Machine-learned prediction of the electronic fields in a crystal

Written by Ying Shi Teh for the following paper:

Ying Shi Teh, Swarnava Ghosh, and Kaushik Bhattacharya. (2021) Machine-learned prediction of the electronic fields in a crystal, JMechMat.

Folder descriptions

• 00_AbinitRun: Data generation and collection of DFT output using the Abinit software.
• 01_ML_fields: Training of PCA/NN models for electronic fields including electron density (DEN), Coulomb potential (VCLMB), band structure energy density (EBAND), and volumetric entropy (ENTR).
• 02_ML_XRED: Training of NN model for atomic positions in fractional (or reduced) coordinates (XRED).
• 03_Plots: Plots.
• 04_EvaluateResults: Evaluate energy values using learned quantities (DEN, EBAND, ENTR, and XRED).
• 05_AbinitConvergenceTest: Convergence test for DFT runs.
• 06_ML_ETOT: Training of NN model for total free energy (ETOT). Generate uniaxial stress-strain relationship using learned energy model.