Prediction of Runtime Parameters of Parallel Chemistry Applications via Active and Generative Learning
This work addresses the need for efficient runtime prediction in high-performance computing for chemistry applications, offering high accuracy with reduced training data.
The authors developed machine learning models to predict runtime parameters of parallel chemistry computations, achieving a mean absolute error percentage as low as 0.023 and a coefficient of determination of 99.9%. Active learning reduced data requirements to 20-25% of the original dataset while maintaining a MAPE of about 0.2.
In this work, we develop two main Machine Learning based approaches to predict the runtime parameters of highly scalable parallel chemistry computations.These approaches employ active and generative learning together with the empirically determined gradient boosted regression tree models chosen among a rich suite of machine learning models. When evaluated on Coupled-Cluster with Singles and Doubles computations, our models achieve a mean absolute error percentage (MAPE) as low as 0.023 and a coefficient of determination as high as 99.9%. Furthermore, when combined with active learning to mitigate the lack of large amounts of training data, our models score a MAPE about 0.2 with 20-25% of the original dataset.