Computing with traceable tensor networks

DOI10.48550/arXiv.2608.02849MaRDI QIDQ7284604arXivFDO

Authors Daniele Venturi, Sarah Ellwein

Publication date 3 August 2026

Published in arXiv

Tldr We present a new SVD-based method for computing with traceable tensor networks, extending classical hierarchical techniques to general connectivity and enabling step-rounding time integration in graph format. We demonstrate the approach on multivariate functions and the Fokker-Planck equation, achieving comparable or better accuracy than traditional formats at lower cost and degrees of freedom.

Full work available at URL https://doi.org/10.48550/arXiv.2608.02849

arXiv classification

math.NAphysics.comp-ph

Cites work

• A review of low-rank methods for time-dependent kinetic simulations
• Adaptive integration of nonlinear evolution equations on tensor manifolds
• Hierarchical Singular Value Decomposition of Tensors
• Implicit integration of nonlinear evolution equations on tensor manifolds
• Numerical methods for kinetic equations
• On Algorithms for and Computing with the Tensor Ring Decomposition
• Rank-adaptive tensor methods for high-dimensional nonlinear PDEs
• Solving high-dimensional Fokker-Planck equation with functional hierarchical tensor
• Spectral methods for nonlinear functionals and functional differential equations
• SVD-based algorithms for fully-connected tensor network decomposition
• SVD-based algorithms for tensor wheel decomposition
• Tensor approximation of functional differential equations
• Tensor methods for the Boltzmann-BGK equation
• Tensor-train decomposition
• The Boltzmann equation and its applications
• The numerical approximation of nonlinear functionals and functional differential equations
• The trouble with tensor ring decompositions

This page was built for publication: Computing with traceable tensor networks

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7284604)