Skip to Main content Skip to Navigation
Journal articles

Capabilities and limits of autoencoders for extracting collective variables in atomistic materials science

Abstract : Free energy calculations in materials science are routinely hindered by the need to provide re- action coordinates that can meaningfully partition atomic configuration space, a prerequisite for most enhanced sampling approaches. Recent studies on molecular systems have highlighted the possibility of constructing appropriate collective variables directly from atomic motions through deep learning techniques. Here we extend this class of approaches to condensed matter problems, using autoencoder neural networks to construct a scalar collective variable for use with the adaptive biasing force method. Particular attention is given to design choices required for application to crystalline systems with defects, including the filtering of thermal motions which otherwise domi- nate the autoencoder input. The machine-learning workflow is tested on body-centered cubic iron and its common defects, such as small vacancy or self-interstitial clusters and screw dislocations. For localized defects, excellent collective variables as well as derivatives, necessary for free energy sampling, are systematically obtained. However, the approach has a limited accuracy when dealing with reaction coordinates that include atomic displacements of a magnitude comparable to thermal motions, e.g. the ones produced by the long-range elastic field of dislocations. We then combine the extraction of collective variables by autoencoders with an adaptive biasing force free energy method based on Bayesian inference. Using a vacancy migration as an example, we demonstrate the performance of coupling these two approaches for simultaneous discovery of reaction coordinates and free energy sampling in systems with localized defects.
Complete list of metadata

https://hal.archives-ouvertes.fr/hal-03783038
Contributor : T D Swinburne Connect in order to contact the contributor
Submitted on : Thursday, September 22, 2022 - 11:00:17 AM
Last modification on : Friday, September 23, 2022 - 5:15:02 AM

File

 Restricted access
To satisfy the distribution rights of the publisher, the document is embargoed until : 2023-09-21

Please log in to resquest access to the document

Identifiers

Citation

Jacopo Baima, Alexandra Goryaeva, Thomas Swinburne, Jean-Bernard Maillet, Maylise Nastar, et al.. Capabilities and limits of autoencoders for extracting collective variables in atomistic materials science. Physical Chemistry Chemical Physics, Royal Society of Chemistry, 2022, https://doi.org/10.1039/D2CP01917E. ⟨10.1039/D2CP01917E⟩. ⟨hal-03783038⟩

Share

Metrics

Record views

0