AI光学神经网络计算芯片发展现状及趋势  

Development Status and Trends of AI Optical Neural Network Computing Chips

作　　者：杨卓樾 钱广 黄梦昊 孔月婵[1] 陈堂胜[1] YANG Zhuoyue;QIAN Guang;HUANG Menghao;KONG Yuechan;CHEN Tangsheng(National Key Laboratory of Solid Microwave Devices and Circuit,Nanjing Electronic Devices Institute,Nanjing 210016,CHN)

机构地区：[1]南京电子器件研究所固态微波器件与电路全国重点实验室,南京210016

出　　处：《光电子技术》2025年第1期1-9,共9页Optoelectronic Technology

基　　金：国家重点研发计划(2022YFB2802502,2022YFB2802702)。

摘　　要：介绍了用于人工智能的光学神经网络芯片工作原理、系统架构以及应用场景,分析了近年来国内外的研究现状,总结了关键技术和未来的发展趋势。为突破传统电子计算“功耗墙”和“算力墙”提供了全新技术路径,在人工智能领域展现出诱人的应用潜力。The working principles,system architectures,and application scenarios of optical neural network(ONN)chips for artificial intelligence were introduced.The recent research progress in both domestic and international contexts were analyzed,while key technologies and future development trends were outlined and summarized.By providing a novel technical pathway to overcome the"power consumption wall"and"computing power wall"inherent in traditional electronic computing,this approach could demonstrate promising application potential in the field of artificial intelligence.

关 键 词：光计算 光学神经网络 人工智能 

分 类 号：TN256[电子电信—物理电子学]