Ten Digits on a Train: AI-Assisted Verification of Two Eigenvalue Problems

DOI10.48550/arXiv.2606.23821MaRDI QIDQ7270032arXivFDO

Authors Matthew J. Colbrook

Publication date 22 June 2026

Published in arXiv

Tldr We describe two case studies of verifying eigenvalue computations with AI assistance, which achieved ten-digit accuracy. In both cases, we reformulated the problem to exploit AI's strengths while avoiding its weaknesses.

Full work available at URL https://doi.org/10.48550/arXiv.2606.23821

Mathematics Subject Classification ID

Eigenvalues, estimation of eigenvalues, upper and lower bounds of ordinary differential operators (34L15) Numerical approximation of eigenvalues and of other parts of the spectrum of ordinary differential operators (34L16) Algorithms with automatic result verification (65G20) Numerical solution of eigenvalue problems involving ordinary differential equations (65L15) Artificial intelligence (68T99) Nonselfadjoint operator theory in quantum theory including creation and destruction operators (81Q12)

arXiv classification

cs.AIcs.HCmath.NAmath.SP

Cited in

(1)

• A new analysis of the randomly pivoted Cholesky algorithm

This page was built for publication: Ten Digits on a Train: AI-Assisted Verification of Two Eigenvalue Problems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7270032)