Correlated random band matrices: localization-delocalization transitions
Martin Janssen (Bochum University), and Krystian Pracz (University of, Cologne)

TL;DR
This paper investigates how correlations in random band matrices influence eigenvector behavior, revealing a localization-delocalization transition akin to quantum Hall phenomena when certain parameters scale with matrix size.
Contribution
It introduces a correlated random band matrix model and demonstrates a localization-delocalization transition driven by the correlation parameter's scaling.
Findings
Localization-delocalization transition observed
Rich phenomenology due to correlations
Transition resembles quantum Hall effect
Abstract
We study the statistics of eigenvectors in correlated random band matrix models. These models are characterized by two parameters, the band width B(N) of a Hermitian N times N matrix and the correlation parameter C(N) describing correlations of matrix elements along diagonal lines. The correlated band matrices show a much richer phenomenology than models without correlation as soon as the correlation parameter scales sufficiently fast with matrix size. In particular, for B(N) and C(N) increasing like the square root of N the model shows a localization-delocalization transition of the quantum Hall type.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
