EUVIMEDEUVIMED entdecken
Crossref

ElemeNet: Multiscale Molecular Machine Learning with Uncertainty Quantification across the Periodic Table

Jacob W Toney iD, Samir Darouich iD, Yiran Wang iD, Aaron G Garrison iD, Johannes Kästner iD, Heather J Kulik iD

Journal of Chemical Information and Modeling · 26.10.2026 · 66 (20) · 13252-13269

DOI: 10.1021/acs.jcim.6c02178 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Chemical Information and Modeling, online erschienen am 01.10.2026. Verfasst von Jacob W Toney und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts02139, United States. Hauptthemen laut MEDLINE: Machine Learning, Software. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 121 Literaturangaben. Gefördert von: Alfred P. Sloan Foundation, Dow Chemical Company, Claussen-Simon-Stiftung.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Advances in deep learning architectures and representations have enabled ML-driven chemical property prediction, but state-of-the-art (SOTA) models have remained largely confined to independent codebases and lack support for diverse chemical species. This work introduces ElemeNet, a unified, general-purpose software package for molecular machine learning. The ElemeNet software package enables the training of advanced ML models for diverse properties and data sets with an enlarged range of elemental compositions. We define molecular representations compatible with elements 1–100, supporting diverse organometallic and biological systems in addition to organic chemistry already well served by the Chemprop ML toolkit. As well as more common atom-, bond-, and molecule-level predictions, we introduce and allow for moiety-level predictions. We also natively define optional conditioning on charge and spin states. Advanced E(3)-equivariant and transformer architectures are supported in addition to 2D models, with all classes including built-in uncertainty quantification through deterministic and statistical measures. We benchmark our protocols for ML model training against representative data sets from organic, inorganic, coordination, and biological chemistry, achieving competitive and SOTA performance relative to literature baselines and favorable scaling to millions of molecules. The entire workflow is exposed through a concise command-line interface, lowering the barrier to entry for nonexpert users. We anticipate ElemeNet will empower noncomputational researchers to leverage modern deep learning methods across the chemical and physical sciences.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Machine LearningSoftwareUncertaintyModels, Molecular

Autor:innen und Institutionen

  • Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts02139, United States.
  • Center for Computational Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts02139, United States.
  • Institute for Theoretical Chemistry, University of Stuttgart, 70569Stuttgart, Germany.
  • Institute for Artificial Intelligence, University of Stuttgart, 70569Stuttgart, Germany.
  • Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts02139, United States.

ORCID-Profile: Jacob W Toney, Samir Darouich, Yiran Wang, Aaron G Garrison, Johannes Kästner, Heather J Kulik

Förderung

  • Alfred P. Sloan Foundation
  • Dow Chemical Company
  • Claussen-Simon-Stiftung
  • Deutsche Forschungsgemeinschaft
  • Ministerium f?r Wirtschaft, Arbeit und Wohnungsbau Baden-W?rttemberg
  • International Max Planck Research School for Intelligent Systems
  • Massachusetts Institute of Technology School of Engineering

Zitieren

APA

Toney, J. W., Darouich, S., Wang, Y., Garrison, A. G., Kästner, J., & Kulik, H. J. (2026). ElemeNet: Multiscale Molecular Machine Learning with Uncertainty Quantification across the Periodic Table. Journal of Chemical Information and Modeling, 66(20), 13252-13269. https://doi.org/10.1021/acs.jcim.6c02178

Vancouver

Toney JW, Darouich S, Wang Y, Garrison AG, Kästner J, Kulik HJ. ElemeNet: Multiscale Molecular Machine Learning with Uncertainty Quantification across the Periodic Table. Journal of Chemical Information and Modeling. 2026;66(20):13252-13269. doi:10.1021/acs.jcim.6c02178

BibTeX
@article{Toney2026,
  author = {Toney, Jacob W and Darouich, Samir and Wang, Yiran and Garrison, Aaron G and Kästner, Johannes and Kulik, Heather J},
  title = {ElemeNet: Multiscale Molecular Machine Learning with Uncertainty Quantification across the Periodic Table},
  journal = {Journal of Chemical Information and Modeling},
  year = {2026},
  volume = {66},
  number = {20},
  pages = {13252--13269},
  doi = {10.1021/acs.jcim.6c02178}
}
Kostenlos starten →