多波长二值光学衍射神经网络  

Multi-Wavelength Binary Optical Diffraction Neural Network

作　　者：陈龙 栾海涛[2] Long Chen;Haitao Luan(School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai;Institute of Photonic Chips,University of Shanghai for Science and Technology,Shanghai)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海 [2]上海理工大学光子芯片研究院,上海

出　　处：《建模与仿真》2024年第5期5013-5020,共8页Modeling and Simulation

基　　金：国家重点研发计划(2022YFB2804301)。

摘　　要：随着人工智能的快速发展,传统基于电子的计算受到摩尔定律的限制。衍射神经网络作为一种新兴的技术,凭借其优越的鲁棒性,已经在许多领域得到广泛应用。在这里,我们提出了一种可以在多波长中泛用的光学衍射神经网络。该神经网络在多个波长的图像识别准确率都能保持在88%以上,为处理各种环境下的光芯片提供可能。With the rapid development of artificial intelligence,traditional electronic-based computation is constrained by Moore’s Law.Diffraction neural networks,as an emerging technology,have been widely applied in many fields due to their superior robustness.Here,we propose an optical diffraction neural network that can be used across multiple wavelengths.The image recognition accuracy of this neural network maintains over 88%across multiple wavelengths,providing possibilities for processing optical chips in various environments.

关 键 词：衍射神经网络 多波长 光芯片 

分 类 号：TP1[自动化与计算机技术—控制理论与控制工程]