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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Predicting electronic energies, densities, and related chemical propertiescan facilitate the discovery of novel catalysts, medicines, and batterymaterials. By developing a physics-inspired equivariant neural network, weintroduce a method to learn molecular representations based on the electronicinteractions among atomic orbitals. Our method, OrbNet-Equi, leveragesefficient tight-binding simulations and learned mappings to recover highfidelity quantum chemical properties. OrbNet-Equi models a wide spectrum oftarget properties with an accuracy consistently better than standard machinelearning methods and a speed orders of magnitude greater than densityfunctional theory. Despite only using training samples collected from readilyavailable small-molecule libraries, OrbNet-Equi outperforms traditional methodson comprehensive downstream benchmarks that encompass diverse main-groupchemical processes. Our method also describes interactions in challengingcharge-transfer complexes and open-shell systems. We anticipate that thestrategy presented here will help to expand opportunities for studies inchemistry and materials science, where the acquisition of experimental orreference training data is costly.

Informing Geometric Deep Learning with Electronic Interactions to Accelerate Quantum Chemistry