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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

We present our latest advancements of machine-learned potentials (MLPs) basedon the neuroevolution potential (NEP) framework introduced in [Fan et al.,Phys. Rev. B 104, 104309 (2021)] and their implementation in the open-sourcepackage GPUMD. We increase the accuracy of NEP models both by improving theradial functions in the atomic-environment descriptor using a linearcombination of Chebyshev basis functions and by extending the angulardescriptor with some four-body and five-body contributions as in the atomiccluster expansion approach. We also detail our efficient implementation of theNEP approach in graphics processing units as well as our workflow for theconstruction of NEP models, and we demonstrate their application in large-scaleatomistic simulations. By comparing to state-of-the-art MLPs, we show that theNEP approach not only achieves above-average accuracy but also is far morecomputationally efficient. These results demonstrate that the GPUMD package isa promising tool for solving challenging problems requiring highly accurate,large-scale atomistic simulations. To enable the construction of MLPs using aminimal training set, we propose an active-learning scheme based on the latentspace of a pre-trained NEP model. Finally, we introduce three separate Pythonpackages, GPYUMD, CALORINE, and PYNEP, which enable the integration of GPUMDinto Python workflows.

GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations