Differentiable sampling of molecular geometries with uncertainty-based adversarial attacks

Neural Networks are known to perform poorly outside of their training domain. Here the authors propose an inverse sampling strategy to train neural network potentials enabling to drive atomistic systems towards high-likelihood and high-uncertainty configurations without the need for molecular dynami...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Daniel Schwalbe-Koda, Aik Rui Tan, Rafael Gómez-Bombarelli
Formato: article
Lenguaje:EN
Publicado: Nature Portfolio 2021
Materias:
Q
Acceso en línea:https://doaj.org/article/1014ecc724a64101ba64cd3dbf9a3690
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:1014ecc724a64101ba64cd3dbf9a3690
record_format dspace
spelling oai:doaj.org-article:1014ecc724a64101ba64cd3dbf9a36902021-12-02T16:35:07ZDifferentiable sampling of molecular geometries with uncertainty-based adversarial attacks10.1038/s41467-021-25342-82041-1723https://doaj.org/article/1014ecc724a64101ba64cd3dbf9a36902021-08-01T00:00:00Zhttps://doi.org/10.1038/s41467-021-25342-8https://doaj.org/toc/2041-1723Neural Networks are known to perform poorly outside of their training domain. Here the authors propose an inverse sampling strategy to train neural network potentials enabling to drive atomistic systems towards high-likelihood and high-uncertainty configurations without the need for molecular dynamics simulations.Daniel Schwalbe-KodaAik Rui TanRafael Gómez-BombarelliNature PortfolioarticleScienceQENNature Communications, Vol 12, Iss 1, Pp 1-12 (2021)
institution DOAJ
collection DOAJ
language EN
topic Science
Q
spellingShingle Science
Q
Daniel Schwalbe-Koda
Aik Rui Tan
Rafael Gómez-Bombarelli
Differentiable sampling of molecular geometries with uncertainty-based adversarial attacks
description Neural Networks are known to perform poorly outside of their training domain. Here the authors propose an inverse sampling strategy to train neural network potentials enabling to drive atomistic systems towards high-likelihood and high-uncertainty configurations without the need for molecular dynamics simulations.
format article
author Daniel Schwalbe-Koda
Aik Rui Tan
Rafael Gómez-Bombarelli
author_facet Daniel Schwalbe-Koda
Aik Rui Tan
Rafael Gómez-Bombarelli
author_sort Daniel Schwalbe-Koda
title Differentiable sampling of molecular geometries with uncertainty-based adversarial attacks
title_short Differentiable sampling of molecular geometries with uncertainty-based adversarial attacks
title_full Differentiable sampling of molecular geometries with uncertainty-based adversarial attacks
title_fullStr Differentiable sampling of molecular geometries with uncertainty-based adversarial attacks
title_full_unstemmed Differentiable sampling of molecular geometries with uncertainty-based adversarial attacks
title_sort differentiable sampling of molecular geometries with uncertainty-based adversarial attacks
publisher Nature Portfolio
publishDate 2021
url https://doaj.org/article/1014ecc724a64101ba64cd3dbf9a3690
work_keys_str_mv AT danielschwalbekoda differentiablesamplingofmoleculargeometrieswithuncertaintybasedadversarialattacks
AT aikruitan differentiablesamplingofmoleculargeometrieswithuncertaintybasedadversarialattacks
AT rafaelgomezbombarelli differentiablesamplingofmoleculargeometrieswithuncertaintybasedadversarialattacks
_version_ 1718383734827253760