Chapter 4: Revisiting Machine Learning Methods for Predicting Hysteresis

Examining one MOF with molecular simulation is OK. Examining a thousand is too expensive. Can we use machine learning to predict hysteresis and vdW loops?

I went back to energy-histogram ML for help. Similar to predicting loadings at different fugacities, we can also predict fugacities at different loadings. Once we have a set of predicted fugacities at certain loadings values, we can connect these points by training an autoencoder that learns the vdW loop. This work is in preparation.