Abstract
We develop a framework for establishing the Law of Large Numbers for the eigenvalues in the random matrix ensembles as the size of the matrix goes to infinity simultaneously with the beta (inverse temperature) parameter going to zero. Our approach is based on the analysis of the (symmetric) Dunkl transform in this regime. As an application we obtain the LLN for the sums of random matrices as the inverse temperature goes to 0. This results in a one-parameter family of binary operations which interpolates between classical and free convolutions of the probability measures. We also introduce and study a family of deformed cumulants, which linearize this operation.
Similar content being viewed by others
Notes
Say, we deal with complex Hermitian matrices. Then this set is an orbit of the unitary group U(N) under the action by conjugations, and the uniform measure on the orbit is the image of the Haar (uniform) measure on U(N) with respect to this action.
If we know that Q is invariant under orthogonal conjugations and we know the values of \(\chi _Q(X)\) for all normal X, then we can uniquely determine the law of Q. In fact it is sufficient to take X to be symmetric (or \(\mathbf {i}\) times symmetric).
For a reader who is not familiar with the theory of multivariate Bessel functions, we remark that at \(N=1\), \(B_{(a)}(z;\, \theta )=\exp (az)\). Hence, choosing \(z=\mathbf {i}t\), the Bessel functions turn into the exponents \(\exp (\mathbf {i}a t)\) and uniqueness turns into the well-known uniqueness of a measure with a given Fourier transform.
The reasons are: widespread independence of the Law of Large Number from the value \(\beta \) for the random matrix \(\beta \)-ensembles, cf. [BAG97, J98, BoG15]; the same answer in Theorem 1.1 for three values \(\theta =\tfrac{\beta }{2}=\tfrac{1}{2},1,2\); existence of \(\theta \)-independent observables for \(\mathbf {a}+_\theta \mathbf {b}\), see [GM20, Theorem 1.1]; \(\theta \)-independence in a discrete version of the same problem, see [Hu18].
One should compare [AP18, (3.1), (4.2)] with our pair of equations (3.8) and notice that the conventions are slightly different: \((d)_n\) is a falling factorial in [AP18] and \((\gamma )_n\) is a rising factorial in our work. One can also directly compare the formulas for the first four cumulants of (3.3) and (3.6) with similar formulas above Corollary 4.3 in the journal version of [AP18]. We are grateful to Octavio Arizmendi and Daniel Perales for pointing this connection to us.
It is plausible that many of the results of our text extend to the situations where this restrictive condition fails.
The space of test-functions f is equipped with a topology: \(f^{n}\) converge to 0 as \(n\rightarrow \infty \), if the supports of all these functions belong to the same compact set and all partial derivatives of \(f^{n}\) converge to 0 uniformly.
In our wordings we stick to the situation when \(\mu _N\) are bona fide probability measures. If they are distributions (i.e. generalized functions possibly without any positivity), then all the random variables produced from them should be interpreted in formal sense: the laws of such random variables can be identified with expectations of various smooth functions of them, which are readily computed as pairings of \(\mu _N\) with appropriate test functions. (One also should divide by N! to adjust for differences between ordered and arbitrary N-tuples).
We omit the dependence on \(\gamma \) from the notation \(W(\pi )\).
In this section the word “distribution” is used in probabilistic meaning, as in “distribution of a random variable”, rather than in functional-analytic meaning, where a distribution is a synonym of a generalized function.
We omit the dependence on \(\gamma \) and on the sequences \(a_1, a_2, \ldots \) and \(\kappa _1, \kappa _2, \ldots \) from the notation \(w(\pi )\).
References
Ahn, A.: Airy Point Process via Supersymmetric Lifts. arXiv:2009.06839
Akemann, G., Byun, S.-S.: The high temperature crossover for general 2D Coulomb gases. J. Stat. Phys. 175(6), 1043–1065 (2019)
Allez, R., Bouchaud, J.-P., Guionnet, A.: Invariant Beta ensembles and the Gauss–Wigner crossover. Phys. Rev. Lett. 109(9), 094–102 (2012)
Allez, R., Bouchaud, J.-P., Majumdar, S.N., Vivo, P.: Invariant \(\beta \)-Wishart ensembles, crossover densities and asymptotic corrections to the Marcenko–Pastur law. J. Phys. A Math. Theor. 46(1), 015001 (2013)
Allez, R., Dumaz, L.: From Sine kernel to Poisson statistics. Electron. J. Probab. 19(114), 1–25 (2014). arXiv:1407.5402
Anderson, G.W.: A Short Proof of Selberg’s Generalized Beta Formula. In Forum Mathematicum 3(3), 415418 (1991)
Andraus, S., Hermann, K., Voit, M.: Limit theorems and soft edge of freezing random matrix models via dual orthogonal polynomials, to appear in Journal of Mathematical Physics. arXiv:2009.01418
Anker, J.-P.: An introduction to Dunkl theory and its analytic aspects. Analytic, Algebraic and Geometric Aspects of Differential Equations. Birkhäuser, Cham, pp. 3–58 (2017)
Arizmendi, O., Perales, D.: Cumulants for finite free convolution. J. Comb. Theory Ser. A 155, 244–266 (2018)
Askey, R., Wimp, J.: Associated Laguerre and Hermite polynomials. Proc. R. Soc. Edinb. Sect. A Math. 96(1–2), 15–37 (1984)
Assiotis, T., Najnudel, J.: The boundary of the orbital beta process. arXiv:1905.08684
Ben Arous, G., Guionnet, A.: Large deviations for Wigner’s law and Voiculescu’s non-commutative entropy. Probab. Theory Relat. Fields 108, 517–542 (1997)
Benaych-Georges, F., Péché, S.: Poisson statistics for matrix ensembles at large temperature. J. Stat. Phys. 161(3), 633–656 (2015)
Bercovici, H., Voiculescu, D.: Superconvergence to the central limit and failure of the Cramér theorem for free random variables. Probab. Theory Relat. Fields 102, 215–222 (1995)
Borodin, A., Gorin, V.: General \(\beta \)-Jacobi Corners Process and the Gaussian Free Field. Commun. Pure Appl. Math. 68(10), 1774–1844 (2015)
Bożejko, M., Dołȩga, M., Ejsmont, W., Gal, Ś.R.: Reflection length with two parameters in the asymptotic representation theory of type B/C and applications. arXiv:2104.14530
Bufetov, A., Gorin, V.: Representations of classical Lie groups and quantized free convolution. Geom. Funct. Anal. 25, 763–814 (2015)
Bufetov, A., Gorin, V.: Fluctuations of particle systems determined by Schur generating functions. Adv. Math. 338, 702–781 (2018)
Bufetov, A., Gorin, V.: Fourier transform on high-dimensional unitary groups with applications to random tilings. Duke Math. J. 168(13), 2559–2649 (2019)
Cuenca, C.: Universal behavior of the corners of orbital beta processes. Int. Math. Res. Not. 19, 14761–14813 (2021)
Collins, B.: Moments and cumulants of Polynomial random variables on unitary groups, the Itzykson Zuber integral and free probability. Int. Math. Res. Not. 2003(17), 953–982 (2003)
Collins, B., Sniady, P.: Integration with respect to the Haar measure on unitary, orthogonal and symplectic group. Commun. Math. Phys. 264, 773–795 (2006)
Drake, D.: The combinatorics of associated Hermite polynomials. Eur. J. Comb. 30(4), 1005–1021 (2009)
Dumitriu, I., Edelman, A.: Global spectrum fluctuations for the \(\beta \)-Hermite and \(\beta \)-Laguerre ensembles via matrix models. J. Math. Phys. 47, 063302 (2006)
Dunkl, C.: Differential-difference operators associated to reflection groups. Trans. Am. Math. Soc. 311(1), 167–183 (1989)
Duy, K.T., Shirai, T.: The mean spectral measures of random Jacobi matrices related to Gaussian beta ensembles. Electron. Commun. Probab. 20 (2015)
Etingof, P., Ma, X.: Lecture notes on Cherednik algebras. Preprint. arXiv:1001.0432 (2010)
Faraut, J., Fourati, F.: Markov–Krein transform. Colloq. Math. 144(1), 137–156 (2016)
Feller, W.: An Introduction to Probability: Theory and its Applications, vol. II, 2nd edn. Wiley (1971)
Forrester, P.J.: High-low temperature dualities for the classical \(\beta \)-ensembles. arXiv:2103.11250
Forrester, P.J., Mazzuca, G.: The classical \(\beta \)-ensembles with \(\beta \) proportional to \(1/N\): from loop equations to Dyson’s disordered chain. arXiv:2102.09201
Fyodorov, Y., Le Doussal, P.: Moments of the position of the maximum for GUE characteristic polynomials and for log-correlated Gaussian processes. J. Stat. Phys. 164, 190–240 (2016)
Gorin, V., Kleptsyn, V.: Universal objects of the infinite beta random matrix theory. arXiv:2009.02006
Gorin, V., Marcus, A.W.: Crystallization of random matrix orbits. Int. Math. Res. Not. 2020(3), 883–913 (2020)
Gorin, V., Panova, G.: Asymptotics of symmetric polynomials with applications to statistical mechanics and representation theory. Ann. Probab. 43(6), 3052–3132 (2015). arXiv:1301.0634
Gorin, V., Shkolnikov, M.: Multilevel Dyson Brownian motions via Jack polynomials. Probab. Theory Relat. Fields 163(3), 413–463 (2015). arXiv:1401.5595
Gorin, V., Sun, Y.: Gaussian fluctuations for products of random matrices, to appear in American Journal of Mathematics. arXiv:1812.06532 (2018)
Guhr, T., Kohler, H.: Recursive construction for a class of radial functions. I. Ordinary space. J. Math. Phys. 43(5), 2707–2740 (2002)
Hardy, A., Lambert, G.: CLT for Circular beta-Ensembles at high temperature. J. Funct. Anal. 280(7), 108869 (2021)
Johansson, K.: On fluctuations of eigenvalues of random Hermitian matrices. Duke Math. J. 91(1), 151–204 (1998)
Huang, J.: Law of large numbers and central limit theorems by jack generating functions. Adv. Math. 380, 107545 (2021)
Kerov, S.: Interlacing measures. American Mathematical Society Translations, 35–84 (1998)
Kerov, S.: Asymptotic Representation Theory of the Symmetric Group and its Applications in Analysis. American Mathematical Society, Providence (2003)
Killip, R., Stoiciu, M.: Eigenvalue statistics for CMV matrices: from Poisson to clock via random matrix ensembles. Duke Math. J. 146, 361–399 (2009)
Kirillov, A., Jr.: Lectures on affine Hecke algebras and Macdonald’s conjectures. Bull. Am. Math. Soc. 34(3), 251–292 (1997)
Knutson, A., Tao, T.: Honeycombs and sums of Hermitian matrices. Not. Am. Math. Soc. 48(2), 175–186 (2001)
Lukacs, E.: Characteristic Functions, 2nd edn. Griffin (1970)
Macdonald, I.G.: Symmetric Functions and Hall Polynomials. Oxford University Press (1998)
Marcus, A.W.: Polynomial convolutions and (finite) free probability, Preprint (2018)
Matsumoto, S., Novak, J.: A moment method for invariant ensembles. Electron. Res. Announc., 25, 60
Matveev, K.: Macdonald-positive specializations of the algebra of symmetric functions: Proof of the Kerov conjecture. Ann. Math. 189(1), 277–316 (2019)
Mergny, P., Potters, M.: Rank one HCIZ at high temperature: interpolating between classical and free convolutions. Preprint. arXiv:2101.01810 (2021)
Mingo, J.A., Speicher, R.: Free Probability and Random Matrices, vol. 35. Springer, New York (2017)
Nakano, F., Trinh, K.D.: Gaussian beta ensembles at high temperature: eigenvalue fluctuations and bulk statistics. J. Stat. Phys. 173(2), 296–321 (2018)
Nakano, F., Trinh, H.D., Trinh, K.D.: Limit theorems for moment processes of beta Dyson’s Brownian motions and beta Laguerre processes. arXiv:2103.09980
Neretin, Y.A.: Rayleigh triangles and non-matrix interpolation of matrix beta integrals. Sbornik: Math. 194(4), 515–540 (2003)
Nica, A., Speicher, R.: Lectures on the Combinatorics of Free Probability, vol. 13. Cambridge University Press (2006)
Okounkov, A., Olshanski, G.: Shifted Jack polynomials, binomial formula, and applications. Math. Res. Lett. 4, 69–78 (1997)
Olshanski, G., Vershik, A.: Ergodic unitarily invariant measures on the space of infinite Hermitian matrices, In: Contemporary Mathematical Physics. F. A. Berezin’s memorial volume. Amer. Math. Transl. Ser. 2, vol. 175 R. L. Dobrushin et al., (eds), pp. 137–175 (1996). arXiv:math/9601215
Opdam, E.M.: Dunkl operators, Bessel functions and the discriminant of a finite Coxeter group. Compos. Math. 85(3), 333–373 (1993)
Pakzad, C.: Poisson statistics at the edge of Gaussian beta-ensembles at high temperature. arXiv:1804.08214 (2018)
Pickrell, D.: Mackey analysis of infinite classical motion groups. Pac. J. Math. 150, 139–166 (1991)
Rösler, M.: A positive radial product formula for the Dunkl kernel. Trans. Am. Math. Soc. 355(6), 2413–2438 (2003). arXiv:math/0210137
Rösler, M.: Dunkl Operators: Theory and Applications. In: Koelink, E., Van Assche, W. (eds) Orthogonal Polynomials and Special Functions. Lecture Notes in Mathematics, vol. 1817, (2003). Springer, Berlin
Rudin, W.: Functional Analysis. International Series in Pure and Applied Mathematics, vol. 8, 2nd edn. McGraw-Hill Science/Engineering/Math, New York, NY (1991)
Shlyakhtenko, D., Tao, T.: (with an appendix by David Jekel), Fractional free convolution powers, arXiv:2009.01882
Sokal, A.D., Zeng, J.: Some multivariate master polynomials for permutations, set partitions, and perfect matchings, and their continued fractions. arXiv:2003.08192
Spohn, H.: Generalized gibbs ensembles of the classical toda chain. J. Stat. Phys. 180, 4–22 (2020)
Stanley, R.P.: Some combinatorial properties of jack symmetric functions. Adv. Math. 77, 76–115 (1989)
Trimèche, K.: Payley–Wiener Theorems for the Dunkl Tramsform and Dunkl Translation Operators. Integral Transform. Spec. Funct. 13(1), 17–38 (2002)
Trinh, H.D., Trinh, H.D.: Beta Laguerre ensembles in global regime. Osaka J. Math. 58, 435–450 (2021)
Trinh, H.D., Trinh, K.D.: Beta Jacobi ensembles and associated Jacobi polynomials. arXiv:2005.01100
Voiculescu, D.: Limit laws for random matrices and free products. Invent. Math. 104, 201–220 (1991)
Weyl, H.: Das asymptotische Verteilungsgesetz der Eigenwerte lineare partieller Differentialgleichungen. Math. Ann. 71, 441–479 (1912)
Acknowledgements
The authors would like to thank Alexey Bufetov and Greta Panova for helpful discussions. We are thankful to Maciej Dołȩga for pointing us to the articles [D09, BDEG21], and for sending us a draft of a new version of the latter paper. We thank Octavio Arizmendi and Daniel Perales for directing us to their work [AP18]. We are grateful to two anonymous referees for their feedback. The work of V.G. was partially supported by NSF Grants DMS-1664619, DMS-1949820, by BSF grant 2018248, and by the Office of the Vice Chancellor for Research and Graduate Education at the University of Wisconsin–Madison with funding from the Wisconsin Alumni Research Foundation.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
9. Appendix: Law of Large Numbers for Fixed Temperature
9. Appendix: Law of Large Numbers for Fixed Temperature
The aim of this “Appendix” is to probe the possibility of a version of Theorem 3.8 in which \(\theta > 0\) is fixed and is not changing with N. The following claim is an analogue of one direction of Theorem 3.8.
Claim 9.1
(LLN for finite temperature). Let \(\{\mu _N\}_{N \ge 1}\) be a sequence of exponentially decaying probability measures on tuples \(a_1\le \cdots \le a_N\). For each N, let \(G_{N; \theta }(x_1, \ldots , x_N) := G_\theta (x_1, \ldots , x_N; \mu _N)\) be the BGF of \(\mu _N\). Assume that the sequence \(\{G_{N; \theta }\}_N\) satisfies the following conditions:
-
(a)
\(\displaystyle \lim _{N\rightarrow \infty }\left. \frac{1}{N}\cdot \frac{\partial ^l}{\partial x_i^l}\ln {(G_{N; \theta })}\right| _{x_1=\dots =x_N=0} = (l - 1)!\cdot c_l\), for all \(l\in \mathbb {Z}_{\ge 1}\).
-
(b)
\(\displaystyle \lim _{N\rightarrow \infty }\left. \frac{1}{N}\cdot \frac{\partial }{\partial x_{i_1}}\cdots \frac{\partial }{\partial x_{i_r}}\ln {(G_{N; \theta })}\right| _{x_1=\dots =x_N=0} = 0\), for all \(i_1, \dots , i_r\in \mathbb {Z}_{\ge 1}\) such that the set \(\{i_1, \dots , i_r\}\) contains at least two distinct indices.
Then the sequence \(\{\mu _N\}_{N \ge 1}\) satisfies the following LLN (compare to Definition 3.1): there exist real numbers \(\{m_{k}\}_{k\ge 1}\) such that for any \(s=1,2,\dots \) and any \(k_1, \dots , k_s\in \mathbb {Z}_{\ge 1}\), we have
If this occurs, \(\{c_l\}_{l\ge 1}\) and \(\{m_k\}_{k\ge 1}\) are related by either
or, equivalently,
We do not present a proof of the claim, but probably one can prove it with the same techniques that we have used in Sect. 5. At \(\theta =1\), another approach is by a degeneration of [BuG15, Theorem 5.1], see also [MN18]. This claim would prove the one-sided implication
Based on our Theorem 3.8, on [BuG19, Theorem 2.6] (which studies the CLT at fixed \(\theta = 1\)), and on the classical theorem that relates the weak convergence of measures to the convergence of their characteristic functions, the reader may be inclined to believe that the reverse implication is also true and that this kind of “if and only if” results are always expected.
However, this turns out to be wrong. The naive analogue of Theorem 3.8 does not hold for fixed \(\theta \): the “expected if and only if statement" is false. Here is a counter-example.
Let us consider a sequence of probability measures \(\mu _N\) such that a random \(\mu _N\)-distributed vector \(a_1 \ge \cdots \ge a_N\) has \(a_1 = \cdots = a_N\) almost surely and this common value a is distributed according to a Gaussian measure of mean 0 and variance N. In this case, the random variable
is distributed as the k-th power of a Gaussian random variable of mean 0 and variance 1/N. Consequently, the sequence \(\{\mu _N\}_N\) satisfies a LLN, and all \(m_k\)’s are equal to zero. On the other hand, by using \(B_{(a, \dots , a)}(x_1, \dots , x_N; \theta ) = \exp (a\sum _{i=1}^N{x_i})\), it follows that the BGF of \(\mu _N\) equals
The sequence \(\{G_{N; \theta }\}_N\) of BGFs then satisfies
therefore contradicting condition (b) from Claim 9.1.
A more refined question is whether some “if and only if for LLN” statement holds, if one modifies somehow the conditions (a) and (b) of Claim 9.1. Based on small calculations (obtained when trying to reverse-engineer the proof of Theorem 3.8), it is plausible that the answer is yes.
Indeed, based on Proposition 2.11, we must study the limits of the expressions
where \(\lambda \) ranges over the set of all partitions of a given size k. We have performed calculations for \(k=2, 3\); they indicate that the conditions on second-order derivatives in (a) and (b) from Claim 9.1 should be replaced by:
and the conditions on third-order derivatives should be replaced by:
These relations are much more involved than conditions (a) and (b) from Claim 9.1, or than the conditions from Definition 3.3. What should be the correct “if and only if” relations for \(k>3\)? This is an interesting open question for future research.
Rights and permissions
About this article
Cite this article
Benaych-Georges, F., Cuenca, C. & Gorin, V. Matrix Addition and the Dunkl Transform at High Temperature. Commun. Math. Phys. 394, 735–795 (2022). https://doi.org/10.1007/s00220-022-04411-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00220-022-04411-z