Pattern recognition in multi-synaptic photonic spiking neural networks based on a DFB-SA chip

Yanan Han; Shuiying Xiang; Ziwei Song; Shuang Gao; Xingxing Guo; Yahui Zhang; Yuechun Shi; Xiangfei Chen; Yue Hao

doi:10.29026/oes.2023.230021

Opto-Electronic Science, Volume. 2, Issue 9, 230021-1(2023)

Pattern recognition in multi-synaptic photonic spiking neural networks based on a DFB-SA chip

Yanan Han1... Shuiying Xiang1,*, Ziwei Song1, Shuang Gao1, Xingxing Guo1, Yahui Zhang1, Yuechun Shi2, Xiangfei Chen3 and Yue Hao1 |Show fewer author(s)

Author Affiliations

¹State Key Laboratory of Integrated Service Networks, State Key Discipline Laboratory of Wide Bandgap Semiconductor Technology, Xidian University, Xi’an 710071, China

²Yongjiang Laboratory, Ningbo 315202, China

³Key Laboratory of Intelligent Optical Sensing and Manipulation, Ministry of Education, the National Laboratory of Solid State Microstructures, the College of Engineering and Applied Sciences, Institute of Optical Communication Engineering, Nanjing University, Nanjing 210023, China

show less

Figures & Tables(7)

Fig. 1. (a) The geometry of multi-synaptic connections. (b) The experimental setup. The optical path is denoted by yellow, and the electrical path is black. The GDS file and chip micrograph of the DFB-SA neuron chip. (c1–c3) The optical spectrum of free-running DFB-SA, marked with different injection wavelengths. (d1–d3) The optical spectrum of DFB-SA with external optical injection corresponding to (c1–c3). (e1–e3) and (f1–f3) The input pattern and the corresponding output.

Download full size

View in Article

Fig. 2. The basic structure for supervised learning of photonic SNNs with temporal coding.

Download full size

View in Article

Fig. 3. (a) The process of digit pattern recognition based on a multi-synaptic photonic spiking neural network. (b) The output classification. (c) The basic principle for temporal coding SNN. (d) The STDP rule. (e) The mean distance of 4 different training methods with different N_syn.

Download full size

View in Article

Fig. 4. The output spike timing of all 10 output neurons (the row) at the input of all 10 patterns (the column) when N_syn=3.

Download full size

View in Article

Fig. 5. (a) The converged Distance with 1, 2, and 3 synapses, respectively. (b1–b2) Normalized weights for each of the two sets of synapses after convergence. (c1–c2) Normalized delays for each of the two sets of synapses after convergence. (d1) The converged Distance in the parameter space of weight mismatch and delay mismatch and (d2) projection of the contour with Distance=1.

Download full size

View in Article

Fig. 6. (a1–j1) The input spike patterns for all the 10 output neurons. (a2–j2) The correlated outputs of DFB-SA. (k1) The output spike interval for all 10 output neurons of 20 experiment trials. (k2) The output spike power for all 10 output neurons of 20 experiment trials.

Download full size

View in Article

Table 1. Accuracy of MNIST classification based on different models with different N_syn.

View table

View in Article

Table 1. Accuracy of MNIST classification based on different models with different N_syn.

Algorithm	Network size	N_syn=1	N_syn=2	N_syn=3
Tempotron	20×10	(25.71±2.43)%	(30.4±3.51)%	(34.80±2.68)%
ReSuMe	20×10	(19.4±2.19)%	(27.75±5.56)%	(29.60±4.10)%
DL-ReSuMe	20×10	(37±7.78)%	(42.80±1.30)%	(47.20±2.39)%
ANN-BP	20×10	(54.60±7.13)%	(56.67±5.28)%	(60.67±3.14)%
CNN-BP	(784×16)×10	(79.4±1.52)%		/

Tools

Get Citation

Copy Citation Text

Yanan Han, Shuiying Xiang, Ziwei Song, Shuang Gao, Xingxing Guo, Yahui Zhang, Yuechun Shi, Xiangfei Chen, Yue Hao. Pattern recognition in multi-synaptic photonic spiking neural networks based on a DFB-SA chip[J]. Opto-Electronic Science, 2023, 2(9): 230021-1

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Research Articles

Received: Jul. 9, 2023

Accepted: Sep. 25, 2023

Published Online: Jan. 22, 2024

The Author Email: Xiang Shuiying (syxiang@xidian.edu.cn)

DOI:10.29026/oes.2023.230021

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Accuracy of MNIST classification based on different models with different Nsyn.

Table 1. Accuracy of MNIST classification based on different models with different Nsyn.

微信扫一扫：分享

Table 1. Accuracy of MNIST classification based on different models with different N_syn.

Table 1. Accuracy of MNIST classification based on different models with different N_syn.