Jove
Visualize
联系我们
JoVE
x logofacebook logolinkedin logoyoutube logo
关于 JoVE
概览领导团队博客JoVE 帮助中心
作者
出版流程编辑委员会范围与政策同行评审常见问题投稿
图书馆员
用户评价订阅访问资源图书馆顾问委员会常见问题
研究
JoVE JournalMethods CollectionsJoVE Encyclopedia of Experiments存档
教育
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab Manual教师资源中心教师网站
使用条款与条件
隐私政策
政策

相关概念视频

Modeling and Similitude01:12

Modeling and Similitude

245
Scaled modeling is a fundamental technique in engineering, enabling the study of large and complex systems by creating smaller, manageable replicas that recreate critical characteristics of the original. In hydrology and civil infrastructure, for example, scaled models of dams help analyze water flow, turbulence, and pressure. This method allows for accurate predictions of real-world behavior within a controlled environment, significantly reducing the cost and time involved in full-scale...
245

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序
Same author

Fundamental absorption bandwidth to thickness limit for transparent homogeneous layers.

Nanophotonics (Berlin, Germany)·2024
Same author

Active and tunable nanophotonic metamaterials.

Nanophotonics (Berlin, Germany)·2024
Same author

Deep Learning Resolves Myovascular Dynamics in the Failing Human Heart.

JACC. Basic to translational science·2024
Same author

Heterogeneity and Memory Effect in the Sluggish Dynamics of Vacancy Defects in Colloidal Disordered Crystals and Their Implications to High-Entropy Alloys.

Advanced science (Weinheim, Baden-Wurttemberg, Germany)·2022
Same author

Inverse deep learning methods and benchmarks for artificial electromagnetic material design.

Nanoscale·2022
Same author

A Triple-Mode Midinfrared Modulator for Radiative Heat Management of Objects with Various Emissivity.

Nano letters·2021
Same journal

Recent Progress in on-Demand Transfer-Enabled Integration of Wavelength-Scale Light Sources.

Nanophotonics (Berlin, Germany)·2026
Same journal

Tunable skyrmion bag textures in surface phonon polariton lattices.

Nanophotonics (Berlin, Germany)·2026
Same journal

All-Optical Diffractive Operators for Rapid, Computer-Free Morphological Transformations.

Nanophotonics (Berlin, Germany)·2026
Same journal

Tunable Skyrmion, Meron, and Skyrmion Bag Textures in Surface Phonon Polariton Lattices.

Nanophotonics (Berlin, Germany)·2026
Same journal

Deep-Subwavelength Slot-Enhanced Broadband Dynamic Camouflage Metasurface Across the S, C, X, and Ku Bands.

Nanophotonics (Berlin, Germany)·2026
Same journal

Machine Learning-Driven Cooling Window Design Beyond Hyperbolic Metamaterials.

Nanophotonics (Berlin, Germany)·2026
查看所有相关文章

相关实验视频

Updated: Jun 5, 2025

Simulation, Fabrication and Characterization of THz Metamaterial Absorbers
13:44

Simulation, Fabrication and Characterization of THz Metamaterial Absorbers

Published on: December 27, 2012

15.3K

转移学习用于元材料设计和模拟.

Rixi Peng1, Simiao Ren1, Jordan Malof2,1

  • 1Electrical and Computer Engineering, Duke University, Durham, NC, USA.

Nanophotonics (Berlin, Germany)
|December 5, 2024
PubMed
概括
此摘要是机器生成的。

转移学习增强了对 metasurface 阵列的深度学习模型训练. 这种方法显著降低了数据需求,通过最小的训练数据实现了高精度.

关键词:
深度学习是一种深度学习.超材料是指金属材料.metasurfaces 是一个表层.散射散射是一种散射.转移学习转移学习

更多相关视频

Characterizing Dissipative Elastic Metamaterials Produced by Additive Manufacturing
09:39

Characterizing Dissipative Elastic Metamaterials Produced by Additive Manufacturing

Published on: June 28, 2024

850
Fabricating Metamaterials Using the Fiber Drawing Method
11:57

Fabricating Metamaterials Using the Fiber Drawing Method

Published on: October 18, 2012

13.8K

相关实验视频

Last Updated: Jun 5, 2025

Simulation, Fabrication and Characterization of THz Metamaterial Absorbers
13:44

Simulation, Fabrication and Characterization of THz Metamaterial Absorbers

Published on: December 27, 2012

15.3K
Characterizing Dissipative Elastic Metamaterials Produced by Additive Manufacturing
09:39

Characterizing Dissipative Elastic Metamaterials Produced by Additive Manufacturing

Published on: June 28, 2024

850
Fabricating Metamaterials Using the Fiber Drawing Method
11:57

Fabricating Metamaterials Using the Fiber Drawing Method

Published on: October 18, 2012

13.8K

科学领域:

  • 电磁主义 电磁主义
  • 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:

  • 深度学习模型,特别是残余神经网络 (ResNets),需要大量的数据来进行有效的培训.
  • 在电磁应用中至关重要的元表面阵列,存在复杂的多尺度问题,模拟需要大量的数据.
  • 数据瓶是将深度学习应用于新型超表面设计的重大挑战.

研究的目的:

  • 展示转移学习作为一种提高培训深度学习模型对 metasurface 阵列的效率的方法.
  • 评估跨不同问题领域的转移学习的有效性,与最初的培训任务有不同程度的相似性.
  • 为了减轻深度学习中的数据瓶,用于电磁超材料研究.

主要方法:

  • 利用剩余神经网络 (ResNets) 进行深度学习模型培训.
  • 采用准分析的离散二极距近似 (DDA) 方法来模拟电气上大型的超表面阵列.
  • 应用转移学习,使预先训练的ResNet模型适应新的超表面设计问题.

主要成果:

  • 转移学习大大减少了训练深度学习模型所需的数据量.
  • 在最佳传输场景中,通过预训练的神经网络实现了3%的测试平均绝对相对误差.
  • 在最佳情况下,转移学习应用程序的数据减少了1000倍.
  • 展示了跨越相关电磁超材料问题的频谱的转移学习的效率.

结论:

  • 转移学习是一种强大的工具,可以加速超表面阵列的设计和分析.
  • 这种方法有效地克服了电磁超材料深度学习中的数据瓶.
  • 通过转移学习利用预先训练的模型,即使使用有限的数据,也可以实现快速训练和高精度.