About me

Keigo NISHIDA (西田 圭吾)

I am a postdoctoral researcher in the Laboratory for Computational Molecular Design at RIKEN BDR, and I also concurrently work in the Approximate Bayesian Inference Team at RIKEN AIP. In RIKEN, I am working on exploring new computers that efficiently process Bayesian neural netrworks and applying Bayesian deep learning to drug discovery. I am also a visiting researcher at the CiNet working on fluctuation-driven information processing.

Reserch Keywords : Bayesian Deep Learning, Variational Inference, Uncertainty, Domain Specific Architectures

Work Experience

Education

Grants

Publications

Journal (peer-reviewed)

  1. Kazufumi Hosoda, Keigo Nishida, Shigeto Seno, Tomohiro Mashita, Hideki Kashioka, Izumi Ohzawa, in Front. Neurosci., A Single Fast Hebbian-like Process Enabling One-Shot Class Addition in Deep Neural Networks Without Backbone Modification, Apr. 17, 2024.
  2. paper
  3. Keigo Nishida, Makoto Taiji, AdamB: Decoupled Bayes by Backprop With Gaussian Scale Mixture Prior, in IEEE Access, vol. 10, pp. 92959-92970, 2022, doi: 10.1109/ACCESS.2022.3203484.
  4. paper code

Proceedings (peer-reviewed)

  1. Gentaro Morimoto, Yohei M. Koyama, Hao Zhang, Teruhisa S. Komatsu, Yousuke Ohno, Keigo Nishida, Itta Ohmura, Hiroshi Koyama, Makoto Taiji, Hardware acceleration of tensor-structured multilevel ewald summation method on MDGRAPE-4A, a special-purpose computer system for molecular dynamics simulations, in Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, pp. 1-15, 2021
  2. paper

Presentations

Invited Talk

  1. 西田 圭吾, AdamB: AdamWのベイズ拡張による安定したベイズ深層学習法 人工知能学会 第125回人工知能基本問題研究会(SIG-FPAI), 盛岡, 2022.8.29-30

Links