An official website of the United States government
Abstract We construct a novel family of difference-permutation operators and prove that they are diagonalized by the wreath MacdonaldP-polynomials; the eigenvalues are written in terms of elementary symmetric polynomials of arbitrary degree. Our operators arise from integral formulas for the action of the horizontal Heisenberg subalgebra in the vertex representation of the corresponding quantum toroidal algebra.
more »
« less
The Large Deviation Principle for W-random spectral measures
The đť‘Š -random graphs provide a flexible framework for modeling large random networks. Using the Large Deviation Principle (LDP) for đť‘Š -random graphs from [19], we prove the LDP for the corresponding class of random symmetric Hilbert-Schmidt integral operators. Our main result describes how the eigenvalues and the eigenspaces of the integral operator are affected by large deviations in the underlying random graphon. To prove the LDP, we demonstrate continuous dependence of the spectral measures associated with integral operators on the corresponding graphons and use the Contraction Principle. To illustrate our results, we obtain leading order asymptotics of the eigenvalues of small-world and bipartite random graphs conditioned on atypical edge counts. These examples suggest several representative scenarios of how the eigenvalues and the eigenspaces are affected by large deviations. We discuss the implications of these observations for bifurcation analysis of Dynamical Systems and Graph Signal Processing.
more »
« less
- PAR ID:
- 10610590
- Publisher / Repository:
- Elsevier
- Date Published:
- Journal Name:
- Applied and Computational Harmonic Analysis
- Volume:
- 77
- Issue:
- C
- ISSN:
- 1063-5203
- Page Range / eLocation ID:
- 101756
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
This article shows that for a large class of discrete periodic Schrödinger operators, most wavefunctions resemble Bloch states. More precisely, we prove quantum ergodicity for a family of periodic Schrödinger operators H on periodic graphs. This means that most eigenfunctions of H on large finite periodic graphs are equidistributed in some sense, hence delocalized. Our results cover the adjacency matrix on Z^d, the triangular lattice, the honeycomb lattice, Cartesian products, and periodic Schrödinger operators on Z^d. The theorem applies more generally to any periodic Schrödinger operator satisfying an assumption on the Floquet eigenvalues.more » « less
-
We prove a sample-path large deviation principle (LDP) with sublinear speed for unbounded functionals of certain Markov chains induced by the Lindley recursion. The LDP holds in the Skorokhod space [Formula: see text] equipped with the [Formula: see text] topology. Our technique hinges on a suitable decomposition of the Markov chain in terms of regeneration cycles. Each regeneration cycle denotes the area accumulated during the busy period of the reflected random walk. We prove a large deviation principle for the area under the busy period of the Markov random walk, and we show that it exhibits a heavy-tailed behavior. Funding: The research of B. Zwart and M. Bazhba is supported by the Nederlandse Organisatie voor Wetenschappelijk Onderzoek [Grant 639.033.413]. The research of J. Blanchet is supported by the National Science Foundation (NSF) [Grants 1915967, 1820942, and 1838576] as well as the Defense Advanced Research Projects Agency [Grant N660011824028]. The research of C.-H. Rhee is supported by the NSF [Grant CMMI-2146530].more » « less
-
Abstract We introduce a framework for repurposing error estimators for sourceproblems to compute an estimator for the gap between eigenspaces and theirdiscretizations. Of interest are eigenspaces of finiteclusters of eigenvalues of unbounded nonselfadjoint linear operatorswith compact resolvent. Eigenspaces and eigenvalues of rationalfunctions of such operators are studied as a first step.Under an assumption of convergence of resolvent approximations inthe operator norm and an assumption on global reliability of sourceproblem error estimators, we show that the gap in eigenspaceapproximations can be bounded by a globally reliable and computableerror estimator. Also included are applications of the theoreticalframework to first-order system least squares(FOSLS) discretizations and discontinuous Petrov–Galerkin (DPG)discretizations, both yielding new estimators for the error gap.Numerical experiments with a selfadjoint model problem and with aleaky nonselfadjoint waveguide eigenproblem show that adaptive algorithmsusing the new estimators give refinement patternsthat target the cluster as a whole instead of individual eigenfunctions.more » « less
-
Designing effective positional encodings for graphs is key to building powerful graph transformers and enhancing message-passing graph neural networks. Although widespread, using Laplacian eigenvectors as positional encodings faces two fundamental challenges: (1) \emph{Non-uniqueness}: there are many different eigendecompositions of the same Laplacian, and (2) \emph{Instability}: small perturbations to the Laplacian could result in completely different eigenspaces, leading to unpredictable changes in positional encoding. Despite many attempts to address non-uniqueness, most methods overlook stability, leading to poor generalization on unseen graph structures. We identify the cause of instability to be a ``hard partition'' of eigenspaces. Hence, we introduce Stable and Expressive Positional Encodings (SPE), an architecture for processing eigenvectors that uses eigenvalues to ``softly partition'' eigenspaces. SPE is the first architecture that is (1) provably stable, and (2) universally expressive for basis invariant functions whilst respecting all symmetries of eigenvectors. Besides guaranteed stability, we prove that SPE is at least as expressive as existing methods, and highly capable of counting graph structures. Finally, we evaluate the effectiveness of our method on molecular property prediction, and out-of-distribution generalization tasks, finding improved generalization compared to existing positional encoding methods.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript
- Journal Article:
- https://doi.org/10.1016/j.acha.2025.101756
