In-span learning: adapting reduced-order models using their own predictions

DOI10.48550/arXiv.2607.02937MaRDI QIDQ7283960arXivFDO

Authors Laura Balzano, Karthik Duraisamy, Amirpasha Hedayat

Publication date 3 July 2026

Published in arXiv

Tldr This paper presents a method called "in-span learning" that adapts reduced-order models online by leveraging their own predictions to improve accuracy when faced with out-of-span dynamics. The key result is that in-span learning uses the model's trajectory-informed spectral preconditioner to reweight and realign its basis, enabling it to better absorb future corrections.

Full work available at URL https://doi.org/10.48550/arXiv.2607.02937

arXiv classification

cs.CEcs.LGmath.NA

This page was built for publication: In-span learning: adapting reduced-order models using their own predictions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7283960)