The rise of machine learning (ML) has created an explosion in the potential strategies for using data to make scientific predictions. For physical scientists wishing to apply ML strategies to a particular domain, it can be difficult to assess in advance what strategy to adopt within a vast space of possibilities. Here we outline the results of an online community-powered effort to swarm search the space of ML strategies and develop algorithms for predicting atomic-pairwise nuclear magnetic resonance (NMR) properties in molecules. Using an open-source dataset, we worked with Kaggle to design and host a 3-month competition which received 47,800 ML model predictions from 2,700 teams in 84 countries. Within 3 weeks, the Kaggle community produced models with comparable accuracy to our best previously published in-house efforts. A meta-ensemble model constructed as a linear combination of the top predictions has a prediction accuracy which exceeds that of any individual model, 7-19x better than our previous state-of-The-Art. The results highlight the potential of transformer architectures for predicting quantum mechanical (QM) molecular properties.
Bibliographical notePublisher Copyright:
© 2021 Public Library of Science. All rights reserved.
FingerprintDive into the research topics of 'A community-powered search of machine learning strategy space to find NMR property prediction models'. Together they form a unique fingerprint.
Polly E Eccleston (Other), Simon H Atack (Other), D A G Williams (Manager), Sadaf R Alam (Manager) & Steven A Chapman (Manager)