GUE conjecture

GUE conjecture
Name	GUE conjecture
Field	Random matrix theory; Analytic number theory
Proposed	1970s
Proposer	Freeman Dyson; Hugh Montgomery (conjectural pairing)
Related	Gaussian unitary ensemble; Riemann zeta function; Montgomery pair correlation conjecture

GUE conjecture

The GUE conjecture posits that statistical properties of high-lying zeros of the Riemann zeta function (and of many families of L-functions) coincide with the eigenvalue statistics of the Gaussian unitary ensemble from Random matrix theory. It emerged from comparisons among conjectures of Hugh Montgomery, numerical work by Andrew Odlyzko, and theoretical insights of Freeman Dyson, linking ideas in Analytic number theory, Mathematical physics, and spectral theory. The conjecture informs research across Princeton University, Institute for Advanced Study, University of Cambridge, and many research groups studying the interplay of Paul Erdős-style experimental mathematics with rigorous analysis.

Statement of the conjecture

The conjecture asserts that the local spacing statistics of suitably normalized nontrivial zeros of the Riemann zeta function on the critical line agree with the local eigenvalue spacing statistics of large random Hermitian matrices drawn from the Gaussian unitary ensemble. More precisely, correlation functions (pair correlation, n-point correlation, spacing distributions) of zeros match those of GUE in the limit as height on the critical line tends to infinity. Montgomery formulated the pair correlation conjecture based on work with Hugh Montgomery and conjectural connections to the Montgomery–Odlyzko law; later formulations by Michael Berry and others extended this to full agreement of n-level correlations analogous to results in Dyson's circular ensembles.

Mathematical background

The conjecture sits at the intersection of Random matrix theory, spectral statistics of ensembles like Gaussian orthogonal ensemble and Gaussian symplectic ensemble, and classical problems in Analytic number theory about zero distributions of Dirichlet L-functions, Dedekind zeta functions, and automorphic L-functions associated to Atle Selberg's trace formula. It uses tools from the theory of large N limits in matrix models developed by Freeman Dyson, asymptotic analysis pioneered by Harold Widom and Craig Tracy, and the theory of orthogonal polynomials connected to work of Freeman Dyson and Miguel de Guzmán (note: de Guzmán as representative of orthogonal polynomial literature). Concepts from the spectral theory of the Laplace operator on arithmetic manifolds, informed by Selberg trace formula techniques and insights from Peter Sarnak and Dennis Hejhal, provide the analytic framework.

Evidence and numerical results

Extensive numerical computations by Andrew Odlyzko provided strong empirical evidence: high-precision computation of zeros near very large heights displayed spacing distributions matching GUE predictions. Odlyzko's work followed earlier heuristic suggestions by Hugh Montgomery and comparisons influenced by Freeman Dyson's observations about eigenvalue statistics. Subsequent numerical studies by groups at University of Illinois, Princeton University, and University of Bristol examined moments and spacing of zeros for various L-functions, including families studied by Brian Conrey, David Farmer, Jon Keating, and Nicolas Katz. These computations measured n-level correlations, nearest-neighbor spacings, and distribution tails, comparing to predictions derived from results by Tracy–Widom and correlations described by Mehta's matrix integral formulas.

Connections to number theory and Riemann zeta zeros

The conjecture provides a probabilistic model for fine-scale statistics of zeros of the Riemann zeta function and of automorphic L-functions associated to groups like GL(2) and GL(n). It complements the Riemann hypothesis by offering a statistical description of zero spacings assuming critical-line behavior; researchers such as Atle Selberg, Andrew Wiles, Peter Sarnak, and Henryk Iwaniec have used random matrix analogies to formulate moment conjectures and value-distribution predictions. Connections to the Montgomery pair correlation conjecture and to conjectures of Keating and Snaith on moments of L-functions link random matrix ensembles to arithmetic statistics studied at institutions such as Clay Mathematics Institute and Royal Society-affiliated groups.

Rigorous results and partial proofs

While the full conjecture remains open, rigorous progress has been made in model settings and averaged families. Results by Kurt Johansson, Terence Tao, and Van Vu on universality in random matrix theory establish that many ensembles share local statistics with GUE under mild hypotheses. In number theory, results by Conrey, Iwaniec, Duke, and Goldston prove variants of pair correlation or spacing results for families of L-functions averaged over parameters or levels. Work by Zelditch and Rudnick on quantum chaos and equidistribution provides conditional links between eigenvalue statistics of quantized systems and zeros of zeta-like functions.

Applications and implications

If true, the conjecture informs predictions for extreme value statistics of the Riemann zeta function relevant to conjectures by Keating and Snaith about moments and maximum size, and impacts heuristics about prime gaps via explicit formulas of G. H. Hardy-style analytic transforms. It influences statistical models used in cryptography and in theoretical physics contexts ranging from quantum chaos to mesoscopic systems studied at CERN-adjacent laboratories. The conjecture guides design of experiments in computational number theory at centers like University of Illinois Urbana–Champaign and shapes expectations about zeros of L-functions associated to modular forms studied by Andrew Wiles and Richard Taylor.

Open problems and research directions

Major open problems include proving the conjecture for the Riemann zeta function or for broad families of automorphic L-functions, establishing universality classes that rigorously connect arithmetic symmetries to specific random matrix ensembles, and deriving precise error terms for convergence of n-level correlations. Active directions involve extending universality proofs by Erdős, Yau, Tao, and Vu to arithmetic settings, refining moment conjectures by Keating and Snaith via input from Alain Connes-inspired noncommutative geometry, and developing new computational verifications à la Andrew Odlyzko at ever larger heights. Progress on these fronts would bridge work at Institute for Advanced Study, Princeton University, and research groups associated with European Research Council grants and national science foundations.

Category:Conjectures in number theory