betaFeedback

WorksMathematics

Almost-optimal bulk regularity conditions in the CLT for GββE

Abstract

We consider linear statistics ∑k=1Nf(xk)\sum_{k=1}^N f(x_k) of a Gaussian \b-ensemble. Following the work of \cite{LaSo22}, which was concerned only with Wigner matrices, we extend their results for general \b. In particular, we show that the CLT holds for any f \in H^{1/2 + \e}. The novel input which was not used in \cite{LaSo22} is the Helffer-Sjöstrand representation for the covariance, first appearing the random matrix context in \cite{ErYa15}, coupled with characteristic flow estimates for DBM, first appearing in \cite{Bo22}.

Introduction

Fluctuations of linear statistics

1.1. Fluctuations of linear statistics. In this work, we are concerned with the fluctuations of linear statistics FN(x)=∑k=1Nf(xk)F_N(x) = \sum_{k=1}^{N} f(x_k) of Gaussian β\beta-ensembles (GβEG\beta E),

dν(x)=dνN,β(x):=1ZN,βexp⁡(−Nβ4∣x∣2+β∑i<jlog⁡∣xi−xj∣) dx(1)d\nu(x) = d\nu_{N,\beta}(x) := \frac{1}{Z_{N,\beta}} \exp\left(-\frac{N\beta}{4}|x|^2 + \beta\sum_{i<j} \log|x_i-x_j|\right)\,dx \tag*{(1)}

In particular, we consider functions f∈H1/2+ϵ(R)f \in H^{1/2+\epsilon}(\mathbb{R}) of low regularity, where we recall that the homogeneous and inhomogeneous (fractional) Sobolev spaces H˙s=H˙s(R)\dot{H}^{s} = \dot{H}^{s}(\mathbb{R}) and Hs=Hs(R)H^{s} = H^{s}(\mathbb{R}), respectively, are the spaces of all functions such that the norms

∥f∥H˙s2:=∫∣f^(ξ)∣2∣ξ∣2s dξ,∥f∥Hs2:=∫∣f^(ξ)∣2(1+∣ξ∣2s) dξ(2)\|f\|_{\dot{H}^{s}}^{2} := \int|\hat{f}(\xi)|^{2}|\xi|^{2s}\,d\xi,\qquad\|f\|_{H^{s}}^{2} := \int|\hat{f}(\xi)|^{2}(1+|\xi|^{2s})\,d\xi \tag*{(2)}

are finite, where f^\hat{f} denotes the Fourier transform (see (40) for our scaling convention). Recall that for the particular case of s=1/2s=1/2, we have the identity,

∥f∥H˙1/22=12π∬R2(f(x)−f(y))2(x−y)2 dxdy.(3)\|f\|_{\dot{H}^{1/2}}^{2}=\frac{1}{2\pi}\iint_{\mathbb{R}^{2}}\frac{(f(x)-f(y))^{2}}{(x-y)^{2}}\,dxdy. \tag*{(3)}

The significance of this form in the context of β\beta-ensembles is that one expects a (centered) linear statistic FNF_N to converge in distribution to a Gaussian random variable with variance [1, 16]

v(f):=12βπ2∫−22∫−22(f(x)−f(y))2(x−y)24−xy4−x24−y2 dxdy.(4)v(f):=\frac{1}{2\beta\pi^{2}}\int_{-2}^{2}\int_{-2}^{2}\frac{(f(x)-f(y))^{2}}{(x-y)^{2}}\frac{4-xy}{\sqrt{4-x^{2}}\sqrt{4-y^{2}}}\,dxdy. \tag*{(4)}

See [19, 20] for an analogous expression in the Wigner matrix setting, where there is an additional term depending on the fourth cumulant of the matrix entries.

The study of fluctuations and central limit theorems (CLT) for linear statistics of β\beta-ensembles was initiated by Johansson [16], whose approach centered around a variational equation that is now commonly referred to as the loop equation. Later works in this direction include [18] which obtained a CLT in the regularity class f∈C3f\in C^{3} for the Circular β\beta-ensembles (Cβ\betaE), and [8] which expanded the loop equation hierarchy to obtain CLTs for the Green’s function and log-characteristic polynomial,

s(z)=sN(z):=1N∑k=1N1xk−z,L(E)=LN(E):=∑k=1Nlog⁡(E−xk),(5)s(z)=s_N(z):=\frac{1}{N}\sum_{k=1}^{N}\frac{1}{x_k-z},\qquad L(E)=L_N(E):=\sum_{k=1}^{N}\log(E-x_k), \tag*{(5)}

for general β\beta-ensembles on the line (with standard one-cut and smoothness assumptions on the potential). On the other hand, for f∈C5f\in C^{5} and V∈C6V\in C^{6} a CLT was obtained in [2] for the mesoscopic linear statistics

∑k=1Nf(Nα(xk−E)),(6)\sum_{k=1}^{N}f\bigl(N^{\alpha}(x_k-E)\bigr), \tag*{(6)}

for fixed 0<α<10<\alpha<1 and EE in the bulk. An alternate line in [1] obtains the CLT for f∈C4f\in C^{4} and more general assumptions (such as multi-cut) on VV, using a transport approach. Johansson [16, Remark 2.5] conjectured that the only condition one needs for the CLT to hold should be that the limiting expression of the variance (4) is finite. Thus a natural line of work has been to push this conjecture for functions in the fractional Sobolev spaces HsH^{s}, with the aim of getting as close to s=1/2s=1/2 as possible. This was carried out in [22] and then later in [19] to obtain CLTs for smooth Wigner matrices, with a critical input being the use of Littlewood-Paley theory to decompose the test function ff (in H1/2+ϵH^{1/2+\epsilon} or C1/2+ϵC^{1/2+\epsilon}) into scales. For a more exhaustive list of works on the fluctuations of linear statistics for other random matrix models and classes of test functions, refer to [19] and the references therein.

The following is our main result. As noted in [19], the result is optimal up to the ϵ>0\epsilon>0 factor, as the CLT does not hold for functions in H1/2−ϵH^{1/2-\epsilon} which can be seen from the example of the indicator function 1x≤E\mathbf{1}_{x\leq E}.

Theorem 1.1. Let ν=νN,β\nu=\nu_{N,\beta} be given as in (1), for any β>0\beta>0. For any ϵ>0\epsilon>0 and k>0k>0, there is a C=C(β,ϵ,k)>0C=C(\beta,\epsilon,k)>0 such that the following holds. For any f:R→Rf:\mathbb{R}\to\mathbb{R} supported in (−2+k,2−k)(-2+k,2-k) we have,

Var⁡ν(∑k=1Nf(xk))≤C∥f∥H1/2+ϵ2.(7)\operatorname{Var}_{\nu}\left(\sum_{k=1}^{N}f(x_k)\right)\leq C\|f\|_{H^{1/2+\epsilon}}^{2}. \tag*{(7)}

Moreover, the centered linear statistic F−E[F]F-\mathbb{E}[F] converges to a Gaussian random variable with variance as given in (4).

It is worth mentioning at this point that our work does not rely on the loop equation method at all1, but depends rather on the more recent dynamical breakthroughs in the understanding of Dyson Brownian motion (DBM),

dxk(t)=2Nβ dBk(t)+(1N∑j≠k1xk(t)−xj(t)−12xk(t))dt,k=1,2,…,N,(8)dx_k(t)=\sqrt{\frac{2}{N\beta}}\,dB_k(t)+\left(\frac{1}{N}\sum_{j\ne k}\frac{1}{x_k(t)-x_j(t)}-\frac{1}{2}x_k(t)\right)dt,\qquad k=1,2,\ldots,N, \tag*{(8)}

where (Bk)k=1N(B_k)_{k=1}^{N} are independent standard Brownian motions. In particular, starting from [4] we now have a fairly robust understanding of the homogenization of the nonlocal dynamics,

ddtuk(t)=([∇2H(xt)]u)(k)=1N∑j≠kuj(t)−uk(t)(xk(t)−xj(t))2,(9)\frac{d}{dt}u_k(t)=\left([\nabla^2\mathcal{H}(x_t)]\mathbf{u}\right)^{(k)}=\frac{1}{N}\sum_{j\ne k}\frac{u_j(t)-u_k(t)}{(x_k(t)-x_j(t))^2}, \tag*{(9)}

which can be seen as a noisy Δ1/2\Delta^{1/2} operator driven by a log-gas random environment. An analysis by characteristics and maximum principle yield the (submicroscopic) estimate (31), which enabled the authors in [19] to push previous techniques (notably, Littlewood-Paley theory) in [22], extending their CLT to more general smooth Wigner matrices and f∈H1/2+ϵf\in H^{1/2+\epsilon}. A large part of our work builds off of the multiscale Littlewood-Paley analysis2 in [19], with our main contribution being a robust way to handle the mesoscopic scales (N−1,1](N^{-1},1], which now works for general β>0\beta>0. In particular, we recall a certain Helffer-Sjöstrand representation3, which has its (mathematical physics) roots in [15, 21], and later saw an application to (random matrix and β\beta-ensembles) universality in [14]. However, until (relatively recent) sharp estimates for the heat kernel of the dynamics (9) were developed [5], one could not leverage the Helffer-Sjöstrand representation fully towards the CLT program. This is the (now feasible) step that we carry out in this paper.

Organization and notation of the paper

In this paper, we denote c,C>0c,C>0 to be generic small and large constants, respectively, which do not depend on NN but may vary from line to line. Also let

φ=eC0(log⁡log⁡N)2(10)\varphi=e^{C_0(\log\log N)^2} \tag*{(10)}

be a subpolynomial error parameter for some fixed C0>0C_0>0. We denote s(z)=sN(z)=1N∑k=1N1xk−zs(z)=s_N(z)=\frac{1}{N}\sum_{k=1}^{N}\frac{1}{x_k-z} the Green’s function associated to the particles of the GβEG\beta E (1), and

ρsc(x)=12π4−x21∣x∣<2,msc(z)=−z+z2−42,(11)\rho_{\mathrm{sc}}(x)=\frac{1}{2\pi}\sqrt{4-x^2}\mathbf{1}_{|x|<2},\qquad m_{\mathrm{sc}}(z)=\frac{-z+\sqrt{z^2-4}}{2}, \tag*{(11)}

the semicircle law and its Stieltjes transform, respectively. Define the kk-th typical location γk\gamma_k through

∫−∞γkdρsc=kN.\int_{-\infty}^{\gamma_k}d\rho_{\mathrm{sc}}=\frac{k}{N}.

References

  1. [1]Florent Bekerman, Thomas Leblé, and Sylvia Serfaty, CLT for fluctuations of β-ensembles with general potential, Electron. J. Probab. 23 (2018), Paper no. 115, 31.DOI
  2. [2]Florent Bekerman and Asad M. Kiessling, Mesoscopic central limit theorem for general β-ensembles, Ann. Inst. Henri Poincaré Probab. Stat. 54 (2018), no. 4, 1917–1938 (English, with English and French summaries).arxiv.org/abs/1605.05206
  3. [4]P. Bourgade, Extreme gaps between eigenvalues of Wigner matrices, J. Eur. Math. Soc. (JEMS) 24 (2022), no. 8, 2823–2873.arxiv.org/abs/1812.10376
  4. [5]P. Bourgade and G. Cipolloni, Universality of spectral form factors (2026).
  5. [8]Paul Bourgade, Krishnan Mody, and Michel Pain, Optimal local law and central limit theorem for β-ensembles, Comm. Math. Phys. 390 (2022), no. 3, 1017–1079.DOI
  6. [14]L. Erdős and H.-T. Yau, Gap universality of generalized Wigner and β-ensembles, J. Eur. Math. Soc. (JEMS) 17 (2015), no. 8, 1927–2036.
  7. [15]Bernard Helffer and Johannes Sjöstrand, On the correlation for Kac-like models in the convex case, J. Statist. Phys. 74 (1994), no. 1-2, 349–409.DOI
  8. [16]Kurt Johansson, On fluctuations of eigenvalues of random Hermitian matrices, Duke Math. J. 91 (1998), no. 1, 151–204.DOI
  9. [18]Gaultier Lambert, Mesoscopic central limit theorem for the circular β-ensembles and applications, Electron. J. Probab. 26 (2021), Paper No. 7, 33.arxiv.org/abs/1902.06611
  10. [19]B. Landon and P. Sosoe, Almost-optimal bulk regularity conditions in the CLT for Wigner matrices, ArXiv:2204.03419 (2022).arxiv.org/abs/2204.03419
  11. [20]A. Lytova and L. Pastur, Central limit theorem for linear eigenvalue statistics of random matrices with independent entries, Ann. Probab. 37 (2009), no. 5, 1778–1840.
  12. [21]Ali Naddaf and Thomas Spencer, On homogenization and scaling limit of some gradient perturbations of a massless free field, Comm. Math. Phys. 183 (1997), no. 1, 55–84.
  13. [22]Philippe Sosoe and Percy Wong, Regularity conditions in the CLT for linear eigenvalue statistics of Wigner matrices, Adv. Math. 249 (2013), 37–87.DOI

Paper details

vlnc:W_qdpbus9y
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 4

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

ChatGPT 5.6

View statement
Looking for related papers…
Contents