Genetic Deep Learning for Photonic Device Inverse Design

We demonstrate an inverse design method by combining deep neural networks and genetic optimization. The method significantly reduces the training data compared with existing methods. The method is adopted to design some ultra-compact photonic devices to show its high computing efficiency and great f...

Full description

Saved in:
Bibliographic Details
Main Authors: Ren, Yangming, Zhang, Lingxuan, Zhang, Wenfu, Sun, Xiaochen
Format: Conference Proceeding
Language:English
Subjects:
Conferences
Deep learning
Design methodology
Genetics
Inverse design
Machine learning
Optimization
Silicon
Silicon photonics
Training data
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We demonstrate an inverse design method by combining deep neural networks and genetic optimization. The method significantly reduces the training data compared with existing methods. The method is adopted to design some ultra-compact photonic devices to show its high computing efficiency and great flexibility.
ISSN:2575-274X
DOI:10.1109/IPC48725.2021.9593021