Microring-Based Multi-Operand Optical Neurons with On-Chip Trainable Nonlinearity
We propose an optical neuron based on multi-operand microring resonators, featuring a compact footprint and high scalability with built-in tunable nonlinearity. Experimental results highlight its superior expressivity in machine learning tasks with exceptional hardware efficiency.
Authors
- David Z. Pan (ORCID: https://orcid.org/0000-0002-5705-2501)
- Ziang Yin (ORCID: https://orcid.org/0009-0009-5308-1100)
- Ray T. Chen (ORCID: https://orcid.org/0000-0002-9181-4266)
- Spencer Denton
- Chenghao Feng (ORCID: https://orcid.org/0000-0002-0751-7681)
- Shupeng Ning (ORCID: https://orcid.org/0000-0002-5780-9711)
- Jiaqi Gu (ORCID: https://orcid.org/0000-0001-8535-7698)
- Hanqing Zhu
Publication Details
- Published
- 2025-01-01
- DOI
- https://doi.org/10.1364/cleo_at.2025.aa120_1
- Primary Topic
- Neural Networks and Reservoir Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00