Low-Order Spectral Moments

• Low-order spectral moments are integrals of a spectral density weighted by powers of frequency, capturing essential measures like total power, centroid, and spread.
• They play a vital role in diverse fields such as quantum many-body physics, random matrix theory, QCD, and signal processing by serving as foundational inputs for reconstruction and inference methods.
• These moments reveal universal statistical behaviors and finite-size corrections, providing practical insights for both theoretical modeling and experimental applications.

Low-order spectral moments are fundamental quantitative descriptors capturing integrals of power or density functions weighted by low powers of frequency, energy, or matrix eigenvalues. They form the backbone of statistical spectral analysis across quantum many-body physics, random matrix theory, quantum field theory, statistical signal processing, and condensed matter. These moments typically encode mean, variance, and higher-order cumulants of spectra, serving both as direct observables and as inputs to reconstructive and inferential schemes.

1. Definitions and Formalism

A low-order spectral moment is defined as an integral of a spectral density (or measure) against a monomial in its argument. For a spectral density $S(\omega)$ of a stationary process or operator, the $n$th spectral moment is

$$m_n = \int_{-\infty}^{\infty} \omega^n S(\omega) d\omega.$$

In quantum many-body systems and random matrix ensembles, analogous definitions are employed. For example:

• For the empirical spectral distribution of an $n\times n$ random matrix $A_n$ with eigenvalues $\{\lambda_j\}$,

$m_k = \frac{1}{n} \sum_{j=1}^n \lambda_j^k.$

• For power spectra estimated from signals, moments are often constraints or targets in estimation procedures (Georgiou et al., 2016).
• In collision-induced absorption, moments of the absorption profile $J(\nu)$ characterize the integrated band intensity and mean frequency (Chistikov et al., 17 Apr 2025).

Low-order typically denotes $n=0$ (total power or normalization), $n=1$ (centroid or dipole), and $n$0 (variance or spread).

2. Spectral Moments in Quantum Many-body Chaos: The SYK Model

In SYK-like disordered quantum systems, low-order spectral moments of the spectral form factor probe both macroscopic ramp/plateau structures and microscopic spectral noise (Legramandi et al., 2024):

• The $n$1th moment of the spectral form factor is $n$2, with $n$3.
• For $n$4 SYK, at leading order in $n$5 and for low $n$6,

$n$7

reflecting random-matrix-type universality. The leading $n$8 corrections scale as $n$9, and become $$m_n = \int_{-\infty}^{\infty} \omega^n S(\omega) d\omega.$$0 at $$m_n = \int_{-\infty}^{\infty} \omega^n S(\omega) d\omega.$$1, signaling breakdown of Gaussianity and the onset of rare fluctuations.

• For sparse SYK (random parameter reduction), these corrections are amplified as $$m_n = \int_{-\infty}^{\infty} \omega^n S(\omega) d\omega.$$2 with $$m_n = \int_{-\infty}^{\infty} \omega^n S(\omega) d\omega.$$3 the fraction of retained couplings.
• In the $$m_n = \int_{-\infty}^{\infty} \omega^n S(\omega) d\omega.$$4 (free) case, the moments exhibit an exponential ramp and dramatically enhanced noise due to an infinite-dimensional zero-mode manifold, departing from any RMT/Gaussian statistics.

These moment calculations reveal how random-matrix universality controls low-order statistics but is lost for higher moments due to edge-of-spectrum phenomena or decreased randomness (Legramandi et al., 2024).

3. Low-Order Spectral Moments in Random Matrix Theory

a. Wigner and Ginibre Ensembles

In classical Wigner matrices and non-Hermitian Ginibre ensembles, the low-order spectral moments underpin laws of large numbers and central limit theorems for individual matrix entries or empirical distributions (Duy, 2014, Byun et al., 2023, Akemann et al., 18 May 2025):

• For Wigner matrices, the even moments asymptotically follow the Catalan sequence, yielding the semicircle law, with explicit CLT corrections calculable for any fixed $$m_n = \int_{-\infty}^{\infty} \omega^n S(\omega) d\omega.$$5 (Duy, 2014).
• For the real Ginibre ensemble, exact expressions for the first few moments of the real spectrum are available (e.g., $$m_n = \int_{-\infty}^{\infty} \omega^n S(\omega) d\omega.$$6, $$m_n = \int_{-\infty}^{\infty} \omega^n S(\omega) d\omega.$$7) in terms of hypergeometric functions, as well as recursive relations and genus expansions showing both integer and half-integer $$m_n = \int_{-\infty}^{\infty} \omega^n S(\omega) d\omega.$$8 corrections (Byun et al., 2023).
• For non-Hermitian random matrices (complex and symplectic elliptic Ginibre, Wishart), the low-order moments are given by explicit formulas involving planar orthogonal polynomials and their recurrence coefficients. Holomorphic moments in the complex ensemble coincide with Hermitian ones up to a scaling by the non-Hermiticity parameter $$m_n = \int_{-\infty}^{\infty} \omega^n S(\omega) d\omega.$$9 (Akemann et al., 18 May 2025). Symplectic moments involve additional correction terms decomposable in terms of the complex ensemble.

b. Matrices with Structure

For symmetric random matrices with a rank-one pattern of variances (variance profile $n\times n$0), the low-order limiting moments are determined by noncrossing pairings and moments of $n\times n$1:

• All odd moments vanish.
• Even moments up to order $n\times n$2 are determined combinatorially via the Kreweras complement and moments $n\times n$3; for example,

$n\times n$4

(Preciado et al., 2014).

4. Low-Order Moments in QCD Spectral Analysis and $n\times n$5-Functions

Low-order spectral moments are central in precision QCD, both for extracting $n\times n$6 from hadronic $n\times n$7 decays and for understanding families of $n\times n$8-functions (Goulden et al., 2012, Beneke et al., 2012, Boito, 2013, Boito et al., 2020, Boito et al., 2010):

• In $n\times n$9-decay sum rules, moments

$A_n$0

are formed with various polynomial weights $A_n$1 (monomials, "pinched" weights, ALEPH/OPAL weights). The low-order choices $A_n$2, $A_n$3, $A_n$4 are most common.

• The perturbative expansions of these moments reveal strong sensitivity to the choice of weight; moments with $A_n$5 terms are destabilized by leading IR renormalons (u=2), and only moments $A_n$6, $A_n$7, or "pinched" weights with no $A_n$8 term exhibit robust, quickly convergent FOPT series (Beneke et al., 2012, Boito, 2013, Boito et al., 2020).
• In families of $A_n$9-functions, conjectured asymptotic expansions for low-order moments at the central point are given in terms of explicit combinatorial determinants (e.g., $\{\lambda_j\}$0), with arithmetic weights, and display perfect agreement with RMT to all orders once these are included (Goulden et al., 2012).

5. Signal Processing and Statistical Inference via Low-Order Spectral Moments

Low-order spectral moments serve as constraints or objectives in modern spectral estimation and signal processing frameworks:

• In likelihood analysis of power spectra, prescribed low-order moments (covariances or Fourier coefficients) serve as constraints in minimizing the Kullback-Leibler divergence to a prior spectrum. The solution is characterized as the unique minimizer in a convex dual problem, with explicit closed forms for classical or autoregressive-type priors (Georgiou et al., 2016).
• Two estimation schemes are widely compared in the context of electromagnetic reverberation: (1) direct time-domain extraction using finite differences, and (2) autocovariance-based methods. Both require correction for finite sampling, noise, and EMI, especially for $\{\lambda_j\}$1 moments (total power, centroid, variance) (Arnaut et al., 2024).
• Correction and normalization protocols can be explicitly computed and are analytic in key experimental parameters (sampling interval, bandwidth ratios, noise amplitudes).

6. Physical and Experimental Contexts

a. Quantum Corrections to Classical Moments

In molecular spectroscopy, low-order spectral moments encode integrated strengths and mean frequencies of collision-induced absorption bands. Exact quantum corrections to classical expressions can be computed via Wigner-Kirkwood expansions (up to $\{\lambda_j\}$2 or higher) or via path integral Monte Carlo, achieving sub-percent accuracy over a range of temperatures (Chistikov et al., 17 Apr 2025).

b. Observability and Protocols

Low-order moments can be efficiently estimated in laboratory and quantum information experiments. In quantum entanglement, a few low-order moments of the partially transposed density matrix allow tight bounds on the entanglement negativity, with explicit deterministic constructions and tight error analysis feasible with as few as three to five-copy measurement circuits (Carteret, 2016).

7. Classification, Universality, and Limitations

Low-order spectral moments reflect universal structures up to thresholds dictated by model complexity, system size, and degree of randomness:

• In SYK and related sparse models, universality holds for $\{\lambda_j\}$3; above that, deviations emerge from finite-$\{\lambda_j\}$4 (edge) corrections (Legramandi et al., 2024).
• In RMT and matrix ensembles, the low-order moments reproduce the universal limiting laws (semicircle, Marchenko-Pastur, etc.), with corrections calculable explicitly.
• In QCD and $\{\lambda_j\}$5-function analogues, low-order spectral moments are robust to perturbative uncertainties and model assumptions, provided weight functions are chosen to suppress duality violations and renormalon contributions.
• Systematic bias from finite sampling, noise, and non-ideal measurement can be analytically characterized and compensated at low orders (Arnaut et al., 2024).

Low-order moments, while powerful, are not generically sufficient for capturing non-Gaussianity, rare events, or fine spectral features, necessitating high-order corrections, careful statistical analysis, or direct estimation protocols when higher fidelity is required.