I'm Hiromichi Kamata (鎌田啓路), a master course student of Harada-Mukuta lab at the University of Tokyo. My research interests are Spiking Neural Networks (SNN), accelerating Neural Networks, speech recognitions, etc.
Keywords
- Spiking Neural Networks
- Neuromorphic computing
- Generative models
- Speech recognition
- Network compression
Publications
- Hiromichi Kamata, Yusuke Mukuta, Tatsuya Harada, "Fully Spiking Variational Autoencoder", Thirty-Sixth AAAI Conference on Artificial Intelligence (AAAI), 2022.
Arxiv: https://arxiv.org/abs/2110.00375
Github: https://github.com/kamata1729/FullySpikingVAE
Biography
Sony R&D center; Apr.2022 - Now
work as an AI researcher
The University of Tokyo; Apr.2020 - Mar.2022
Graduate School of Information Science and Technology. I am working on expanding the applicability of Spiking Neural Networks.
The University of Tokyo; Apr.2016 - Mar.2020
Department of mechano informatics. I studied to improve the efficiency of neural architecture search (NAS).
Internships
Sony R&D center; Sep.2020 (1 month)
I researched methods for recovering images in datasets from trained models.
Recruit; Mar.2021 (1 month)
I analyzed data for the real estate app called suumo.
Coding Skills
- Python use in my research or various applications like Deep learning (Pytorch), image processing, data analysis, etc.
- Swift use in the internship and create some ios apps.
- C++ implement pytorch models in c++ to speed up.
- Java create some Android apps.