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

Machine learning assisted modeling of the inter-atomic potential energysurface (PES) is revolutionizing the field of molecular simulation. With theaccumulation of high-quality electronic structure data, a model that can bepretrained on all available data and finetuned on downstream tasks with a smalladditional effort would bring the field to a new stage. Here we propose DPA-1,a Deep Potential model with a novel attention mechanism, which is highlyeffective for representing the conformation and chemical spaces of atomicsystems and learning the PES. We tested DPA-1 on a number of systems andobserved superior performance compared with existing benchmarks. Whenpretrained on large-scale datasets containing 56 elements, DPA-1 can besuccessfully applied to various downstream tasks with a great improvement ofsample efficiency. Surprisingly, for different elements, the learned typeembedding parameters form a $spiral$ in the latent space and have a naturalcorrespondence with their positions on the periodic table, showing interestinginterpretability of the pretrained DPA-1 model.

DPA-1: Pretraining of Attention-based Deep Potential Model for Molecular Simulation