Training Deep Neural Networks for the Inverse Design of Nanophotonic Structures,ACS Photonics

当前位置： X-MOL 学术 › ACS Photonics › 论文详情

Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)

Training Deep Neural Networks for the Inverse Design of Nanophotonic Structures
ACS Photonics ( IF 6.5 ) Pub Date : 2018-02-25 00:00:00 , DOI: 10.1021/acsphotonics.7b01377
Dianjing Liu ₁ , Yixuan Tan ₁ , Erfan Khoram ₁ , Zongfu Yu ₁

Affiliation

Data inconsistency leads to a slow training process when deep neural networks are used for the inverse design of photonic devices, an issue that arises from the fundamental property of nonuniqueness in all inverse scattering problems. Here we show that by combining forward modeling and inverse design in a tandem architecture, one can overcome this fundamental issue, allowing deep neural networks to be effectively trained by data sets that contain nonunique electromagnetic scattering instances. This paves the way for using deep neural networks to design complex photonic structures that require large training data sets.

中文翻译：

训练深层神经网络进行纳米光子结构逆设计

当将深度神经网络用于光子设备的逆设计时，数据不一致会导致训练过程变慢，这是由所有逆散射问题中非唯一性的基本属性引起的。在这里，我们表明，通过将前向建模和逆向设计组合到一前一后的体系结构中，可以克服这一基本问题，从而可以通过包含非唯一电磁散射实例的数据集来有效地训练深度神经网络。这为使用深度神经网络设计需要大量训练数据集的复杂光子结构铺平了道路。

更新日期：2018-02-25

点击分享查看原文

点击收藏

公开下载

阅读更多本刊最新论文本刊介绍/投稿指南11