain-computer interface (BCD research), is elucidated and tricks of the trade for achieving a powerful CSP performance are revealed.
Proceedings Article
QSGD: Communication-Efficient SGD via Gradient Quantization and Encoding
TL;DR: Quantized SGD (QSGD) as discussed by the authors is a family of compression schemes for gradient updates which provides convergence guarantees for convex and nonconvex objectives, under asynchrony, and can be extended to stochastic variance-reduced techniques.
Proceedings Article
f -GAN: training generative neural samplers using variational divergence minimization
TL;DR: In this paper, the generative adversarial training (GAN) approach is shown to be a special case of an existing more general variational divergence estimation approach, and any f-divergence can be used for training GANs.
Proceedings Article
Norm-Based Capacity Control in Neural Networks
TL;DR: In this article, the capacity, convexity and characterization of a general family of norm-constrained feed-forward networks are investigated, and the capacity of the network is characterized.
Posted Content
QSGD: Communication-Efficient SGD via Gradient Quantization and Encoding
TL;DR: Quantized SGD is proposed, a family of compression schemes for gradient updates which provides convergence guarantees and leads to significant reductions in end-to-end training time, and can be extended to stochastic variance-reduced techniques.