What question did this study set out to answer?

July 26, 2026The Journal of Chemical PhysicsOpen Access

Variational estimation of generator invariant subspaces

View Full Paper

Ask AI

Bookmark

Authors

LDLuca DonatiZuse Institute Berlin SFSafarov FKhZuse Institute Berlin SCSurahit ChewleZuse Institute Berlin

Discussion

Member takes

Overview

Randomized trial demonstrates a variational method for approximating invariant subspaces in diffusion processes, suggesting improvements in sampling strategies.

Key Points

This study aims to develop and validate VEGIS, a variational method for approximating invariant subspaces of reversible diffusion processes.
Utilizes neural networks to represent trial subspaces
Optimizes a Dirichlet-form trace objective using equilibrium samples and network gradients
Introduces VEGIS-driven sampling to modify effective diffusivity while preserving invariant density.
Accuracy and flexibility of VEGIS are validated through numerical results on low-dimensional and molecular systems.
VEGIS successfully diagonalizes the learned trial space to recover generator eigenfunctions and eigenvalues.
Membership functions associated with metastable sets are obtained using PCCA+ transformations.

Cite This Study

Donati et al. (2026) studied this question.

synapsesocial.com/papers/6a65a468d3aea3239cd76faa https://doi.org/10.1063/5.0341109

View Full Paper

Ask AI

Bookmark

July 26, 2026The Journal of Chemical PhysicsOpen Access

Variational estimation of generator invariant subspaces

View Full Paper

Ask AI

Bookmark

Authors

LDLuca DonatiZuse Institute Berlin SFSafarov FKhZuse Institute Berlin SCSurahit ChewleZuse Institute Berlin

Discussion

Member takes

Overview

Randomized trial demonstrates a variational method for approximating invariant subspaces in diffusion processes, suggesting improvements in sampling strategies.

Key Points

This study aims to develop and validate VEGIS, a variational method for approximating invariant subspaces of reversible diffusion processes.
Utilizes neural networks to represent trial subspaces
Optimizes a Dirichlet-form trace objective using equilibrium samples and network gradients
Introduces VEGIS-driven sampling to modify effective diffusivity while preserving invariant density.
Accuracy and flexibility of VEGIS are validated through numerical results on low-dimensional and molecular systems.
VEGIS successfully diagonalizes the learned trial space to recover generator eigenfunctions and eigenvalues.
Membership functions associated with metastable sets are obtained using PCCA+ transformations.

Cite This Study

Donati et al. (2026) studied this question.

synapsesocial.com/papers/6a65a468d3aea3239cd76faa https://doi.org/10.1063/5.0341109

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark