This project applies delta-learning to predict the difference between force field and DFT-level interaction energies for water-ion and water-water dimers.
Why it matters: Classical force fields fail in reactive/interfacial systems. This model learns quantum corrections, enabling high accuracy at low cost.
- Parse xyz file, generate dimer sets
- Predict delta energies using PyTorch feedforward NN
- Combine with force field baseline
- Evaluate with MAE / Pearson metrics
- MAE: 0.28 kcal/mol (test set)
- Pearson r: 0.95
- Supports F⁻, Cl⁻, Br⁻, I⁻, K⁺, Na⁺, Mg²⁺, Ca²⁺ clusters
python predict.py input.xyz model.pt