Neural networks determination of nematic elastic constants - Jupyter Notebook

The uploaded Jupyter Notebook is a collection of Python functions that are used in a method for the determination of nematic elastic constants, which is described in [Zaplotnik et al. SciRep 2023].
Running all the notebook cells one after another, one can create a training data set (i.e. generate hundreds of random initial states of liquid crystal samples with random elastic constants, numerically simulate the liquid crystal dynamics and light transmission), train a simple sequential neural network, import experimentally measured data, and at the end, use the neural network to determine elastic constants of the 5CB liquid crystal used in the experiments.

Required Python libraries: \(\texttt{numpy}, \texttt{ scipy}, \texttt{ numba}, \texttt{ matplotlib}, \texttt{ tensorflow}, \texttt{ pandas}, \texttt{ csv}, \texttt{ time}.\)

The most recent version of this notebook can be found on Google Colab, where the Python code can be very easily executed online without any installations of required libraries. 

Supplementary material (training data sets) is uploaded on Zenodo.