Reconfigurable Low-Threshold All-Optical Nonlinear Activation Functions Based on an Add-Drop Silicon Microring Resonator
The realization of optical nonlinear activation functions (NAFs) is essential for integrated optical neural networks (ONNs). Here, we propose and experimentally demonstrate a photonic method to implement reconfigurable and low-threshold all-optical NAFs based on a compact and high-Q add-drop microri...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Photonics Journal |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9940587/ |
| _version_ | 1811319096881446912 |
|---|---|
| author | Weizhen Yu Shuang Zheng Zhenyu Zhao Bin Wang Weifeng Zhang |
| author_facet | Weizhen Yu Shuang Zheng Zhenyu Zhao Bin Wang Weifeng Zhang |
| author_sort | Weizhen Yu |
| collection | DOAJ |
| description | The realization of optical nonlinear activation functions (NAFs) is essential for integrated optical neural networks (ONNs). Here, we propose and experimentally demonstrate a photonic method to implement reconfigurable and low-threshold all-optical NAFs based on a compact and high-Q add-drop microring resonator (MRR) on silicon. In the experiment, four different NAFs including softplus, radial basis, clamped ReLU, and sigmoid functions are realized by exploiting the thermo-optical (TO) effect of the MRR. The threshold to implement NAFs is as low as 0.08 mW. As a demonstration, a handwritten digit classification benchmark task is simulated based on a convolutional neural network (CNN) using the obtained activation functions, where an accuracy of 98% is realized. Thanks to the unique advantages of ultra-compact footprint and ultralow threshold, the proposed nonlinear unit is promising to be widely used in large-scale integrated ONNs. |
| first_indexed | 2024-04-13T12:37:04Z |
| format | Article |
| id | doaj.art-4de699fcdd7d4a30a2f816cb73495ee6 |
| institution | Directory Open Access Journal |
| issn | 1943-0655 |
| language | English |
| last_indexed | 2024-04-13T12:37:04Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Photonics Journal |
| spelling | doaj.art-4de699fcdd7d4a30a2f816cb73495ee62022-12-22T02:46:37ZengIEEEIEEE Photonics Journal1943-06552022-01-011461710.1109/JPHOT.2022.32192469940587Reconfigurable Low-Threshold All-Optical Nonlinear Activation Functions Based on an Add-Drop Silicon Microring ResonatorWeizhen Yu0https://orcid.org/0000-0003-2122-353XShuang Zheng1https://orcid.org/0000-0002-4825-5163Zhenyu Zhao2Bin Wang3https://orcid.org/0000-0001-6431-4365Weifeng Zhang4https://orcid.org/0000-0001-6454-1314Radar Research Lab, School of Information and Electronics, Beijing institute of Technology, Beijing, ChinaRadar Research Lab, School of Information and Electronics, Beijing institute of Technology, Beijing, ChinaRadar Research Lab, School of Information and Electronics, Beijing institute of Technology, Beijing, ChinaRadar Research Lab, School of Information and Electronics, Beijing institute of Technology, Beijing, ChinaRadar Research Lab, School of Information and Electronics, Beijing institute of Technology, Beijing, ChinaThe realization of optical nonlinear activation functions (NAFs) is essential for integrated optical neural networks (ONNs). Here, we propose and experimentally demonstrate a photonic method to implement reconfigurable and low-threshold all-optical NAFs based on a compact and high-Q add-drop microring resonator (MRR) on silicon. In the experiment, four different NAFs including softplus, radial basis, clamped ReLU, and sigmoid functions are realized by exploiting the thermo-optical (TO) effect of the MRR. The threshold to implement NAFs is as low as 0.08 mW. As a demonstration, a handwritten digit classification benchmark task is simulated based on a convolutional neural network (CNN) using the obtained activation functions, where an accuracy of 98% is realized. Thanks to the unique advantages of ultra-compact footprint and ultralow threshold, the proposed nonlinear unit is promising to be widely used in large-scale integrated ONNs.https://ieeexplore.ieee.org/document/9940587/Optical neural networkmicroring resonatorall-optical nonlinear activation functionsthermo-optical effect |
| spellingShingle | Weizhen Yu Shuang Zheng Zhenyu Zhao Bin Wang Weifeng Zhang Reconfigurable Low-Threshold All-Optical Nonlinear Activation Functions Based on an Add-Drop Silicon Microring Resonator IEEE Photonics Journal Optical neural network microring resonator all-optical nonlinear activation functions thermo-optical effect |
| title | Reconfigurable Low-Threshold All-Optical Nonlinear Activation Functions Based on an Add-Drop Silicon Microring Resonator |
| title_full | Reconfigurable Low-Threshold All-Optical Nonlinear Activation Functions Based on an Add-Drop Silicon Microring Resonator |
| title_fullStr | Reconfigurable Low-Threshold All-Optical Nonlinear Activation Functions Based on an Add-Drop Silicon Microring Resonator |
| title_full_unstemmed | Reconfigurable Low-Threshold All-Optical Nonlinear Activation Functions Based on an Add-Drop Silicon Microring Resonator |
| title_short | Reconfigurable Low-Threshold All-Optical Nonlinear Activation Functions Based on an Add-Drop Silicon Microring Resonator |
| title_sort | reconfigurable low threshold all optical nonlinear activation functions based on an add drop silicon microring resonator |
| topic | Optical neural network microring resonator all-optical nonlinear activation functions thermo-optical effect |
| url | https://ieeexplore.ieee.org/document/9940587/ |
| work_keys_str_mv | AT weizhenyu reconfigurablelowthresholdallopticalnonlinearactivationfunctionsbasedonanadddropsiliconmicroringresonator AT shuangzheng reconfigurablelowthresholdallopticalnonlinearactivationfunctionsbasedonanadddropsiliconmicroringresonator AT zhenyuzhao reconfigurablelowthresholdallopticalnonlinearactivationfunctionsbasedonanadddropsiliconmicroringresonator AT binwang reconfigurablelowthresholdallopticalnonlinearactivationfunctionsbasedonanadddropsiliconmicroringresonator AT weifengzhang reconfigurablelowthresholdallopticalnonlinearactivationfunctionsbasedonanadddropsiliconmicroringresonator |