光卷积网络噪声量化仿真分析

IF 0.7 4区物理与天体物理 Q4 OPTICS

Optica Applicata Pub Date : 2023-01-01 DOI:10.37190/oa230311

None Ye Zhang, None Saining Zhang, None Danni Zhang, None Yanmei Su, None Junkai Yi, None Pengfei Wang, None Ruiting Wang, None Guangzhen Luo, None Xuliang Zhou, None Jiaoqing Pan

{"title":"光卷积网络噪声量化仿真分析","authors":"None Ye Zhang, None Saining Zhang, None Danni Zhang, None Yanmei Su, None Junkai Yi, None Pengfei Wang, None Ruiting Wang, None Guangzhen Luo, None Xuliang Zhou, None Jiaoqing Pan","doi":"10.37190/oa230311","DOIUrl":null,"url":null,"abstract":"Optical neural network (ONN) has been regarded as one of the most prospective techniques in the future, due to its high-speed and low power cost. However, the realization of optical convolutional neural network (CNN) in non-ideal cases still remains a big challenge. In this paper, we propose an optical convolutional networks system for classification problems by applying general matrix multiply (GEMM) technology. The results show that under the influence of noise, this system still has good performance with low TOP-1 and TOP-5 error rates of 44.26% and 14.51% for ImageNet. We also propose a quantization model of CNN. The noise quantization model reaches a sufficient prediction accuracy of about 96% for MNIST handwritten dataset.","PeriodicalId":19589,"journal":{"name":"Optica Applicata","volume":"9 1","pages":"0"},"PeriodicalIF":0.7000,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Noise quantization simulation analysis of optical convolutional networks\",\"authors\":\"None Ye Zhang, None Saining Zhang, None Danni Zhang, None Yanmei Su, None Junkai Yi, None Pengfei Wang, None Ruiting Wang, None Guangzhen Luo, None Xuliang Zhou, None Jiaoqing Pan\",\"doi\":\"10.37190/oa230311\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Optical neural network (ONN) has been regarded as one of the most prospective techniques in the future, due to its high-speed and low power cost. However, the realization of optical convolutional neural network (CNN) in non-ideal cases still remains a big challenge. In this paper, we propose an optical convolutional networks system for classification problems by applying general matrix multiply (GEMM) technology. The results show that under the influence of noise, this system still has good performance with low TOP-1 and TOP-5 error rates of 44.26% and 14.51% for ImageNet. We also propose a quantization model of CNN. The noise quantization model reaches a sufficient prediction accuracy of about 96% for MNIST handwritten dataset.\",\"PeriodicalId\":19589,\"journal\":{\"name\":\"Optica Applicata\",\"volume\":\"9 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.7000,\"publicationDate\":\"2023-01-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Optica Applicata\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.37190/oa230311\",\"RegionNum\":4,\"RegionCategory\":\"物理与天体物理\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q4\",\"JCRName\":\"OPTICS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Optica Applicata","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.37190/oa230311","RegionNum":4,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"OPTICS","Score":null,"Total":0}

引用次数: 0

摘要

光神经网络(ONN)以其高速、低功耗的特点被认为是未来最有发展前景的技术之一。然而，在非理想情况下实现光学卷积神经网络(CNN)仍然是一个很大的挑战。本文提出了一种基于广义矩阵乘法技术的光学卷积网络分类系统。结果表明，在噪声的影响下，该系统仍然具有良好的性能，对ImageNet的TOP-1和TOP-5错误率分别为44.26%和14.51%。我们还提出了CNN的量化模型。噪声量化模型对MNIST手写数据集的预测精度达到96%左右。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

Noise quantization simulation analysis of optical convolutional networks

Optical neural network (ONN) has been regarded as one of the most prospective techniques in the future, due to its high-speed and low power cost. However, the realization of optical convolutional neural network (CNN) in non-ideal cases still remains a big challenge. In this paper, we propose an optical convolutional networks system for classification problems by applying general matrix multiply (GEMM) technology. The results show that under the influence of noise, this system still has good performance with low TOP-1 and TOP-5 error rates of 44.26% and 14.51% for ImageNet. We also propose a quantization model of CNN. The noise quantization model reaches a sufficient prediction accuracy of about 96% for MNIST handwritten dataset.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

Optica Applicata 物理-光学

CiteScore

1.00

自引率

16.70%

发文量

审稿时长

4 months

期刊介绍： Acoustooptics, atmospheric and ocean optics, atomic and molecular optics, coherence and statistical optics, biooptics, colorimetry, diffraction and gratings, ellipsometry and polarimetry, fiber optics and optical communication, Fourier optics, holography, integrated optics, lasers and their applications, light detectors, light and electron beams, light sources, liquid crystals, medical optics, metamaterials, microoptics, nonlinear optics, optical and electron microscopy, optical computing, optical design and fabrication, optical imaging, optical instrumentation, optical materials, optical measurements, optical modulation, optical properties of solids and thin films, optical sensing, optical systems and their elements, optical trapping, optometry, photoelasticity, photonic crystals, photonic crystal fibers, photonic devices, physical optics, quantum optics, slow and fast light, spectroscopy, storage and processing of optical information, ultrafast optics.