Predicting Optical Property in One-Dimensional Axionic Photonic Crystals by Deep Learning Approach

De Araújo Faria, Anny Caroline (2025) Predicting Optical Property in One-Dimensional Axionic Photonic Crystals by Deep Learning Approach. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (12MB) | Preview
Preview	PDF (Configuration Manual) Download (3MB) | Preview

Abstract

Axionic photonic crystals (APCs) are a generalized class of topological photonic materials characterized by physical parameters ϵ, µ, and θ. Controlling their transmission spectra is crucial for developing future axionic photonic devices. The central problem is that traditional computational methods (like TMM) are slow and require substantial memory. Furthermore, the literature lacked a deep learning prediction focused specifically on the axionic crystal domain, and it was unproven if MLPs could accurately predict phenomena characterized by abrupt oscillations, such as their transmission spectra. This study employs a Multi-Layer Perceptron (MLP) to predict the transmission spectra of 1D APCs, solving the traditional method’s cost limitations. The study was conducted in three experiments with increasing input complexity (X, R, δ). Datasets were generated via TMM and an MLP was trained to map inputs to the output spectrum T(ω), with performance evaluated primarily by Euclidean Distance (ED). The results show the MLP is highly efficient on calculation time. The model achieved its best accuracy in the one-parameter experiment, with a mean ED = 0.0365. However, in the three experiments the results indicate the MLP works well in low-complexity regimes but struggles to model abrupt oscillations that occur in high-non-linearity parameter regimes near δ = 1.0, where performance was poor and unstable. This study validates the MLP as an effective method to accelerate 1DAPC analysis and provides the foundation for all future research into the inverse design of these complex axionic devices.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Hamill, David UNSPECIFIED
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	07 Sep 2026 10:17
Last Modified:	07 Sep 2026 10:17
URI:	https://norma.ncirl.ie/id/eprint/9854

Actions (login required)

View Item