The Gaussian Wave for Graphs of Finite Cone Type

Theo McKenzie · 2026

Read The Gaussian Wave for Graphs of Finite Cone Type by Theo McKenzie (2026) 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-2603-04386.txt.

Introduction

Dembo and McKenzie prove the Gaussian wave is the only typical process on finite cone type trees with mild expansion. Their covariance comes from the Green's function on tree vertices. The result generalizes Backhausz and Szegedy, who proved it for regular trees of degree at least 3. A reduction to inner products with Green's function columns, which are easy to compute, drives the proof. Consequences follow for random bipartite biregular graphs and for generic configuration models including random lifts, where local eigenvector distributions approach the Gaussian wave.

Catalog record

CREATORTheo McKenzie
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE81530
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.04386

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The Gaussian Wave for Graphs of Finite Cone Type Theo McKenzie Abstract We show that for any infinite tree of finite cone type satisfying a mild expansion condition, the only typical process on its vertices with covariance induced by the Green’s function is the Gaussian wave. This generalizes a result of Backhausz and Szegedy, who proved this for the infinite regular tree of degree d ≥ 3 d\geq 3 . We do this by giving a reduction to a statement concerning the distribution of the inner product of our process with columns of the Green’s function, which in turn are straightforward to calculate. As a consequence, for random bipartite biregular graphs, the distribution of local neighborhoods of eigenvectors must approximate the Gaussian wave. Moreover, for generic configuration models including random lifts, the local distribution of a uniformly chosen eigenvector from any arbitrarily small spectral window likewise converges to the Gaussian wave. Stanford University [email protected] Stanford University [email protected] 1 Introduction Berry’s random wave conjecture is a foundational prediction in the study of quantum chaotic systems [ 10 , 11 ] . It theorizes that the local statistics of eigenfunctions converge to those of a Gaussian random wave. The discrete version of this question is well studied, and large sparse graphs have long been used as a system of quantum chaos [ 32 ] . Whereas a general heuristic is that these graphs will have their spectral measure converge to their local weak limit, the idea of Berry’s conjecture is that, for chaotic models, we can move past this weak convergence to a stronger notion dealing with local statistics. We will make this explicit through the Green’s function . Define \mathbb{C}^{+}\mathrel{\mathop{:}}=\{z\in\mathbb{C}:\Im[z]>0\} and its closure \overline{\mathbb{C}^{+}}\mathrel{\mathop{:}}=\{z\in\mathbb{C}:\Im[z]\geq 0\} . We define the Green’s function for a graph \mathcal{G} with adjacency operator A_{\mathcal{G}} and z\in\mathbb{C}^{+} as G(z,\mathcal{G})\mathrel{\mathop{:}}=[A_{\mathcal{G}}-z\mathds{I}]^{-1}. We can then extend the definition to the real line by saying that for \lambda\in\mathbb{R} , G(\lambda,\mathcal{G})\mathrel{\mathop{:}}=\lim_{\eta\rightarrow 0^{+}}G(\lambda+\mathrm{i}\eta,\mathcal{G}). If \mathcal{G} is a finite graph on vertex set [N] , we define the Stieltjes transform m(z,\mathcal{G}) to be the normalized trace of the Green’s function, \frac{1}{N}\sum_{x\in[N]}G_{xx} . We can think of this as the expected value of the diagonal entry of the Green’s function for a vertex selected uniformly at random. More generally, if \mathcal{G} is a unimodular random graph (see [ 2 ] for a thorough overview), we can take the expectation over the root of the graph. Therefore, we define the Stieltjes transform of \mathcal{G} for z\in\overline{\mathbb{C}^{+}} to be m(z,\mathcal{G})\mathrel{\mathop{:}}=\mathbb{E}_{o}[G(z,\mathcal{G})_{oo}], where \mathbb{E}_{o}[\cdot] refers to the expectation over the randomly selected root according to the unimodular measure. We say that \lambda\in\mathbb{R} is in the purely absolutely continuous spectrum if 0<\Im(m(\lambda))<\infty. (1.1) Whereas point spectrum corresponds to localized states, the absolutely continuous spectrum corresponds to dynamical delocalization of the quantum operator (e.g. [ 1 , Theorem 2.6] ), so it is where we witness chaotic behavior of eigenfunctions from Berry’s conjecture. The impressive work of Backhausz and Szegedy [ 8 ] proved that every eigenvector, besides the trivial eigenvector of eigenvalue d , has the local statistics of a Gaussian process on the infinite regular tree \mathcal{T}_{d} (see 1.4 ). As the spectrum of the adjacency operator on the infinite regular tree is purely absolutely continuous, this gives evidence towards Berry’s conjecture. The d -regular graph benefits from a key simplifying feature: its local weak limit is the infinite regular tree, which has a large automorphism group. This is due to it being vertex transitive, and, more strongly, distance regular (for any pair of paths of the same length, there is an automorphism sending one path to the other). Any limiting wave must respect this symmetry, and this invariance plays a decisive role in identifying and solving the associated formulas of a limiting process. Nevertheless, there is no reason to expect the symmetry of the regular tree to be essential to chaotic behavior, and we prove a result toward Berry’s conjecture for a broader class of random graph models generalizing the random regular graph. Specifically, for a large family of configuration models—random graphs with a prescribed degree distribution—we show that eigenvectors exhibit Gaussian-type statistics, driven solely by local expansion properties. A principal innovation is a structural decomposition of the Gaussian process on the infinite tree via the Green’s function, which makes this mechanism transparent. The Gaussian wave can be realized as a Gaussian process obtained by pushing forward i.i.d. Gaussian noise through the resolvent. Namely, for z=\lambda+\mathrm{i}\eta\in\overline{\mathbb{C}^{+}} , graph \mathcal{T} , and a family \{\xi_{x}\}_{x\in V(\mathcal{T})} of i.i.d. random variables with \xi_{x}\sim\mathcal{N}(0,1) , the Gaussian wave \Psi^{z}(\mathcal{T}) (see 1.1 ) admits the representation \Psi^{z}(\mathcal{T})\overset{d}{=}\sqrt{\eta}\sum_{x\in V(\mathcal{T})}\xi_{x}\,G(z,\mathcal{T})_{x}, (1.2) where G(z,\mathcal{T})_{x}(\cdot)\mathrel{\mathop{:}}=G(z,\mathcal{T})_{\cdot x} denotes the column of the kernel of G(z,\mathcal{T}) corresponding to x\in V(\mathcal{T}) . To show any candidate distribution must be Gaussian, we examine its entropy, and one key insight is that to find the entropy of the Gaussian wave, it is sufficient to consider moments of the form \{\langle\Psi^{z}(\mathcal{T}),G(z,\mathcal{T})_{x}\rangle\}_{x\in V(\mathcal{T})}. Of course, because \Psi^{z} has the decomposition from ( 1.2 ), the above is easy to compute. While the Green’s function representation of covariance is known (see [ 26 , Section 1.1] ), treating this decomposition as a primary analytical tool allows for a direct comparison between the cases z\in\mathbb{C}^{+} and z\in\mathbb{R} and provides a transparent framework for analyzing the entropy of the Gaussian wave. ( 1.2 ) can be compared to the heuristic related to Berry’s conjecture that on a hyperbolic manifold, the Fourier coefficients of the Poisson kernel should be i.i.d. Gaussian [ 6 , 35 ] . Indeed, we would expect this type of decomposition to hold in many models and to be useful in future work. In order to precisely state our results, we give the following definition of an eigenvector process with covariance determined by the Green’s function. This was defined for the regular tree in [ 21 ] , but a more general definition is as follows. Definition 1.1 . An eigenvector process \Phi^{z}(\mathcal{T}) on the (potentially infinite) graph \mathcal{T} for z\in\overline{\mathbb{C}^{+}} , is a real-valued process on V(\mathcal{T}) that is automorphism invariant, is centered, and for every x,y\in V(\mathcal{T}) , \mathbb{E}[\Phi^{z}(\mathcal{T})_{x}\Phi^{z}(\mathcal{T})_{y}]=\Im[G(z,\mathcal{T})_{xy}] . The Gaussian wave \Psi^{z}(\mathcal{T}) is the unique real valued Gaussian eigenvector process. By the fact that z\in\overline{\mathbb{C}^{+}} , the imaginary part of the Green’s function is positive semidefinite, and therefore forms a valid covariance matrix. Moreover, ( 1.2 ) has the correct covariance by the Ward identity ( B.4 ). For z\in\mathbb{R} , the Gaussian wave is nontrivial with finite variance if and only if z is in the purely absolutely continuous spectrum. For families of finite random graphs, we will be interested in the Gaussian wave on their universal cover, defined as follows. Definition 1.2 . For any finite graph \mathcal{G} , its universal cover \mathcal{T} is the tree whose vertices are nonbacktracking walks in \mathcal{G} , with adjacency given by one-step extensions of such walks. There is a canonical projection \pi:V(\mathcal{T})\to V(\mathcal{G}) that maps each walk to its endpoint in \mathcal{G} . To obtain a unimodular random rooted tree, we select a root o uniformly at random from V(\mathcal{G}) and consider the component of \mathcal{T} corresponding to walks starting at o . A labeling of \mathcal{G} is a function

Read the complete text