Multicore Photonic Complex-Valued Neural Network with Transformation Layer
Photonic neural network chips have been widely studied because of their low power consumption, high speed and large bandwidth. Using amplitude and phase to encode, photonic chips can accelerate complex-valued neural network computations. In this article, a photonic complex-valued neural network (PCN...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-05-01
|
| Series: | Photonics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2304-6732/9/6/384 |
| _version_ | 1827657451931959296 |
|---|---|
| author | Ruiting Wang Pengfei Wang Chen Lyu Guangzhen Luo Hongyan Yu Xuliang Zhou Yejin Zhang Jiaoqing Pan |
| author_facet | Ruiting Wang Pengfei Wang Chen Lyu Guangzhen Luo Hongyan Yu Xuliang Zhou Yejin Zhang Jiaoqing Pan |
| author_sort | Ruiting Wang |
| collection | DOAJ |
| description | Photonic neural network chips have been widely studied because of their low power consumption, high speed and large bandwidth. Using amplitude and phase to encode, photonic chips can accelerate complex-valued neural network computations. In this article, a photonic complex-valued neural network (PCNN) chip is designed. The scale of the single-core PCNN chip is limited because of optical losses, and the multicore architecture of the chip is used to improve computing capability. Further, for improving the performance of the PCNN, we propose the transformation layer, which can be implemented by the designed photonic chip to transform real-valued encoding to complex-valued encoding, which has richer information. Compared with real-valued input, the transformation layer can effectively improve the classification accuracy from 93.14% to 97.51% of a 64-dimensional input on the MNIST test set. Finally, we analyze the multicore computation of the PCNN. Compared with the single-core architecture, the multicore architecture can improve the classification accuracy by implementing larger neural networks and has better phase noise robustness. The proposed architecture and algorithms are beneficial to promote the accelerated computing of photonic chips for complex-valued neural networks and are promising for use in many applications, such as image recognition and signal processing. |
| first_indexed | 2024-03-09T22:43:22Z |
| format | Article |
| id | doaj.art-3bafd7d4bd99414696ded83cfa107708 |
| institution | Directory Open Access Journal |
| issn | 2304-6732 |
| language | English |
| last_indexed | 2024-03-09T22:43:22Z |
| publishDate | 2022-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Photonics |
| spelling | doaj.art-3bafd7d4bd99414696ded83cfa1077082023-11-23T18:32:43ZengMDPI AGPhotonics2304-67322022-05-019638410.3390/photonics9060384Multicore Photonic Complex-Valued Neural Network with Transformation LayerRuiting Wang0Pengfei Wang1Chen Lyu2Guangzhen Luo3Hongyan Yu4Xuliang Zhou5Yejin Zhang6Jiaoqing Pan7The Key Laboratory of Semiconductor Materials Science, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, ChinaThe Key Laboratory of Semiconductor Materials Science, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, ChinaThe Key Laboratory of Semiconductor Materials Science, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, ChinaThe Key Laboratory of Semiconductor Materials Science, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, ChinaThe Key Laboratory of Semiconductor Materials Science, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, ChinaThe Key Laboratory of Semiconductor Materials Science, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, ChinaThe Center of Materials Science and Optoelectronics Engineering, University of Chinese Academy of Sciences, Beijing 100049, ChinaThe Key Laboratory of Semiconductor Materials Science, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, ChinaPhotonic neural network chips have been widely studied because of their low power consumption, high speed and large bandwidth. Using amplitude and phase to encode, photonic chips can accelerate complex-valued neural network computations. In this article, a photonic complex-valued neural network (PCNN) chip is designed. The scale of the single-core PCNN chip is limited because of optical losses, and the multicore architecture of the chip is used to improve computing capability. Further, for improving the performance of the PCNN, we propose the transformation layer, which can be implemented by the designed photonic chip to transform real-valued encoding to complex-valued encoding, which has richer information. Compared with real-valued input, the transformation layer can effectively improve the classification accuracy from 93.14% to 97.51% of a 64-dimensional input on the MNIST test set. Finally, we analyze the multicore computation of the PCNN. Compared with the single-core architecture, the multicore architecture can improve the classification accuracy by implementing larger neural networks and has better phase noise robustness. The proposed architecture and algorithms are beneficial to promote the accelerated computing of photonic chips for complex-valued neural networks and are promising for use in many applications, such as image recognition and signal processing.https://www.mdpi.com/2304-6732/9/6/384photonic neural networkmulticore architecturecomplex-valued neural network |
| spellingShingle | Ruiting Wang Pengfei Wang Chen Lyu Guangzhen Luo Hongyan Yu Xuliang Zhou Yejin Zhang Jiaoqing Pan Multicore Photonic Complex-Valued Neural Network with Transformation Layer Photonics photonic neural network multicore architecture complex-valued neural network |
| title | Multicore Photonic Complex-Valued Neural Network with Transformation Layer |
| title_full | Multicore Photonic Complex-Valued Neural Network with Transformation Layer |
| title_fullStr | Multicore Photonic Complex-Valued Neural Network with Transformation Layer |
| title_full_unstemmed | Multicore Photonic Complex-Valued Neural Network with Transformation Layer |
| title_short | Multicore Photonic Complex-Valued Neural Network with Transformation Layer |
| title_sort | multicore photonic complex valued neural network with transformation layer |
| topic | photonic neural network multicore architecture complex-valued neural network |
| url | https://www.mdpi.com/2304-6732/9/6/384 |
| work_keys_str_mv | AT ruitingwang multicorephotoniccomplexvaluedneuralnetworkwithtransformationlayer AT pengfeiwang multicorephotoniccomplexvaluedneuralnetworkwithtransformationlayer AT chenlyu multicorephotoniccomplexvaluedneuralnetworkwithtransformationlayer AT guangzhenluo multicorephotoniccomplexvaluedneuralnetworkwithtransformationlayer AT hongyanyu multicorephotoniccomplexvaluedneuralnetworkwithtransformationlayer AT xuliangzhou multicorephotoniccomplexvaluedneuralnetworkwithtransformationlayer AT yejinzhang multicorephotoniccomplexvaluedneuralnetworkwithtransformationlayer AT jiaoqingpan multicorephotoniccomplexvaluedneuralnetworkwithtransformationlayer |