English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In the field of Machine Learning Interatomic Potentials (MLIPs),understanding the intricate relationship between data biases, specificallyconformational and structural diversity, and model generalization is criticalin improving the quality of Quantum Mechanics (QM) data generation efforts. Weinvestigate these dynamics through two distinct experiments: a fixed budgetone, where the dataset size remains constant, and a fixed molecular set one,which focuses on fixed structural diversity while varying conformationaldiversity. Our results reveal nuanced patterns in generalization metrics.Notably, for optimal structural and conformational generalization, a carefulbalance between structural and conformational diversity is required, butexisting QM datasets do not meet that trade-off. Additionally, our resultshighlight the limitation of the MLIP models at generalizing beyond theirtraining distribution, emphasizing the importance of defining applicabilitydomain during model deployment. These findings provide valuable insights andguidelines for QM data generation efforts.

Role of Structural and Conformational Diversity for Machine Learning Potentials