Recurrent Neural Network-based Model for Accelerated Trajectory Analysis in AIMD Simulations

Understand

The presented work demonstrates the training of recurrent neural networks (RNNs) from distributions of atom coordinates in solid state structures that were obtained using ab initio molecular dynamics (AIMD) simulations.

• AIMD simulations on solid state structures are treated as a multi-variate time-series problem.
• By referring interactions between atoms over the simulation time to temporary correlations among them, RNNs find patterns in the multi-variate time-dependent data, which enable forecasting trajectory paths and potential energy profiles.
• Two types of RNNs, namely gated recurrent unit and long short-term memory networks, are considered.