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

Obtaining accurate solutions to the Schr\"odinger equation is the keychallenge in computational quantum chemistry. Deep-learning-based VariationalMonte Carlo (DL-VMC) has recently outperformed conventional approaches in termsof accuracy, but only at large computational cost. Whereas in many domainsmodels are trained once and subsequently applied for inference, accurate DL-VMCso far requires a full optimization for every new problem instance, consumingthousands of GPUhs even for small molecules. We instead propose a DL-VMC modelwhich has been pre-trained using self-supervised wavefunction optimization on alarge and chemically diverse set of molecules. Applying this model to newmolecules without any optimization, yields wavefunctions and absolute energiesthat outperform established methods such as CCSD(T)-2Z. To obtain accuraterelative energies, only few fine-tuning steps of this base model are required.We accomplish this with a fully end-to-end machine-learned model, consisting ofan improved geometry embedding architecture and an existing SE(3)-equivariantmodel to represent molecular orbitals. Combining this architecture withcontinuous sampling of geometries, we improve zero-shot accuracy by two ordersof magnitude compared to the state of the art. We extensively evaluate theaccuracy, scalability and limitations of our base model on a wide variety oftest systems.

Variational Monte Carlo on a Budget -- Fine-tuning pre-trained Neural Wavefunctions