计算机科学
编码(内存)
人工神经网络
巨量平行
功能(生物学)
深度学习
人工智能
信息处理
反向传播
控制(管理)
并行处理
深层神经网络
光学计算
钥匙(锁)
计算机体系结构
分布式计算
传播延迟
作者
Guannan Wang,Xiaofei Zang,Teng Zhang,Zhiyu Tan,Ziqing Guo,Yiming Zhu,Xianzhong Chen,Songlin Zhuang
出处
期刊:Science Advances
[American Association for the Advancement of Science]
日期:2025-11-21
卷期号:11 (47): eadu0904-eadu0904
被引量:15
标识
DOI:10.1126/sciadv.adu0904
摘要
We propose a directional-diffractive deep neural network (D-D 2 NN) by encoding the wave propagation direction into a D 2 NN, introducing a new degree of freedom and enable high-capacity information processing. The unprecedented capability of metasurfaces in controlling light propagation provides a new opportunity to realize such a D-D 2 NN. We propose and demonstrate a versatile D-D 2 NN using three spin-decoupled metasurfaces. Simultaneous control of geometric and propagation phases is used to realize direction-dependent functionality. High-capacity information processing is demonstrated by controlling the distances between neighboring metasurfaces. The efficacy of our approach is exemplified through the classification of digits and fashion products in two channels and the calculation-like function realized in four channels. Moreover, the proposed D-D 2 NN can divide rich information into multiple channels, enabling high-volume data encryption. The metasurface-enabled deep learning networks with versatile directional functionalities provide a flexible route toward massively parallel processing, pattern recognition, encryption, and artificial intelligence systems.
科研通智能强力驱动
Strongly Powered by AbleSci AI