Bridging machine learning and neuroscience with a silicon nanosheet neuromorphic device
In brief 2D properties of the silicon nanosheets (SiNSs) enable the neuromorphic behaviors of the devices, which are rigorously analyzed by mathematical deduction and modeling.The prototype chip demonstrated on-chip pattern recognition capabilities.The quasi-ReLU activation function of the SiNS device network maps onto artificial neural networks used in machine learning.
Authors
- Anzhe Chen (ORCID: https://orcid.org/0009-0005-5921-3160)
- Wei Sun (ORCID: https://orcid.org/0000-0002-8525-619X)
- Zhenyi Ni (ORCID: https://orcid.org/0000-0002-8379-9182)
- X. Yuan (ORCID: https://orcid.org/0009-0000-3521-2837)
- Chenhao Wang (ORCID: https://orcid.org/0000-0002-6471-7962)
- Xinyi Xu (ORCID: https://orcid.org/0000-0003-0523-7982)
- Feng Tian (ORCID: https://orcid.org/0000-0002-9579-5023)
- Deren Yang
- Yang Xu (ORCID: https://orcid.org/0009-0000-9798-0903)
- Bodepudi Srikrishna Chanakya
Institutions
- Westlake University (CN)
- Institute of Semiconductors (CN)
- Zhejiang University (CN)
Publication Details
- Journal
- Cell Reports Physical Science
- Published
- 2025-10-22
- DOI
- https://doi.org/10.1016/j.xcrp.2025.102915
- Citations
- 1
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.73
Funders
- Fundamental Research Funds for the Central Universities
- Natural Science Foundation of Zhejiang Province