🔢 mathematics

The central heat trace on large compact classical groups

This paper establishes a full large- $N$ asymptotic expansion for the central heat trace on compact classical groups by utilizing a highest weights/partitions correspondence and a random surface representation, leading to new results on Casimir spectrum growth and rigorous formulations of Yang--Mills/Hurwitz and Yang--Mills/Gromov--Witten dualities on the two-dimensional torus.

Original authors: Thibaut Lemoine, Mylène Maïda

Published 2026-06-19

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Thibaut Lemoine, Mylène Maïda

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: Listening to the "Hum" of Giant Shapes

Imagine you have a giant, complex musical instrument. In mathematics, these instruments are called compact classical groups (like the Unitary, Orthogonal, or Symplectic groups). As the paper says, we are looking at these instruments as they get huge (as the size $N$ goes to infinity).

Every instrument has a specific "hum" or vibration pattern. In physics and math, this hum is described by something called the heat kernel. If you could take a snapshot of the instrument's vibration at a specific moment in time, you would get a number called the central heat trace. This number tells you about the overall energy and structure of the instrument.

The authors of this paper wanted to answer a big question: What happens to this "hum" when the instrument becomes infinitely large?

1. The Recipe for a Perfect Expansion (The "Large-N" Expansion)

Usually, when things get very big, they get messy and hard to calculate. However, the authors discovered a beautiful pattern. They proved that you can predict the "hum" of these giant instruments with extreme precision using a recipe book.

The Analogy: Imagine trying to describe the taste of a giant cake. Instead of eating the whole thing, you can describe it by adding up layers: "It's mostly vanilla, with a tiny bit of chocolate, a pinch of salt, and a whisper of nutmeg."
The Math: The authors showed that the "hum" can be written as a sum of terms:
$\text{Hum} = \text{Main Flavor} + \frac{\text{Second Flavor}}{N} + \frac{\text{Third Flavor}}{N^2} + \dots$
As the instrument gets bigger ( $N$ increases), the extra flavors get smaller and smaller, making the prediction incredibly accurate. This is called an asymptotic expansion.

2. The Secret Code: Partitions and Young Diagrams

How did they crack this code? They realized that the complex vibrations of these giant groups are actually controlled by something much simpler: integer partitions.

The Analogy: Think of a partition as a way of stacking blocks. You have a certain number of blocks, and you stack them in rows that get shorter as you go down. This creates a shape called a Young Diagram (which looks like a staircase).
The Discovery: The authors found a "translation dictionary" between the complex math of the giant groups and these simple block-stacking shapes.
- The "energy" of the instrument (called the Casimir number) turns out to be a simple calculation based on how these blocks are stacked.
- When the instrument gets huge, the rules for stacking these blocks become very stable and predictable. This stability is what allows the authors to write down their perfect recipe (the expansion) mentioned above.

3. The "Random Surface" Story

The paper takes a surprising turn by connecting these abstract vibrations to random surfaces (like crumpled pieces of paper or soap bubbles).

The Analogy: Imagine you are trying to wrap a gift (the giant group) using a piece of wrapping paper. Sometimes the paper has to be crumpled or folded (ramified coverings).
The Connection: The authors proved that the "hum" of the giant instrument is exactly the same as the average behavior of all possible ways to wrap a torus (a donut shape) with these crumpled papers.
- They created a "random surface" model where the number of folds and the size of the paper are chosen randomly.
- Surprisingly, if you average out all these random, crumpled donuts, you get the exact same number as the "hum" of the giant mathematical instrument.
- This confirms a famous idea in physics called Gauge/String Duality, which suggests that the math of forces (gauge theory) is secretly the same as the math of strings and surfaces. The authors made this idea rigorous and proved it works for these specific giant groups.

4. A New Kind of "Counting Law"

Finally, the paper looks at the spectrum (the list of all possible notes the instrument can play).

The Old Way (Weyl's Law): For normal shapes (like a drum or a sphere), the number of notes grows in a predictable, polynomial way (like $x^2$ or $x^3$ ). It's a steady, slow climb.
The New Discovery: For these giant groups, the number of notes grows exponentially fast.
- The Analogy: Imagine a normal drum where the notes double every time you hit it harder. Now imagine a "super-drum" where the number of notes doesn't just double; it explodes like a firework, growing much faster than anyone expected.
- The authors found that the number of notes follows a pattern similar to the famous Hardy-Ramanujan formula (which counts how many ways you can break a number into smaller parts). This is a "Cardy growth," a phenomenon usually seen in quantum physics, but here it appears in the geometry of these giant groups.

Summary

In short, this paper does three main things:

It writes a perfect recipe for calculating the vibrations of giant mathematical shapes as they get larger and larger.
It translates these complex vibrations into a simple game of stacking blocks (partitions) and crumpled paper (random surfaces).
It discovers that these giant shapes have a "spectrum" of notes that grows explosively fast, revealing a deep connection between pure geometry, random surfaces, and string theory.

The authors didn't just guess; they provided rigorous mathematical proofs that link these seemingly different worlds together.

Technical Summary: The Central Heat Trace on Large Compact Classical Groups

Problem Statement

This paper investigates the large- $N$ asymptotics of the central heat trace on compact classical groups $G_N \subset GL_N(\mathbb{C})$ . Specifically, the authors study the trace of the heat kernel $p_t = e^{\frac{t}{2}\Delta_{G_N}}$ acting on the Hilbert space of square-integrable central functions, $H_{G_N} = \{f \in L^2(G_N) : f(hgh^{-1}) = f(g)\}$ .

Unlike the standard heat trace on $L^2(G_N)$ , which admits known short-time asymptotics and large deviation results, the central heat trace has been less explored analytically. While theoretical physicists have studied its large- $N$ behavior via formal power series (notably in the context of Yang–Mills theory), rigorous quantitative expansions were lacking. The central object of study is:
$\text{Tr}(p_t) = \sum_{\lambda \in \widehat{G_N}} e^{-\frac{t}{2} c_2(\lambda)}$
where $c_2(\lambda)$ are the Casimir eigenvalues associated with irreducible representations $\lambda$ .

Methodology

The authors employ a multi-faceted approach combining representation theory, asymptotic analysis of random partitions, and enumerative geometry.

1. Highest Weights and Partition Correspondence

The core technical innovation is a highest weights/partitions correspondence adapted to the large-rank regime.

Stable Realization: The authors map the set of irreducible representations $\widehat{G_N}$ to tuples of integer partitions (and integers for unitary groups). For example, for $U(N)$ , representations are mapped to $(\alpha, \beta, n)$ where $\alpha, \beta$ are partitions and $n \in \mathbb{Z}$ .
Casimir Expansion: In these coordinates, the Casimir operator $-\Delta_{G_N}$ admits a stable expansion in powers of $1/N$ . The leading term corresponds to an "energy" observable, while the first-order correction is governed by the cut-and-join operator acting on the algebra of shifted symmetric functions.
Probabilistic Reformulation: The trace is rewritten as an expectation with respect to $q$ -uniform random partitions (where $q = e^{-t/2}$ ) and, in the unitary case, a discrete Gaussian distribution.

2. Asymptotic Expansion via Tail Estimates

Using the probabilistic reformulation, the authors derive a full asymptotic expansion.

They expand the exponential term $e^{-\frac{t}{2N} L_\tau}$ (where $L_\tau$ is the first-order correction) using Taylor series.
Rigorous control of the remainder terms is achieved through deviation inequalities for $q$ -uniform partitions (specifically, bounds on the tail probabilities of partition sizes and lengths).
This yields a full asymptotic series in powers of $1/N$ (or $1/N^2$ for unitary groups), where coefficients are explicit expectations of polynomials in the partition content $K(\lambda)$ and size $|\lambda|$ .

3. Random Surface Representation

The moments of the partition content $K(\lambda)$ are identified with Hurwitz numbers, which count ramified coverings of the torus.

The authors construct a measure $\rho_t$ on the space of ramified coverings of a torus, where the degree $n$ and number of ramification points $k$ are random variables (geometric and Poisson distributions, respectively).
This allows the heat trace to be expressed as an integral over these "random surfaces," providing a probabilistic realization of the Yang–Mills/Hurwitz duality.

Key Results

1. Full Large- $N$ Asymptotic Expansion (Theorem 1.1 / Theorem 3.1)

For any compact classical group $G_N$ (Unitary, Special Unitary, Special Orthogonal, Symplectic), the central heat trace admits an expansion:
$\text{Tr}(e^{\frac{t}{2}\Delta_{G_N}}) = a_0(t) + \frac{a_1(t)}{N} + \dots + \frac{a_p(t)}{N^p} + O_t(N^{-p-1})$

Coefficients: The coefficients $a_k(t)$ are explicitly expressed as expectations with respect to the $q$ -uniform measure on partitions.
Parity: For unitary groups ($U(N), SU(N)$), the expansion contains only even powers of $1/N$ . For orthogonal and symplectic groups, all powers appear.
Convergence: The leading term $a_0(t)$ recovers known limits from previous works [43, 17].

2. Random Surface Representation (Theorem 1.2 / Theorem 4.3)

The heat trace is represented as an integral over Hurwitz spaces (spaces of ramified coverings of the torus):

For $U(N)$ and $SU(N)$, the trace involves a coupling of two random coverings $X_1, X_2$ .
For other groups, it involves a single random covering $X$ .
The integrand depends on the Euler characteristic $\chi(X)$ and the degree of the covering, weighted by the measure $\rho_t$ . This provides a rigorous probabilistic counterpart to the formal Gross–Taylor Yang–Mills/Hurwitz duality.

3. Yang–Mills/Gromov–Witten Duality (Theorem 1.3)

The authors establish a direct link between the coefficients of the heat trace expansion and the generating function of Gromov–Witten (GW) invariants on a torus.

The coefficients $a_k(t)$ are explicit functionals of the GW generating function $Z^T_{q_t}$ .
This connects the spectral data of the heat kernel to topological sigma-models coupled to gravity, a relationship not previously observed in the literature.

4. Spectral Counting Law (Theorem 1.4)

The paper derives a new asymptotic counting law for the Casimir spectrum as $N \to \infty$ .

Unlike Weyl's law for fixed-rank manifolds (which predicts polynomial growth of the counting function $N(\Lambda) \sim \Lambda^{n/2}$ ), the large- $N$ Casimir spectrum exhibits exponential growth of the form:
$N_\tau(\Lambda) \sim a \Lambda^{-b} e^{c\sqrt{\Lambda}}$
This behavior is analogous to the Hardy–Ramanujan asymptotic for integer partitions and Cardy's formula in Conformal Field Theory.

Significance and Claims

The paper claims to provide the first rigorous, quantitative large- $N$ asymptotic expansion for the central heat trace on all classical families of compact groups. Its significance lies in three main areas:

Analytical Rigor: It moves beyond the formal power series approaches common in theoretical physics to provide convergent matrix integral results with explicit error bounds.
Unification of Dualities: It rigorously establishes the Yang–Mills/Hurwitz duality on a torus at the level of convergent integrals and extends this to a Yang–Mills/Gromov–Witten duality, linking spectral theory to enumerative geometry and topological string theory.
Spectral Theory: It reveals a distinct spectral regime for large-rank groups, characterized by exponential growth in the density of states, contrasting sharply with the polynomial behavior of Weyl's law in fixed dimensions.

The authors emphasize that while the duality for higher genus surfaces is known at a formal level, this work focuses on the torus to provide a self-contained, condensed proof that controls the duality at the level of spectral data. The extension to general surfaces is noted as a subject for future work.