基于芯片的高维光神经网络

IF 31.6 1区材料科学 Q1 MATERIALS SCIENCE, MULTIDISCIPLINARY

Nano-Micro Letters Pub Date : 2022-11-14 DOI:10.1007/s40820-022-00957-8

Xinyu Wang, Peng Xie, Bohan Chen, Xingcai Zhang

{"title":"基于芯片的高维光神经网络","authors":"Xinyu Wang, Peng Xie, Bohan Chen, Xingcai Zhang","doi":"10.1007/s40820-022-00957-8","DOIUrl":null,"url":null,"abstract":"<div><h2>Highlights</h2><div>\n \n \n <ul>\n <li>\n <p>High-dimensional optical neural network is achieved by introducing an on-chip soliton microcomb source and wavelength division multiplexing technique.</p>\n </li>\n <li>\n <p>The programmable electro-optic nonlinear layer and optical meshes promote the implementation of a multi-layer optical neural network.</p>\n </li>\n <li>\n <p>Ultra-low coupling loss is realized between functional chips and fiber array, which is around 1 dB per facet.</p>\n </li>\n </ul>\n \n </div></div>","PeriodicalId":48779,"journal":{"name":"Nano-Micro Letters","volume":"14 1","pages":""},"PeriodicalIF":31.6000,"publicationDate":"2022-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9663775/pdf/","citationCount":"0","resultStr":"{\"title\":\"Chip-Based High-Dimensional Optical Neural Network\",\"authors\":\"Xinyu Wang, Peng Xie, Bohan Chen, Xingcai Zhang\",\"doi\":\"10.1007/s40820-022-00957-8\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><h2>Highlights</h2><div>\\n \\n \\n <ul>\\n <li>\\n <p>High-dimensional optical neural network is achieved by introducing an on-chip soliton microcomb source and wavelength division multiplexing technique.</p>\\n </li>\\n <li>\\n <p>The programmable electro-optic nonlinear layer and optical meshes promote the implementation of a multi-layer optical neural network.</p>\\n </li>\\n <li>\\n <p>Ultra-low coupling loss is realized between functional chips and fiber array, which is around 1 dB per facet.</p>\\n </li>\\n </ul>\\n \\n </div></div>\",\"PeriodicalId\":48779,\"journal\":{\"name\":\"Nano-Micro Letters\",\"volume\":\"14 1\",\"pages\":\"\"},\"PeriodicalIF\":31.6000,\"publicationDate\":\"2022-11-14\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9663775/pdf/\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Nano-Micro Letters\",\"FirstCategoryId\":\"88\",\"ListUrlMain\":\"https://link.springer.com/article/10.1007/s40820-022-00957-8\",\"RegionNum\":1,\"RegionCategory\":\"材料科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"MATERIALS SCIENCE, MULTIDISCIPLINARY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Nano-Micro Letters","FirstCategoryId":"88","ListUrlMain":"https://link.springer.com/article/10.1007/s40820-022-00957-8","RegionNum":1,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"MATERIALS SCIENCE, MULTIDISCIPLINARY","Score":null,"Total":0}

引用次数: 0

摘要

亮点通过引入片上孤子微梳源和波分复用技术，实现了高维光神经网络。可编程电光非线性层和光网格促进了多层光神经网络的实现。功能芯片与光纤阵列之间实现了超低耦合损耗，每个面的耦合损耗约为 1 dB。

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

Chip-Based High-Dimensional Optical Neural Network

Highlights

High-dimensional optical neural network is achieved by introducing an on-chip soliton microcomb source and wavelength division multiplexing technique.
The programmable electro-optic nonlinear layer and optical meshes promote the implementation of a multi-layer optical neural network.
Ultra-low coupling loss is realized between functional chips and fiber array, which is around 1 dB per facet.

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

Nano-Micro Letters NANOSCIENCE & NANOTECHNOLOGY-MATERIALS SCIENCE, MULTIDISCIPLINARY

CiteScore

42.40

自引率

4.90%

发文量

715

审稿时长

13 weeks

期刊介绍： Nano-Micro Letters is a peer-reviewed, international, interdisciplinary and open-access journal that focus on science, experiments, engineering, technologies and applications of nano- or microscale structure and system in physics, chemistry, biology, material science, pharmacy and their expanding interfaces with at least one dimension ranging from a few sub-nanometers to a few hundreds of micrometers. Especially, emphasize the bottom-up approach in the length scale from nano to micro since the key for nanotechnology to reach industrial applications is to assemble, to modify, and to control nanostructure in micro scale. The aim is to provide a publishing platform crossing the boundaries, from nano to micro, and from science to technologies.