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

Computational materials discovery has continually grown in utility over thepast decade due to advances in computing power and crystal structure predictionalgorithms (CSPA). However, the computational cost of the \textit{ab initio}calculations required by CSPA limits its utility to small unit cells, reducingthe compositional and structural space the algorithms can explore. Past studieshave bypassed many unneeded \textit{ab initio} calculations by utilizingmachine learning methods to predict formation energy and determine thestability of a material. Specifically, graph neural networks display highfidelity in predicting formation energy. Traditionally graph neural networksare trained on large data sets of relaxed structures. Unfortunately, thegeometries of unrelaxed candidate structures produced by CSPA often deviatefrom the relaxed state, which leads to poor predictions hindering the model'sability to filter energetically unfavorable prior to \textit{ab initio}evaluation. This work shows that the prediction error on relaxed structuresreduces as training progresses, while the prediction error on unrelaxedstructures increases, suggesting an inverse correlation between relaxed andunrelaxed structure prediction accuracy. To remedy this behavior, we propose asimple, physically motivated, computationally cheap perturbation technique thataugments training data to improve predictions on unrelaxed structuresdramatically. On our test set consisting of 623 Nb-Sr-H hydride structures, wefound that training a crystal graph convolutional neural networks, utilizingour augmentation method, reduced the MAE of formation energy prediction by 66\%compared to training with only relaxed structures. We then show how this errorreduction can accelerates CSPA by improving the model's ability to filter outenergetically unfavorable structures accurately.

Data-Augmentation for Graph Neural Network Learning of the Relaxed Energies of Unrelaxed Structures