---
title: Asymptotic Bootstrap in Matrix Integrals
url: https://www.emergentmind.com/topics/asymptotic-bootstrap-method
type: topic
---

# Asymptotic Bootstrap in Matrix Integrals

Searching arXiv for the cited work and closely related papers.
The asymptotic bootstrap method is a non-perturbative computational procedure for reconstructing observables of matrix integrals from exact recursion relations supplemented by asymptotic information at large mode number. In the formulation developed for unitary matrix integrals at complex coupling, the method is a linear truncation-and-tail-control scheme that combines exact finite-difference recursions for Fourier moments with asymptotic control of their rapidly decaying tails, thereby producing very high numerical precision without relying on positivity or semidefinite programming [2602.18559]. In this sense it is a “bootstrap” because it propagates a finite amount of boundary data across infinitely many modes by means of exact identities, but it differs fundamentally from positivity-based bootstrap programs: it remains applicable when the coupling is complex and positivity is absent [2602.18559]. Closely related antecedents include the “shoestring bootstrap” for circle problems [2503.00104], while a parallel development for Hermitian matrix models shows analogous exponentially fast convergence in a cone of complex couplings [2606.30895].

## 1. Definition and conceptual role

In unitary matrix models, the asymptotic bootstrap estimate is designed to reconstruct Fourier moments of the one-eigenvalue weight and, from them, orthogonal polynomials on the unit circle (OPUC), partition functions, and Wilson loops [2602.18559]. The method starts from exact Schwinger–Dyson or loop-equation recursions for the moments, truncates the infinite system at a large mode \(M\), and closes the system by imposing an asymptotic tail condition informed by the exact large-mode decay rate [2602.18559]. This yields a banded linear system whose solution determines the low modes with exponentially small truncation error.

Its distinctive feature is independence from positivity. Positivity or semidefinite programming (SDP) methods constrain moment matrices through positive semidefiniteness, but such constraints become unavailable or inapplicable at complex ’t Hooft coupling [2602.18559]. The asymptotic bootstrap instead exploits the dichotomy between the physical decaying solution of the recursion and unphysical growing solutions. This suggests that the essential non-perturbative information is encoded in selecting the decaying branch by appropriate boundary conditions at infinity.

The method is therefore both numerical and structural. Numerically, it is a fast linear solver for moments and observables. Structurally, it identifies the physically admissible solution among all formal recursion solutions by enforcing the correct asymptotic sector [2602.18559]. A plausible implication is that the method belongs to a broader class of bootstrap procedures in which asymptotic admissibility replaces positivity as the primary selection principle.

## 2. Matrix-integral setting and exact recursion structure

The unitary models considered are single-trace unitary matrix integrals with \(U(N)\)-Haar measure,
$$
Z_N(\boldsymbol t) = \frac{1}{\mathrm{Vol}(U(N))}\int_{U(N)} dU\, \exp\!\Big(V(U)\Big),
$$
with potential
$$
V(U)= \sum_{k=1}^{K} \frac{t_k}{k}\,\Big(\mathrm{Tr}\, U^k + \mathrm{Tr}\, U^{-k}\Big),
$$
where the couplings \(t_k\) may be complex [2602.18559]. Two emphasized examples are the Gross–Witten–Wadia model, \(V(U)=t(\mathrm{Tr}\,U+\mathrm{Tr}\,U^{-1})\), and a quadratic single-trace model with first and second harmonics [2602.18559].

After diagonalization \(U=\mathrm{diag}(e^{i\theta_1},\dots,e^{i\theta_N})\), the basic objects become Fourier moments of the one-eigenvalue weight
$$
a_n=\int_{-\pi}^{\pi}\frac{d\theta}{2\pi}\,e^{i n \theta}\, e^{V(e^{i\theta})},
$$
which determine Toeplitz moment matrices and OPUC [2602.18559]. With monic OPUC \(p_n(z)\) defined by
$$
\langle f|g\rangle=\int_{S^1}\frac{dz}{2\pi i z}\, e^{V(z)}\, g(z)\,f(z^{-1}),
\qquad \langle p_m|p_n\rangle=h_n\,\delta_{mn},
$$
one has
$$
Z_N=\prod_{n=0}^{N-1} h_n,\qquad
W_n=\frac{1}{N}\sum_{k=0}^{N-1}\frac{\langle p_k|z^n p_k\rangle}{h_k},
$$
so the moment sequence is the computational backbone of the full finite-\(N\) problem [2602.18559].

The exact recursions arise from integration by parts on the circle. Writing
$$
V(\theta)=c_0+\sum_{k=1}^K \frac{2 c_k}{k}\cos(k\theta),\qquad c_k\equiv t_k,
$$
the moments satisfy, for all integers \(n\),
$$
-\sum_{k=1}^K c_k\, a_{n-k} + n\, a_n + \sum_{k=1}^K c_k\, a_{n+k}=0,
\qquad a_{-n}=a_n,\quad a_0=1
$$
[2602.18559]. Thus \(a_0\) and the \(K\) seeds \(a_1,\dots,a_K\) determine the full sequence. For \(K=1\),
$$
-c_1 a_{n-1}+ n a_n + c_1 a_{n+1}=0,
$$
and for \(K=2\),
$$
-c_2 a_{n-2}-c_1 a_{n-1}+ n a_n + c_1 a_{n+1}+c_2 a_{n+2}=0
$$
[2602.18559]. These identities are exact and do not invoke positivity.

## 3. Large-mode asymptotics and tail control

The decisive input is the asymptotic structure of the recursion at large mode number. For nondegenerate highest coupling \(c_K\), the physical decaying solution satisfies
$$
a_n \sim d_{n\,\mathrm{mod}\,K}\,\left(\frac{c_K}{K}\right)^{\lfloor n/K\rfloor}\frac{1}{\Gamma\!\left(\frac{n}{K}+1\right)},
$$
where the coefficients \(d_{(\cdot)}\) do not grow with \(n\) [2602.18559]. The decay is therefore faster than exponential. By contrast, a basis of growing solutions behaves as
$$
a_n^{\mathrm{grow}} \sim d'_{n\,\mathrm{mod}\,K}\,\left(\frac{K}{|c_K|}\right)^{\frac{n}{K}-1}\Gamma\!\left(\frac{n}{K}\right)
$$
[2602.18559]. This separation between rapidly decaying and rapidly growing sectors is the central asymptotic mechanism.

Because a small seed error \(\delta\) excites the growing solution with amplification roughly proportional to
$$
|a_n| \sim \delta \left(\frac{K}{|c_K|}\right)^{\frac{n}{K}}\Gamma\!\left(\frac{n}{K}\right),
$$
the boundary condition imposed at large \(n\) determines low-mode accuracy [2602.18559]. Two truncation strategies are contrasted. A “cheap truncation” merely enforces boundedness of the tail, for example \(|a_M|,\dots,|a_{M-K+1}|\le 1\), which leads to low-mode error scaling
$$
\delta_{\text{cheap}} \sim \left(\frac{|c_K|}{K}\right)^{\frac{M}{K}}\frac{1}{\Gamma\!\left(\frac{M}{K}\right)}.
$$
The asymptotic bootstrap, or “shoestring,” instead sets
$$
a_M=\cdots=a_{M-K+1}=0,
$$
matching the exponentially small physical tail, and produces a low-mode error parametrically equal to the square of the cheap one,
$$
\delta_{\text{asymp}} \sim \left(\frac{|c_K|}{K}\right)^{\frac{2M}{K}}\frac{1}{\Gamma\!\left(\frac{M}{K}\right)^2}
$$
[2602.18559].

This is the formal basis for the observed gain in precision. The low-mode estimates typically acquire about twice as many correct digits as under the cheap truncation [2602.18559]. Pre-asymptotic drift can occur because subleading recursion terms are only suppressed by fractional inverse powers of \(M/K\), but the exponential rate predicted by the large-mode analysis remains correct [2602.18559].

This tail-control viewpoint was foreshadowed by the “shoestring bootstrap” for distributions on the circle, which similarly used the asymptotic exponential smallness of high Fourier modes to bypass a full SDP search [2503.00104]. The unitary-matrix asymptotic bootstrap turns that observation into a systematic finite-\(N\) algorithm [2602.18559].

## 4. Algorithmic realization

The implementation is entirely linear. One fixes the number \(K\) of nonzero couplings and a large cutoff \(M\). The procedure begins with \(a_0=1\) and the symmetry \(a_{-n}=a_n\), imposes asymptotic tail boundary conditions \(a_M=\cdots=a_{M-K+1}=0\), and solves the exact recursion over modes \(n=1,\dots,M-K\) as a banded linear system [2602.18559]. Because the recursion bandwidth is \(2K+1\), the solve scales as \(O(MK)\) or \(O(MK^2)\), with small constants [2602.18559].

Once the moments are obtained, one forms the Toeplitz matrix \((a_{i-j})_{0\le i,j\le S}\), with \(S\ge N_{\max}+K\), and computes the monic OPUC by Gram–Schmidt [2602.18559]. This determines \(h_n\), the three-term recurrences if desired, and Wilson loops through the overlap formula. The naive complexity of OPUC construction up to degree \(N_{\max}\) is \(O(N_{\max}^3)\), though Toeplitz or OPUC structure can improve this [2602.18559].

Convergence is monitored by increasing \(M\) until changes in \(a_1,\dots,a_K\) or in \(W_n\) fall below a target tolerance [2602.18559]. The asymptotic error model predicts that the required precision is typically reached once \(M\gtrsim |c_K|\) [2602.18559]. Stability diagnostics include consistency under \(M\to M+\Delta M\), comparison to the asymptotic error envelope
$$
\sim \frac{(|c_K|/K)^{2M/K}}{\Gamma(M/K)^2},
$$
and conditioning estimates from eigenvalue ratios of cumulative transfer matrices \(R_n\) [2602.18559].

The method is particularly amenable to high-precision arithmetic. In the worked example \(V(\theta)=20\cos\theta+20\cos2\theta\), the computation achieved more than \(2{,}800\) correct digits at \(M\approx 1900\) [2602.18559]. The paper also notes the practical use of rational couplings \(t_k\) to avoid roundoff [2602.18559].

## 5. Wilson loops, instantons, and non-perturbative resolution

A central application is the computation of Wilson loop expectation values with sensitivity to exponentially small instanton corrections. In the strong-coupling limit with fixed \(t_k\) as \(N\to\infty\), perturbation theory predicts
$$
\widehat W_n=t_n,\qquad \widehat W_n=0\ (n>K),
$$
equivalently \(W_n=\tilde t_n\) in the ’t Hooft scaling \(\tilde t_k=t_k/N\) [2602.18559]. The moments reconstructed by asymptotic bootstrap and then fed into OPUC reproduce these large-\(N\) limits with residuals decaying as \(e^{-\mathrm{const}\cdot N}\), enabling direct numerical access to instanton sectors [2602.18559].

In the ungapped phase, the partition function has perturbative part
$$
Z_N^{(0)}(\boldsymbol t)=\prod_{k=1}^{K}\exp\!\Big(\frac{t_k^2}{k}\Big),
$$
and non-perturbative completion
$$
Z_N(\boldsymbol t)=Z_N^{(0)}(\boldsymbol t)+\sum_{n\ge 1} Z_N^{(n)}(\boldsymbol t),
$$
with \(n\) labeling tunneling eigenvalue or anti-eigenvalue pairs [2602.18559]. For the Gross–Witten–Wadia model in the ungapped regime \(\tilde t<1/2\), the two saddle points are
$$
z_{1,2}^\star=\frac{-1\pm\sqrt{1-4\tilde t^2}}{2\tilde t},
$$
and the instanton action is
$$
A(\tilde t)= -\sqrt{1-4\tilde t^2}+\operatorname{arccosh}\!\Big(\frac{1}{2\tilde t}\Big)
$$
[2602.18559]. The leading one-instanton contribution obeys
$$
\frac{Z_N^{(1)}(t)}{Z_N^{(0)}(t)}=e^{-2N A(\tilde t)}\sum_{m\ge 1} C_m(\tilde t)\,N^{-m},
$$
with explicit coefficients \(C_1,C_2,C_3\) given in the paper [2602.18559].

Wilson loops inherit a trans-series expansion,
$$
\langle \mathcal{O}\rangle =\langle \mathcal{O}\rangle_{\mathrm{pert}} +\sum_{n\ge 1}\mathcal{C}_n(\boldsymbol t)\, e^{-2nN A(\boldsymbol t)}\left(1+\sum_{m\ge 1} c_{n,m}(\boldsymbol t)\,N^{-m}\right),
$$
and the numerical bootstrap confirms this structure by the ratio
$$
Q_n(N)=\frac{\widehat W_n^{(1)}(N)}{\widehat W_n^{\mathrm{bootstrap}}(N)-\widehat W_n^{(0)}(N)},
$$
for which \(Q_n(N)\to1\) as \(N\) increases, with systematic improvement upon inclusion of successive \(1/N\) corrections [2602.18559]. The same pattern persists in the quadratic model, even when multiple instanton channels interfere [2602.18559].

This instanton sensitivity is one of the defining achievements of the method. Because the truncation error is itself exponentially small, the numerics can resolve exponentially tiny non-perturbative contributions rather than merely reproduce perturbative data [2602.18559].

## 6. Complex coupling, phase structure, and OPUC zeros

The absence of positivity becomes especially consequential in the complex ’t Hooft plane. The asymptotic bootstrap remains applicable there because it never required positivity in the first place [2602.18559]. In this setting the ungapped-phase instanton action \(A(g)\) defines two distinguished loci. Anti-Stokes lines satisfy \(\mathrm{Re}\,A(g)=0\), and Stokes lines satisfy \(\mathrm{Im}\,A(g)=0\) [2602.18559]. Anti-Stokes lines track where eigenvalue transport becomes energetically neutral and accurately follow phase boundaries, while Stokes lines signal changes in Lefschetz-thimble decomposition. The unitary-matrix analysis finds that some Stokes lines also serve as useful proxies for additional phase boundaries, especially near critical points [2602.18559].

Phase diagrams are mapped numerically by the zeros of \(p_N\) at large \(N\). These zeros accumulate on the planar support of the eigenvalue spectral density in the ’t Hooft limit, and plotting them reveals the topology of the spectral support, such as the number of cuts [2602.18559]. The OPUC–density correspondence is rigorous in Hermitian settings and is reported to work empirically and robustly for the unitary models studied [2602.18559].

For the GWW model, the method reproduces the known complex-coupling phase diagram: an ungapped phase, two one-cut phases centered at \(z=\pm1\), and a two-cut phase [2602.18559]. Anti-Stokes lines arise from explicit actions
$$
A_{\text{one-cut}^{\theta=0}}(g)=\sqrt{\frac{2}{g}\Big(\frac{2}{g}-1\Big)}-\tfrac{1}{2}\operatorname{arccosh}\!\Big(\tfrac{4}{g}-1\Big),
$$
and
$$
A_{\text{one-cut}^{\theta=\pi}}(g)=A_{\text{one-cut}^{\theta=0}}(-g)
$$
[2602.18559]. For the quadratic model, the observed phases include ungapped, one-, two-, three-, and four-cut regions, inferred from OPUC zero distributions guided by anti-Stokes and complementary Stokes lines [2602.18559].

A plausible implication is that the asymptotic bootstrap furnishes a non-perturbative mapping from exact recursion data to spectral geometry in coupling space, with OPUC zeros serving as the practical interface between finite-\(N\) computation and planar phase structure.

## 7. Relation to other bootstrap traditions, advantages, and limitations

The term “asymptotic bootstrap method” has multiple meanings across disciplines, but the matrix-integral version is technically distinct from statistical resampling bootstraps in econometrics or semiparametric inference. Here the “bootstrap” refers to recursive propagation of exact identities closed by asymptotic data, not to sampling with replacement. Its nearest relatives are circle and Hermitian moment bootstrap methods that likewise use asymptotic tail behavior to replace or supplement positivity-based feasibility conditions [2503.00104; 2606.30895].

Relative to positivity or SDP bootstrap approaches, the asymptotic bootstrap for unitary matrix integrals is linear, extremely fast, and applicable at complex couplings [2602.18559]. It produces high-precision estimates rather than certified bounds, although positivity checks on Toeplitz matrices can be performed a posteriori when the couplings are real [2602.18559]. Relative to exact integrability or resurgence methods, it is numerical rather than symbolic: saddle-point and instanton data are used for validation and interpretation, not for the primary computation [2602.18559].

The method also has explicit scope conditions. The analysis exploits the symmetry \(V(\theta)=V(-\theta)\), which implies \(a_{-n}=a_n\) and permits a one-sided tail condition [2602.18559]. Fully generic complex potentials with \(V(\theta)\neq V(-\theta)\) would require simultaneous control of positive and negative tails, a case not implemented in the paper [2602.18559]. Precision also depends on entering the asymptotic regime \(M\gtrsim |c_K|\) and on avoiding extreme hierarchies among the couplings \(c_k\) [2602.18559].

Finally, the method should not be conflated with positivity-free asymptotic bootstrap schemes in unrelated contexts unless the underlying mechanism is specified. In the matrix-model literature, the defining mechanism is the imposition of asymptotically correct decaying tails on exact moment recursions. That mechanism first appeared in simplified circle bootstrap problems [2503.00104], was developed here for unitary matrix integrals [2602.18559], and has since been mirrored for Hermitian matrix models, where convergence is asymptotically exponential in a coupling cone determined by the first subleading coupling [2606.30895]. This suggests a broader methodological principle: when exact recursions admit both physical decaying and unphysical growing branches, asymptotic selection can replace positivity as the organizing principle of non-perturbative bootstrap computation.

Source: https://www.emergentmind.com/topics/asymptotic-bootstrap-method