Charting electronic-state manifolds across molecules with multi-state learning and gap-driven dynamics via efficient and robust active learning

Abstract We present a robust protocol for affordable learning of electronic states to accelerate photophysical and photochemical molecular simulations. The protocol solves several issues precluding the widespread use of machine learning (ML) in excited-state simulations. We introduce a novel physics...

Full description

Saved in:
Bibliographic Details
Main Authors: Mikołaj Martyka, Lina Zhang, Fuchun Ge, Yi-Fan Hou, Joanna Jankowska, Mario Barbatti, Pavlo O. Dral
Format: Article
Language:English
Published: Nature Portfolio 2025-05-01
Series:npj Computational Materials
Online Access:https://doi.org/10.1038/s41524-025-01636-z
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850272926222254080
author Mikołaj Martyka
Lina Zhang
Fuchun Ge
Yi-Fan Hou
Joanna Jankowska
Mario Barbatti
Pavlo O. Dral
author_facet Mikołaj Martyka
Lina Zhang
Fuchun Ge
Yi-Fan Hou
Joanna Jankowska
Mario Barbatti
Pavlo O. Dral
author_sort Mikołaj Martyka
collection DOAJ
description Abstract We present a robust protocol for affordable learning of electronic states to accelerate photophysical and photochemical molecular simulations. The protocol solves several issues precluding the widespread use of machine learning (ML) in excited-state simulations. We introduce a novel physics-informed multi-state ML model that can learn an arbitrary number of excited states across molecules, with accuracy better or similar to the accuracy of learning ground-state energies, where information on excited-state energies improves the quality of ground-state predictions. We also present gap-driven dynamics for accelerated sampling of the small-gap regions, which proves crucial for stable surface-hopping dynamics. Together, multi-state learning and gap-driven dynamics enable efficient active learning, furnishing robust models for surface-hopping simulations and helping to uncover long-time-scale oscillations in cis-azobenzene photoisomerization. Our active-learning protocol includes sampling based on physics-informed uncertainty quantification, ensuring the quality of each adiabatic surface, low error in energy gaps, and precise calculation of the hopping probability.
format Article
id doaj-art-04bb013473d04589a6bea6f09e7ee47c
institution OA Journals
issn 2057-3960
language English
publishDate 2025-05-01
publisher Nature Portfolio
record_format Article
series npj Computational Materials
spelling doaj-art-04bb013473d04589a6bea6f09e7ee47c2025-08-20T01:51:39ZengNature Portfolionpj Computational Materials2057-39602025-05-0111111210.1038/s41524-025-01636-zCharting electronic-state manifolds across molecules with multi-state learning and gap-driven dynamics via efficient and robust active learningMikołaj Martyka0Lina Zhang1Fuchun Ge2Yi-Fan Hou3Joanna Jankowska4Mario Barbatti5Pavlo O. Dral6University of Warsaw, Faculty of ChemistryState Key Laboratory of Physical Chemistry of Solid Surfaces, Department of Chemistry, College of Chemistry and Chemical Engineering, and Fujian Provincial Key Laboratory of Theoretical and Computational Chemistry, Xiamen UniversityState Key Laboratory of Physical Chemistry of Solid Surfaces, Department of Chemistry, College of Chemistry and Chemical Engineering, and Fujian Provincial Key Laboratory of Theoretical and Computational Chemistry, Xiamen UniversityState Key Laboratory of Physical Chemistry of Solid Surfaces, Department of Chemistry, College of Chemistry and Chemical Engineering, and Fujian Provincial Key Laboratory of Theoretical and Computational Chemistry, Xiamen UniversityUniversity of Warsaw, Faculty of ChemistryAix Marseille University, CNRS, ICRState Key Laboratory of Physical Chemistry of Solid Surfaces, Department of Chemistry, College of Chemistry and Chemical Engineering, and Fujian Provincial Key Laboratory of Theoretical and Computational Chemistry, Xiamen UniversityAbstract We present a robust protocol for affordable learning of electronic states to accelerate photophysical and photochemical molecular simulations. The protocol solves several issues precluding the widespread use of machine learning (ML) in excited-state simulations. We introduce a novel physics-informed multi-state ML model that can learn an arbitrary number of excited states across molecules, with accuracy better or similar to the accuracy of learning ground-state energies, where information on excited-state energies improves the quality of ground-state predictions. We also present gap-driven dynamics for accelerated sampling of the small-gap regions, which proves crucial for stable surface-hopping dynamics. Together, multi-state learning and gap-driven dynamics enable efficient active learning, furnishing robust models for surface-hopping simulations and helping to uncover long-time-scale oscillations in cis-azobenzene photoisomerization. Our active-learning protocol includes sampling based on physics-informed uncertainty quantification, ensuring the quality of each adiabatic surface, low error in energy gaps, and precise calculation of the hopping probability.https://doi.org/10.1038/s41524-025-01636-z
spellingShingle Mikołaj Martyka
Lina Zhang
Fuchun Ge
Yi-Fan Hou
Joanna Jankowska
Mario Barbatti
Pavlo O. Dral
Charting electronic-state manifolds across molecules with multi-state learning and gap-driven dynamics via efficient and robust active learning
npj Computational Materials
title Charting electronic-state manifolds across molecules with multi-state learning and gap-driven dynamics via efficient and robust active learning
title_full Charting electronic-state manifolds across molecules with multi-state learning and gap-driven dynamics via efficient and robust active learning
title_fullStr Charting electronic-state manifolds across molecules with multi-state learning and gap-driven dynamics via efficient and robust active learning
title_full_unstemmed Charting electronic-state manifolds across molecules with multi-state learning and gap-driven dynamics via efficient and robust active learning
title_short Charting electronic-state manifolds across molecules with multi-state learning and gap-driven dynamics via efficient and robust active learning
title_sort charting electronic state manifolds across molecules with multi state learning and gap driven dynamics via efficient and robust active learning
url https://doi.org/10.1038/s41524-025-01636-z
work_keys_str_mv AT mikołajmartyka chartingelectronicstatemanifoldsacrossmoleculeswithmultistatelearningandgapdrivendynamicsviaefficientandrobustactivelearning
AT linazhang chartingelectronicstatemanifoldsacrossmoleculeswithmultistatelearningandgapdrivendynamicsviaefficientandrobustactivelearning
AT fuchunge chartingelectronicstatemanifoldsacrossmoleculeswithmultistatelearningandgapdrivendynamicsviaefficientandrobustactivelearning
AT yifanhou chartingelectronicstatemanifoldsacrossmoleculeswithmultistatelearningandgapdrivendynamicsviaefficientandrobustactivelearning
AT joannajankowska chartingelectronicstatemanifoldsacrossmoleculeswithmultistatelearningandgapdrivendynamicsviaefficientandrobustactivelearning
AT mariobarbatti chartingelectronicstatemanifoldsacrossmoleculeswithmultistatelearningandgapdrivendynamicsviaefficientandrobustactivelearning
AT pavloodral chartingelectronicstatemanifoldsacrossmoleculeswithmultistatelearningandgapdrivendynamicsviaefficientandrobustactivelearning