Prediction of Runtime Parameters of Parallel Chemistry Applications via Active and Generative Learning

2026-06-15

AI Summary

This study addresses the challenges of runtime parameter prediction and scarce training data in parallel quantum chemical computations by proposing a gradient-boosted regression tree method that integrates active learning with generative learning. To the best of our knowledge, this is the first work to synergistically combine these two learning paradigms in this domain, effectively mitigating data scarcity constraints. The proposed approach achieves a mean absolute percentage error (MAPE) of approximately 0.2 using only 20–25% of the training data, and further attains a remarkably low MAPE of 0.023 with a coefficient of determination (R²) of 99.9% on CCSD tasks. These results substantially outperform existing methods, offering robust support for efficient resource scheduling and performance optimization in computational chemistry workflows.

Abstract

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.

Problem

• runtime prediction
• parallel chemistry applications
• active learning
• generative learning
• machine learning

Innovation

• active learning
• generative learning
• gradient boosted regression trees
• runtime prediction
• parallel chemistry computations