We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

q-bio.QM

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Quantitative Biology > Quantitative Methods

Title: Molecular modeling with machine-learned universal potential functions

Abstract: Molecular modeling is an important topic in drug discovery. Decades of research have led to the development of high quality scalable molecular force fields. In this paper, we show that neural networks can be used to train a universal approximator for energy potential functions. By incorporating a fully automated training process we have been able to train smooth, differentiable, and predictive potential functions on large-scale crystal structures. A variety of tests have also been performed to show the superiority and versatility of the machine-learned model.
Subjects: Quantitative Methods (q-bio.QM); Machine Learning (cs.LG)
Cite as: arXiv:2103.04162 [q-bio.QM]
  (or arXiv:2103.04162v2 [q-bio.QM] for this version)

Submission history

From: Junqiu Wu [view email]
[v1] Sat, 6 Mar 2021 17:36:39 GMT (1460kb)
[v2] Mon, 19 Apr 2021 06:30:52 GMT (1966kb,D)

Link back to: arXiv, form interface, contact.