摘要
随着人工智能技术的快速发展,全球计算需求呈现指数级增长,这对数据处理的速度和效率提出了更高要求。然而在后摩尔时代,传统电子计算技术因受电子器件物理极限和冯·诺依曼架构的限制,面临着速度和能效难以突破的技术瓶颈,因此急需一种新的技术范式革新。在这一背景下,光计算技术因其独特优势受到广泛关注,而马赫-曾德尔干涉仪作为光计算领域的重要关键器件之一,展现出突破传统计算限制的潜力。本文综述了马赫-曾德尔干涉仪的基本原理及其构建酉矩阵的架构,探讨了其在不同波导网络构建中的应用,并重点分析了其在矩阵运算、光学神经网络和可编程光子电路等领域的典型应用。最后,本文对马赫-曾德尔干涉仪的未来发展方向进行了展望。
With the rapid development of artificial intelligence technology,the global computing demand has shown exponential growth,which places higher requirements on the speed and efficiency of data processing.However,in the post-Moore era,the traditional electronic computing technology is facing the technical bottlenecks in speed and energy efficiency due to the physical limits of electronic devices and the constraints of the von Neumann architecture.So,a new technical paradigm innovation is urgently needed.In this context,optical computing technology has received extensive attention due to its unique advantages,and the Mach-Zehnder interferometer,as one of the important key devices in the field of optical computing,shows the potential to break through the limitations of traditional computing.In this paper,we review the basic principle of Mach-Zehnder interferometer and its architecture for constructing unitary matrices,discuss its applications in the construction of different waveguide networks,and focus on analyzing its typical applications in the fields of matrix arithmetic,optical neural networks and programmable photonic circuits.Finally,this paper gives an outlook on the future development directions of Mach-Zehnder interferometer.
作者
王圣尧
佀国翔
路翠翠
Wang Shengyao;Si Guoxiang;Lu Cuicui(School of Physics,Beijing Institute of Technology,100081,Beijing,China)
出处
《山东师范大学学报(自然科学版)》
2025年第3期193-218,共26页
Journal of Shandong Normal University(Natural Sciences Edition)
基金
国家自然科学基金资助项目(12274031)
北京理工大学特立青年学者高层次人才科研启动计划资助项目。
关键词
马赫-曾德尔干涉仪
光计算
光子集成电路
光学神经网络
Mach-Zehnder interferometer
optical computing
photonic integrated circuits
optical neural networks
