Laplace Approximation For Tensor Train Kernel Machines In System Identification footnoteinfo This work is supported by the Dutch Research Council (NWO).

Albert Saiapin, Kim Batselier · 2025

Read Laplace Approximation For Tensor Train Kernel Machines In System Identification footnoteinfo This work is supported by the Dutch Research Council (NWO). by Albert Saiapin, Kim Batselier (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-02532.txt.

Description

Albert Saiapin and Kim Batselier work on faster stand ins for Gaussian process regression, a flexible but costly method. Tensor network methods offer a fix: they use a huge number of basis functions without a huge computing bill. Turning that trick into a fully probabilistic model raises fresh design problems, though.

Their 2025 study applies the Laplace approximation, a standard way to approximate a probability distribution, to tensor train kernel machines. The setting is system identification, where a model of a process is learned from data. A sound probabilistic version would give not just predictions, but honest uncertainty around them. With sound error bars, teams can trust the twin to guide real tests on new cell types.

Catalog record

CREATORAlbert Saiapin, Kim Batselier
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE33 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2512-02532

Browse the full library · About the Tower of Babel