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

We study a fundamental problem in computational chemistry known as molecularconformation generation, trying to predict stable 3D structures from 2Dmolecular graphs. Existing machine learning approaches usually first predictdistances between atoms and then generate a 3D structure satisfying thedistances, where noise in predicted distances may induce extra errors during 3Dcoordinate generation. Inspired by the traditional force field methods formolecular dynamics simulation, in this paper, we propose a novel approachcalled ConfGF by directly estimating the gradient fields of the log density ofatomic coordinates. The estimated gradient fields allow directly generatingstable conformations via Langevin dynamics. However, the problem is verychallenging as the gradient fields are roto-translation equivariant. We noticethat estimating the gradient fields of atomic coordinates can be translated toestimating the gradient fields of interatomic distances, and hence develop anovel algorithm based on recent score-based generative models to effectivelyestimate these gradients. Experimental results across multiple tasks show thatConfGF outperforms previous state-of-the-art baselines by a significant margin.

Learning Gradient Fields for Molecular Conformation Generation