CiNet Hosoda Group

  • Japanese
  • Contact
スライドショー
スライドショー
スライドショー
スライドショー
スライドショー

Welcome to Hosoda group, CiNet

We explore "What is self, what is human, and what is life," and develop brain-inspired AI.

  • Our group was launched in September 2021. We are currently working mainly on mathematical models and fMRI experiments.

Research

The specific research themes are as follows. We welcome bringing themes and collaborative research.

  • Hebbian-inspired One-shot learning for deep-neural-network models
  • Simulating Eureka effect in visual object recognition using artificial neural network.
  • Whole-brain fMRI analysis of Eureka effect in visual object recognition
  • Enhancing Eureka ability of humans (collaborative research with Nittono Laboratory, Osaka University)
  • Fluctuation-driven super energy-saving computer
  • Construction of AI with “Self”

    • Members

    We enjoy deep discussions with these members

  • Kazufumi Hosoda; Senior Researcher (PI) [Prof in Kobe-U, Osaka-U]
  • Komaki Kunishige; Research technician
  • Keigo Nishida; Collaborative researcher [Postdoc in Riken]
  • Wataru Miyamori; Collaborative researcher [M student in Osaka-U]
  • Misako Kimura; Collaborative researcher [M student in U-Hyogo]
  • Tsutomu Murata; Senior Researcher (Advisor)
  • Izumi Ohzawa; Invited Specialist (Advisor) [Prof Emeritus in Osaka-U]
  • Kunihiko Kaneko; Invited Specialist (Advisor) [Prof Emeritus in Tokyo-U, Prof in Niels Bohr Inst]

    • Publications

    Preprints

  • Hiroaki Ohishi, Soya Shinkai, Hitoshi Owada, Takeru Fujii, Kazufumi Hosoda, Shuichi Onami, Takashi Yamamoto, Yasuyuki Ohkawa and Hiroshi Ochiai (2023). "Transcription-coupled changes in higher-order genomic structure and transcription hub viscosity prolong enhancer-promoter connectivity." bioRxiv: 2023.2011. 2027.568629.

    • Peer-reviewed

  • Kazufumi Hosoda, Keigo Nishida, Shigeto Seno, Tomohiro Mashita, Hideki Kashioka and Izumi Ohzawa (2024). "A single fast Hebbian-like process enabling one-shot class addition in deep neural networks without backbone modification." Frontiers in Neuroscience 18.
  • Kosuke Tomimatsu, Takeru Fujii, Ryoma Bise, Kazufumi Hosoda, Yosuke Taniguchi, Hiroshi Ochiai, Hiroaki Ohishi, Kanta Ando, Ryoma Minami, Kaori Tanaka, Taro Tachibana, Seiichi Mori, Akihito Harada, Kazumitsu Maehara, Masao Nagasaki, Seiichi Uchida, Hiroshi Kimura, Masashi Narita, Yasuyuki Ohkawa (2024). "Precise immunofluorescence canceling for highly multiplexed imaging to capture specific cell states." Nature Communications 15: 3657
  • Kazufumi Hosoda, Shigeto Seno, Naomi Murakami, Hideo Matsuda, Yutaka Osada, Rikuto Kamiura and Michio Kondoh (2024). "Synthetic model ecosystem of 12 cryopreservable microbial species allowing for a noninvasive approach." Biosystems: 105087.
  • Kazufumi Hosoda, Shigeto Seno, Rikuto Kamiura, Naomi Murakami and Michio Kondoh (2023). "Biodiversity and Constrained Information Dynamics in Ecosystems: A Framework for Living Systems." Entropy 25(12): 1624. (Feature Paper)
  • Kazufumi Hosoda, Shigeto Seno and Tsutomu Murata (2023). "Simulating Reaction Time for Eureka Effect in Visual Object Recognition Using Artificial Neural Network." IIAI Letters on Informatics and Interdisciplinary Research 3. 60, (KICSS Best Crowd Award)
  • Yusuke Morito, Tsutomu Murata, “Accumulation System: Distributed Neural Substrates of Perceptual Decision Making Revealed by fMRI Deconvolution,” J. Neurosci., vol. 42, no. 24, pp. 4891?4912, Jun. 2022, doi: 10.1523/JNEUROSCI.1062-21.2022. (*draw journal attention: J Neurosci 42: 8596?8598, 2022)

    • Patents

  • International: PCT/JP2023/30480; Autoencoding system, program, and method
  • Japanese: 2022-107734; Learning systems, machine learning programs, and machine learning methods

    • Access & Contact

    E-mail: hosodak at nict.go.jp
    〒565-0871  1-4 Yamadaoka, Suita City, Osaka, 565-0871
    Center for Information and Neural Networks (CiNet)
    (in Osaka university)