READ ME File For 'THz-TDS: extracting complex conductivity of two-dimensional materials via neural networks trained on synthetic and experimental data - Dataset'

Dataset DOI: https://doi.org/10.5258/SOTON/D3433 

ReadMe Author: Ben Beddoes, University of Southampton

This dataset supports the publication:
Ben Beddoes, Nick Klokkou, Jon Gorecki, Patrick Whelan, Peter Boggild, Peter Jepsen and Vasilis Apostolopoulos

THz-TDS: extracting complex conductivity of two-dimensional materials via neural networks trained on synthetic and experimental data
Optics Express
DOI: https://doi.org/10.1364/OE.557580

This dataset contains:

Training and validation data used for 4 networks:
	- Large synthetically trained network using 2e5 training sets 
Comparison networks for different training sets
	- Small synthetic network 1600 data sets 
	- Experimental network 800 data sets
	- Hybrid trained network with 1600 datasets (800 experimental sets, 800 synthetic sets)


The files (contained in ZIP folder) are as follows:

**LargeSynthetic_TrainingData** Training data for network used in figure 3
**ReducedSynthetic_TrainingData** Training data for network trained using purely synthetic data figure 4a
**Exp_TrainingData** Training data for network trained using purely experimental data, figure 4b
**Hybrid_TrainingData** Training data for hybrid trained network, figure 4c
**Synthetic_ValidationData** Synthetic validation data, figure 4
**ExpOOD_ValidationData** Out-of-distribution experimental validation data, figure 4

Dates of data collection: Oct-Dec 2024 

Information about geographic location of data collection: University of Southampton, U.K.
Licence - CC-BY


Date that the file was created: 20/03/2025