相关实验视频

Updated: Jun 1, 2026

Design, Fabrication, and Experimental Characterization of Plasmonic Photoconductive Terahertz Emitters

Design, Fabrication, and Experimental Characterization of Plasmonic Photoconductive Terahertz Emitters

Published on: July 8, 2013

特拉赫兹假冒等离子体神经网络用于衍射信息识别和处理

Xinxin Gao^1,2, Ze Gu², Qian Ma³

¹State Key Laboratory of Terahertz and Millimeter Waves, City University of Hong Kong, Hong Kong, China.

Nature communications

|August 6, 2024

概括

我们开发了一种紧的太赫兹伪造等离子体神经网络,用于高效的多目标识别. 这种衍射人工智能加速器可以直接处理复杂的数据,如手写数字.

科学领域:

物理物理学物理
材料科学材料科学材料科学
计算机科学计算机科学

背景情况:

全光衍射神经网络提供模拟AI加速,但由于空间传输效率低,面临小型化挑战.
现有的衍射网络需要很大的空间尺寸,这限制了它们的实际应用和集成.

研究的目的:

提出一个紧而高效的太赫兹欺骗等离子体神经网络,用于直接的多目标识别.
为了证明使用伪造表面等离子体极子对衍射神经网络层的可行性.

主要方法:

利用一个伪造的表面等离子体极立子合器阵列在平面平台上创建一个衍射网络层.
设计和测试了三个分类方案:基础向量分类,多用户识别和MNIST手写数字分类.
采用金属格子阵列,发射器和接收器来输入和处理数据.

主要成果:

通过使用太赫兹衍射网络成功分类基础向量并识别多用户定向信息.
用拟议的架构展示了MNIST手写数字的直接处理和分类.
实现了一个紧的,高效的和可集成的衍射神经网络架构.

结论:

太赫兹伪造等离子神经网络为微型模拟AI加速器提供了可行的解决方案.
这项工作推进了太赫兹等离子体超材料在芯片集成,智能通信和计算方面的应用.
拟议的平台允许在太赫兹域内高效的直接数据处理和识别.

更多相关视频

Plasmonic Trapping and Release of Nanoparticles in a Monitoring Environment

Plasmonic Trapping and Release of Nanoparticles in a Monitoring Environment

Published on: April 4, 2017

Monitoring Conformational Dynamics of Single Unmodified Proteins using Plasmonic Nanotweezers

Monitoring Conformational Dynamics of Single Unmodified Proteins using Plasmonic Nanotweezers

Published on: March 21, 2025

相关概念视频

Positron Emission Tomography

Positron Emission Tomography

Positron emission tomography (PET) is a medical imaging technique involving radiopharmaceuticals — substances that emit short-lived radiation. Although the first PET scanner was introduced in 1961, it took 15 more years before radiopharmaceuticals were combined with the technique and revolutionized its potential.
One of the main requirements of a PET scan is a positron-emitting radioisotope, which is produced in a cyclotron and then attached to a substance used by the part of the body being...