MMI-Based Pseudo-Real-Value Silicon Photonic Neural Networks
We present a pseudo-real-value silicon photonic chip based on cascaded 4×4 multimode interference couplers and phase shifter arrays for convolutional neural networks, which experimentally demonstrated a classification accuracy of 98.4% on the MNIST dataset.
Authors
- Dapeng Liu
- Guoguang Yao (ORCID: https://orcid.org/0000-0001-7385-8487)
- Naidi Cui
- Wenjie Yuan
- Shang Gao (ORCID: https://orcid.org/0009-0007-7783-0431)
- Dawei Wang (ORCID: https://orcid.org/0000-0002-6651-4261)
- Yang Zhao (ORCID: https://orcid.org/0000-0002-7916-8687)
- Qiang Li (ORCID: https://orcid.org/0000-0001-9344-6682)
- Shengping Liu (ORCID: https://orcid.org/0000-0002-0603-9972)
- Junbo Feng
Publication Details
- Published
- 2024-11-02
- DOI
- https://doi.org/10.1109/acp/ipoc63121.2024.10810083
- Citations
- 1
- Primary Topic
- Neural Networks and Reservoir Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.35
Funders
- National Natural Science Foundation of China
- China Postdoctoral Science Foundation