--- title: Bergman Determinantal Point Process url: https://www.emergentmind.com/topics/bergman-determinantal-point-process type: topic --- # Bergman Determinantal Point Process A Bergman determinantal point process is a determinantal point process whose correlation kernel is a Bergman kernel, or more generally a Bergman projection kernel, attached to a Hilbert space of holomorphic functions or holomorphic sections. The term covers several distinct but structurally parallel models: infinite-rank processes on the unit disk and on weighted Bergman spaces over bounded pseudoconvex domains, finite-rank “Bergman ensembles” on compact complex manifolds obtained from \(H^0(X,L^k)\), and hyperbolic or complex-hyperbolic variants governed by invariant Bergman kernels [2405.09203] [2112.15557] [2404.14793]. In all cases the core mechanism is the same: an orthogonal projection in an \(L^2\)-space produces a reproducing kernel, and that kernel defines the determinantal correlations. ## 1. Projection-kernel definition Let \((E,\lambda)\) be a measure space, and let \(K\) be a kernel such that the \(n\)-point correlation functions satisfy \[ \rho_n(x_1,\dots,x_n)=\det\big(K(x_i,x_j)\big)_{1\le i,j\le n}. \] A point process with this property is determinantal. When the associated integral operator is Hermitian, locally trace class, and satisfies \(0\le K\le I\), the Macchi–Soshnikov / Shirai–Takahashi theorem yields existence and uniqueness of the corresponding DPP; in particular, any locally trace class orthogonal projection defines a DPP [2503.19197]. Bergman DPPs are projection DPPs. If \(H\subset L^2(E,\lambda)\) is a reproducing kernel Hilbert space of holomorphic functions, and \(P_H\) is the orthogonal projection onto \(H\), then the integral kernel of \(P_H\) is the Bergman kernel. The reproducing property has the form \[ f(x)=\int_E K(x,y)f(y)\,d\lambda(y), \] and the kernel is Hermitian and idempotent at the operator level. This is the basic construction on the unit disk, on weighted Bergman spaces over domains in \(\mathbb C^n\), and on spaces of holomorphic sections of positive line bundles [2112.15557] [2404.14793]. In the compact-manifold setting, the process is finite. If \(\{s_i\}_{i=1}^{N_k}\) is an orthonormal basis of \(H^0(X,L^k)\), the Bergman kernel is \[ B_{(k\phi,\mu)}(x,y)=\sum_{i=1}^{N_k}s_i(x)\otimes \overline{s_i(y)}, \] and the associated Bergman ensemble has joint density \[ d\mathcal P_{k\phi}(x_1,\dots,x_{N_k}) =\frac{1}{Z_{N_k}(k\phi)} \left|\det(s_i(x_j))_{1\le i,j\le N_k}\right|_{k\phi}^2 \,d\mu^{\otimes N_k}(x_1,\ldots,x_{N_k}), \] with \(Z_{N_k}(k\phi)=N_k!\). Because the kernel is the orthogonal projector onto a finite-dimensional subspace, the process has exactly \(N_k\) points [2405.09203]. ## 2. Principal geometric realizations The phrase “Bergman determinantal point process” does not denote a single model. It denotes a family of determinantal processes attached to Bergman-type spaces in several geometries [1703.08978] [2211.06955]. | Setting | Hilbert space / kernel | Process type | |---|---|---| | Unit disk \(\mathbb D\) | \(A^2(\mathbb D)\), \(K(z,w)=(1-z\overline w)^{-2}\) with \(du=\pi^{-1}dx\,dy\), or \(K(z,w)=\pi^{-1}(1-z\overline w)^{-2}\) with Lebesgue measure | Infinite-rank | | Weighted disk | \(A^2_\alpha(\mathbb D)\), \(B_w(z,\zeta)=\frac{\alpha+1}{\pi}(1-z\overline\zeta)^{-(\alpha+2)}\) | Infinite-rank | | Bounded pseudoconvex \(\Omega\subset\mathbb C^n\) | \(H(k\phi)=L^2(e^{-k\phi}d\lambda)\cap\mathcal O(\Omega)\), kernel \(K_{k\phi}\) | Infinite-rank | | Compact complex manifold \(X\) | \(H^0(X,L^k)\), kernel \(B_{(k\phi,\mu)}\) | Finite-rank, exactly \(N_k\) points | | Hyperbolic disk / complex hyperbolic space | Weighted Bergman kernels and Bergman projection kernels | Infinite-rank | On the unit disk, the standard Bergman space \(A^2(\mathbb D)\) with normalized Lebesgue measure \(du=\pi^{-1}dx\,dy\) has kernel \[ K(z,w)=\frac{1}{(1-z\overline w)^2}, \] whereas with Lebesgue measure as background measure the kernel is \[ K(z,w)=\frac{1}{\pi(1-z\overline w)^2}. \] This normalization dependence is standard. The corresponding process is also the zero set of the Gaussian analytic function \[ f(z)=\sum_{n=0}^\infty a_n z^n \] with i.i.d. standard complex Gaussian coefficients [2404.14978] [2112.15557]. Weighted disk models include the classical weighted Bergman spaces \(A^2_\alpha(\mathbb D)\), \(\alpha>-1\), with kernel \[ B_w(z,\zeta)=\frac{\alpha+1}{\pi}\frac{1}{(1-z\overline{\zeta})^{\alpha+2}}. \] More general generalized Bergman spaces on \(\mathbb D\) also define DPPs under the integrability condition \[ \int_{\mathbb D}(1-|z|)^2\,B_w(z,z)\,w(z)\,d\lambda(z)<\infty \] [1411.4951]. On bounded pseudoconvex domains \(\Omega\subset\mathbb C^n\), one considers \[ H(k\phi):=L^2(e^{-k\phi}d\lambda)\cap \mathcal O(\Omega), \] where \(\phi\in SPSH(\Omega)\). Its reproducing kernel \(K_{k\phi}\) defines a DPP \(\Lambda_k\) with respect to \(e^{-k\phi}d\lambda\). This is an infinite-rank Bergman DPP with a semiclassical parameter \(k\) [2404.14793]. On compact complex manifolds, Bergman DPPs arise from a positive holomorphic line bundle \(L\to X\) and its tensor powers \(L^k\). The Bergman kernel of \(H^0(X,L^k)\) yields a finite determinantal ensemble with exactly \(N_k=\dim H^0(X,L^k)\) points [2211.06955]. Complex-hyperbolic and hyperbolic models include the Bergman DPP on \(\mathbb D_d\) induced by the kernel \[ K_{\mathbb D_d}(z,w)=\frac{d!}{\pi^d}\frac{1}{(1-\langle z,w\rangle)^{d+1}}, \] as well as weighted Bergman kernels on the Poincaré disk such as \[ G_0^\nu(z,w)=\frac{2\nu-1}{\pi}(1-z\overline w)^{-2\nu}, \] with reference measure \((1-|z|^2)^{2\nu-2}dz\) [2101.09622] [1712.04824]. ## 3. Asymptotic regimes and universal limits For Bergman ensembles on compact complex manifolds, the principal asymptotic regime is \(k\to\infty\). In local normal coordinates at a point \(x\), the Bergman kernel admits a near-diagonal expansion whose leading term is a Gaussian kernel on \(\mathbb C^n\). More precisely, after rescaling by \(u/\sqrt{k}\) and \(v/\sqrt{k}\), \[ k^{-n}B_{(k\phi+\psi,\mu)}\Big(\frac{u}{\sqrt{k}},\frac{v}{\sqrt{k}}\Big) = B_\infty^\phi(u,v)+O(k^{-1/2}), \] where \[ B_\infty^\phi(u,v) =\frac{\det\lambda}{\pi^n} \exp\bigg(\sum_{i=1}^n \lambda_i\Big(u_i\overline{v_i}-\tfrac12|u_i|^2-\tfrac12|v_i|^2\Big)\bigg), \] and \(\lambda_1,\dots,\lambda_n>0\) are the eigenvalues of the curvature matrix at the base point. The rescaled \(m\)-point correlation functions converge to \[ \det\big(B_\infty^\phi(u_i,u_j)\big)_{1\le i,j\le m}, \] so the local limit is a multidimensional generalization of the infinite Ginibre ensemble [2211.06955]. The same compact-manifold framework yields global laws. If \[ \hat\mu_k=\frac1{N_k}\sum_{i=1}^{N_k}\delta_{X_i}, \] then \(\hat\mu_k\) converges in probability, in the weak topology, to the equilibrium measure \[ \mu_{eq}^\phi=\frac{1}{\mathrm{vol}(L)}(\omega+dd^c\phi)^n. \] For weighted variants of the process, the empirical measures satisfy a large deviation principle with speed \(kN_k\sim N_k^{1+1/n}\), and the rate function is the Legendre–Fenchel transform of the Mabuchi functional increment \[ \mathcal L_{eq}(\phi,\psi)=\mathcal E_{eq}(\phi+\psi)-\mathcal E_{eq}(\phi) \] [2211.06955]. A distinct asymptotic regime appears on bounded pseudoconvex domains. There the emphasis is not fixed-cardinality ensembles but the large-\(k\) behavior of the infinite-rank weighted Bergman kernels \(K_{k\phi}\). For \(\phi\)-admissible \(u\in C_c^\infty(\Omega)\), the scaled cumulant generating function satisfies \[ \lim_{k\to\infty}\frac{1}{k^{n+1}} \log \mathbb E\left[e^{-k\langle u,\Lambda_k\rangle}\right] = -\mathcal E(\phi+u), \] where \[ \mathcal E(\phi+u)= \frac{1}{(n+1)!}\sum_{j=0}^{n}\int_\Omega u\,(dd^c(\phi+u))^j\wedge(dd^c\phi)^{n-j}. \] This connects weighted Bergman DPPs on bounded domains to pluripotential energy functionals [2404.14793]. At the flat-model level, the Heisenberg family on \(\mathbb C^d\) has kernel \[ K^{(d)}_{\mathrm{Heisenberg}}(x,x')=e^{x\cdot\overline{x'}}, \] with respect to Gaussian background measure. In the operator-theoretic framework of partial isometries and reproducing kernels, this is the Bargmann–Fock model that underlies Ginibre-type limits [1903.04945]. ## 4. Hyperbolic Bergman processes and fluctuation theory On the Poincaré disk, the lowest hyperbolic Landau level \(m=0\) gives the weighted Bergman kernel \[ G_0^\nu(z,w)=\frac{2\nu-1}{\pi}(1-z\overline w)^{-2\nu}, \] and the associated hyperbolic-type point process \(\mathbb P_0^\nu\) is a Bergman DPP. More generally, higher hyperbolic Landau levels yield kernels \(G_m^\nu\) involving Jacobi polynomials and define generalized Bergman-type DPPs [1712.04824]. For the number \(N_r\) of particles inside the Euclidean disk \(D_r=\{|z|0\) such that \[ \frac{\mathrm{var}(\Xi(B_R))}{\mathbb E[\Xi(B_R)]}\ge C \quad\text{for all }R>0. \] Since projection DPPs also satisfy \(\mathrm{var}(\Xi(A))\le \mathbb E[\Xi(A)]\), the variance is of the same order as the expectation, and such processes are never hyperuniform [2503.19197]. A potential source of confusion is that another hyperbolic line of work interprets the asymptotic \(\operatorname{Var}\sim c/(1-R^2)\) for affine ensembles as growth like boundary length [2210.08214], whereas the general homogeneous theorem concludes non-hyperuniformity [2503.19197]. This suggests that fluctuation terminology in hyperbolic geometry is sensitive to how window size is parameterized and compared with ambient volume growth. ## 5. Palm measures, conditional laws, and quasi-invariance For generalized Bergman DPPs on the unit disk, a central structural result is the equivalence of reduced Palm measures. Under the condition \[ \int_{\mathbb D}(1-|z|)^2\,B_w(z,z)\,w(z)\,d\lambda(z)<\infty, \] all reduced Palm measures of all orders are mutually equivalent, and their Radon–Nikodym derivatives are expressed by regularized multiplicative functionals built from finite Blaschke products. If \[ b_p(z)=\prod_{j=1}^\ell \frac{1-\overline{p_j}z}{z-p_j}, \] then the derivative of the Palm measure at \(p\) has the form \[ \frac{d\mathbb P_{B_w}^p}{d\mathbb P_{B_w}}(X) = \frac{e^{2S_p(X)}}{\mathbb E_{\mathbb P_{B_w}}[e^{2S_p}]}, \] with \(S_p\) a regularized additive functional. In this Bergman setting the process is explicitly described as non-rigid, in contrast with the generalized Fock case on \(\mathbb C\) [1411.4951]. The higher-dimensional analogue holds on domains \(U\subset\mathbb C^d\) for weighted Bergman spaces \(A^2(U,\omega)\). If \(U\) contains a non-constant bounded holomorphic function, then the determinantal measure \(P_{K_\omega}\) is equivalent to its arbitrary reduced Palm measure of any order. A corollary is quasi-invariance under compactly supported diffeomorphisms of \(U\) [1703.08978]. On the unit disk with the classical Bergman kernel \[ K(z,w)=\frac{1}{\pi(1-z\bar w)^2}, \] the conditional law inside a relatively compact region \(B\subset\mathbb D\), given the exterior configuration \(Y\), admits an explicit \(L\)-ensemble form. The conditional kernel is \[ L_Y(q_1,q_2)=\frac{\Phi_{q_1}(Y)\Phi_{q_2}(Y)}{1-q_1\overline{q_2}}, \] where \(\Phi_q\) is a regularized Blaschke-type multiplicative functional, and the reduced Palm measure satisfies \[ P_q=\Phi_q\,P_K. \] This gives a concrete Gibbs-like description of the Bergman DPP conditioned on the outside configuration [2112.15557]. ## 6. Reconstruction, quadrature, and simulation Bergman DPPs have been used as numerical-integration nodes on compact complex manifolds. For a holomorphic line bundle \(L\to X\) and Bergman ensemble with kernel \(B_{(k\phi+\psi,\mu)}\), the estimator \[ I_k(f)=\sum_{i=1}^{N_k}\frac{f(X_i)}{B_{(k\phi+\psi,\mu)}(X_i,X_i)} \] is unbiased: \[ \mathbb E[I_k(f)]=\int_X f(x)\,d\mu(x). \] For Lipschitz \(f\) supported in the weak bulk, \[ \sqrt{N_k^{1+\frac1d}\left( \sum_{i=1}^{N_k}\frac{f(X_i)}{B_{(k\phi+\psi,\mu)}(X_i,X_i)} -\int_X f(x)\,d\mu(x)\right)} \xrightarrow[k\to\infty]{\mathcal D} \mathcal N(0,\sigma_{f,\phi}^2), \] and the mean squared error decays as \[ O\big(N^{-1-2/d_{\mathbb R}}\big), \qquad d_{\mathbb R}=2d. \] This is faster than the \(O(N^{-1})\) rate of i.i.d. Monte Carlo [2405.09203]. On the unit disk, vector-valued linear statistics in the Bergman space provide another analytical use. For \[ \Theta_N^{(s,z)}(X) = \sum_{x\in X\cap U_N(z)} e^{-s d_h(z,x)}K(\cdot,x), \qquad 1