Author:
Kim Changdae,Lee Inho,Yoon Gwanho
Abstract
Burgeoning demand on the analysis of multilayered optical thin films yields an
advent of the simulation methods to calculate the optical response of
structures. Machine learning ensures expedite calculation once a model has been
trained. Here, we present a basic structure of deep neural network to
demonstrate a validity of machine learning-based approach by showing how well
the trained models approximate the target optical spectral responses. The
numerical simulations with several tunable conditions show the overall
inclination of performance of the models. This study paves the way for using
machine learning-based approaches to predict the spectral responses of optical
thin films.
Funder
Seoul National University of Science and
Technology
Publisher
Korea Flexible & Printed Electronics Society