Molecular dynamics is typically based on one of the two most popular interatomic interaction models, namely (1) quantum-mechanical (QM) models that are very accurate but very computationally expensive, and (2) empirical interatomic potentials that typically offer only a qualitative accuracy, but are very computationally efficient.
Machine learning is viewed as a promising tool for developing interatomic potentials with QM accuracy, but orders of magnitude more computationally efficient than the QM models. The challenge, however, is to make such potentials reliable - it requires fitting hundreds to thousands of parameters and making sure that they produce reasonable results in the entire region of interest in the phase space (which could be given only implicitly, e.g., all configurations with energy below a certain threshold).
In the first part of my talk I will present an example of accurate and computationally efficient machine learning interatomic potentials. In the second part I will advocate using active learning as a way to ensure reliability of such potentials.
Back to Workshop III: Collective Variables in Quantum Mechanics