Language selection

Government of Canada / Gouvernement du Canada

Search

Adversarial generation of mesoscale surfaces from small-scale chemical motifs

From National Research Council Canada

Download
  1. (PDF, 4.3 MiB)
DOIResolve DOI: https://doi.org/10.1021/acs.jpcc.0c06673
AuthorSearch for: 1ORCID identifier: https://orcid.org/0000-0003-1768-6873; Search for: ; Search for: 1ORCID identifier: https://orcid.org/0000-0002-8146-6667
Affiliation
  1. National Research Council of Canada. Security and Disruptive Technologies
FormatText, Article
Subjectenergy; neural networks; layers; lattices; computer simulations
Abstract
We demonstrate the use of a regressive upscaling generative adversarial network (RUGAN) as an effective way to sample state space for hexagonal porous graphene sheets. The RUGAN can, after being trained on a set of small-scale examples, generate new, energetically relevant microstates (atomic configurations). The RUGAN can generate configurations across a continuum of total energy values and produce configurations at requested energy values. The microstates produced respect periodic boundary conditions, and importantly, the fully convolutional nature of the generator allows the generation of arbitrarily large microstates, after being trained on only a small-scale data set.
Publication date
PublisherAmerican Chemical Society
In
  • The Journal of Physical Chemistry C 124, no. 42: 23158–23163.
LanguageEnglish
Peer reviewedYes
Export citationExport as RIS
Report a correctionReport a correction (opens in a new tab)
Record identifier513a20ea-ab11-4cad-9056-8795d4f1f5b2
Record created2021-05-11
Record modified2021-09-21
Date modified: