Applications of Machine Learning to Molecular Quantum Dynamics Calculations

In this work, we apply machine learning (ML) algorithms to molecular quantum dynamics calculations to develop methods that can be used to expand the databases while mitigating the computational expense of the explicit calculations. First, we compare two different machine learning algorithms- Gaussian Process Regression (GPR) and an Artificial Neural Network (ANN)- on their viability as prediction models in quantum parameter space in order to be used as solution generators of the Schr\"odinger equation for inelastic molecular scattering. Next, we present a data generation scheme using the probabilistic nature of GPR that reduces the need to compute explicit quantum dynamics simulations by predicting state-to-state rate coefficients for transitions where the values of initial molecular rotational state $j$ are not used to train the machine learning algorithm.