Persistence diagrams of eigenvalue processes: definitions, universality, and detection of the BBP transition

Read full text Summarise with VIRAT
Authors: Matthew Loftus
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

We introduce the eigenvalue process persistence diagram, a new topological invariant for parametric matrix families. For a matrix path M(t), eigenvalues trace curves with avoided crossings; we define three persistence diagram formulations capturing these dynamics via sublevel-set persistence of pairwise gap functions. The strongest formulation (per-pair gap union) achieves coefficient of variation below 0.05 across four test cases (GOE-to-GUE, GOE-to-diagonal, GOE-to-GOE null, BBP rank-1 spiked model), establishing universality. The diagram discriminates repulsive from non-repulsive dynamics at 10.8 sigma and detects the Baik-Ben Arous-Peche spectral phase transition with ensemble bias 0.003 from the theoretical threshold. Persistence entropy scales as PE ~ 1.55 ln(n) with R^2 = 0.9998 over matrix sizes n = 20 to 60.

Fields of Study & Subjects

Topological and Geometric Data Analysis Random Matrices and Applications Quantum chaos and dynamical systems

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
30 March 2026
Language
EN
Document type
Preprint