Active learning of neural network potentials for rare events

Developing an automated active learning framework for Neural Network Potentials, focusing on accurately simulating bond-breaking in hexane chains through steered molecular dynamics sampling and assessing model transferability.