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

The use of machine learning for material property prediction and discoveryhas traditionally centered on graph neural networks that incorporate thegeometric configuration of all atoms. However, in practice not all thisinformation may be readily available, e.g.~when evaluating the potentiallyunknown binding of adsorbates to catalyst. In this paper, we investigatewhether it is possible to predict a system's relaxed energy in the OC20 datasetwhile ignoring the relative position of the adsorbate with respect to theelectro-catalyst. We consider SchNet, DimeNet++ and FAENet as basearchitectures and measure the impact of four modifications on modelperformance: removing edges in the input graph, pooling independentrepresentations, not sharing the backbone weights and using an attentionmechanism to propagate non-geometric relative information. We find that whileremoving binding site information impairs accuracy as expected, modified modelsare able to predict relaxed energies with remarkably decent MAE. Our worksuggests future research directions in accelerated materials discovery whereinformation on reactant configurations can be reduced or altogether omitted.

On the importance of catalyst-adsorbate 3D interactions for relaxed energy predictions