Variational estimation of generator invariant subspaces
The Journal of Chemical Physics · 28.07.2026 · 165 (4)
DOI: 10.1063/5.0341109 ↗Kurz gefasst
Fachartikel in der Zeitschrift The Journal of Chemical Physics, online erschienen am 23.07.2026. Verfasst von Luca Donati und 3 weiteren Autor:innen. Beteiligte Einrichtung: Freie Universität Berlin, Department of Mathematics and Computer Science, Arnimallee 22, D-14195 Berlin, Germany. Die Arbeit nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Berlin Mathematics Research Center MATH+, Deutsche Forschungsgemeinschaft, Bundesministerium für Forschung, Technologie und Raumfahrt.
Abstract (Originalsprache)
We present VEGIS (Variational Estimation of Generator Invariant Subspaces), a variational method to approximate invariant subspaces of the infinitesimal generator of reversible diffusion processes. The method represents a trial subspace by neural networks and optimizes a Dirichlet-form trace objective that can be evaluated using only equilibrium samples and gradients of the network outputs. After training, the learned trial space can be diagonalized to recover generator eigenfunctions and eigenvalues or transformed by PCCA+ to obtain membership functions associated with metastable sets. In addition, we introduce a VEGIS-driven sampling strategy in which a rough approximation of the dominant slow mode is used to modify the effective diffusivity while preserving the invariant density. Numerical results on low-dimensional and molecular systems demonstrate the accuracy and flexibility of VEGIS.
Autor:innen und Institutionen
- Freie Universität Berlin, Department of Mathematics and Computer Science, Arnimallee 22, D-14195 Berlin, Germany.
- Zuse Institute Berlin, Takustr. 7, D-14195 Berlin, Germany.
ORCID-Profile: Luca Donati, Surahit Chewle, Marcus Weber
Förderung
- Berlin Mathematics Research Center MATH+
- Deutsche Forschungsgemeinschaft
- Bundesministerium für Forschung, Technologie und Raumfahrt
Zitieren
APA
Donati, L., Safarov, F., Chewle, S., & Weber, M. (2026). Variational estimation of generator invariant subspaces. The Journal of Chemical Physics, 165(4). https://doi.org/10.1063/5.0341109
Vancouver
Donati L, Safarov F, Chewle S, Weber M. Variational estimation of generator invariant subspaces. The Journal of Chemical Physics. 2026;165(4). doi:10.1063/5.0341109
BibTeX
@article{Donati2026,
author = {Donati, Luca and Safarov, Fazil and Chewle, Surahit and Weber, Marcus},
title = {Variational estimation of generator invariant subspaces},
journal = {The Journal of Chemical Physics},
year = {2026},
volume = {165},
number = {4},
doi = {10.1063/5.0341109}
}
EUVIMED