Operator level soft edge to bulk transition in $\beta$-ensembles via canonical systems

This paper establishes a unified framework using canonical systems to prove that the stochastic Airy operator, which characterizes the soft edge of $\beta$-ensembles, converges to the stochastic sine operator, characterizing the bulk, under a suitable high-energy scaling limit via a coupling of their driving Brownian paths.

The Gist

Imagine you are a detective trying to understand the hidden patterns of a chaotic city. In the world of mathematics, this "city" is a system of random numbers called $\beta$-ensembles. These systems appear everywhere, from the energy levels of atoms to the spacing of prime numbers.

For a long time, mathematicians had two different maps for two different neighborhoods of this city:

1. The "Soft Edge" (The Airy Operator): This is the chaotic fringe of the city, where things are wild and unpredictable. It's like the edge of a storm.
2. The "Bulk" (The Sine Operator): This is the calm, dense center of the city, where things are orderly and repetitive. It's like the steady rhythm of a heartbeat.

Usually, these two neighborhoods were studied using completely different tools. It was like trying to compare a hurricane to a metronome using different rulers. The big question was: How does the wild edge smoothly turn into the calm center?

This paper, by Vincent Painchaud and Elliot Paquette, solves that mystery by building a universal translator.

The Universal Translator: "Canonical Systems"

The authors realized that both the wild edge and the calm center can be described using the same underlying language, called Canonical Systems.

Think of a Canonical System as a musical score.

• The Airy Operator (the edge) is a score written for a chaotic, improvising jazz band.
• The Sine Operator (the bulk) is a score written for a precise, marching orchestra.

Even though the instruments and the style are different, the paper shows that if you look at the sheet music closely enough, they are actually variations of the same song.

The Magic Trick: The "Time Warp"

To prove that the jazz band turns into the orchestra, the authors had to perform a magic trick called a scaling limit.

Imagine you are watching a movie of the jazz band playing.

• The Problem: The jazz band plays on a stage that stretches to infinity, while the orchestra plays on a tiny stage that ends at a wall. You can't compare them directly.
• The Solution: The authors invented a "Time Warp" (a mathematical time-change). They sped up the jazz band's time and compressed the stage.

As they sped up the time (letting a variable called $E$ go to infinity), something magical happened:

1. The chaotic, jagged notes of the jazz band began to smooth out.
2. The wild, unpredictable swings of the music started to average out.
3. Suddenly, the jazz band's improvisation morphed perfectly into the precise, rhythmic marching of the orchestra.

The "Brownian Motion" Connection

How did they make this happen? They used Brownian Motion (the mathematical model for random jitter, like a drunk person walking or pollen dancing in water).

• The Airy Operator is driven by a simple, one-dimensional random walk (like a drunk walking in a straight line).
• The Sine Operator is driven by a Hyperbolic Brownian Motion (like a drunk walking on the surface of a hyperbolic saddle, which curves away from them).

The authors created a coupling. Imagine they tied the drunk walking in the straight line to the drunk walking on the saddle with an invisible elastic rope. As they sped up the time, they showed that the "jitter" of the straight-line walker, when viewed through the right lens, looks exactly like the "jitter" of the saddle-walker.

Why Does This Matter?

Before this paper, if you wanted to prove that the edge of the city looks like the center, you had to use heavy, complicated machinery involving random matrices (giant grids of numbers).

This paper proves the connection at the level of the operators themselves. It's like proving that a river naturally flows into the ocean by watching the water molecules, rather than just measuring the water level at the start and end points.

The Big Takeaway:
The authors showed that the "Soft Edge" and the "Bulk" are not two different species of mathematical animals. They are the same animal wearing different masks. By using the "Canonical System" framework and a clever time-warp, they proved that as you zoom out and look at the big picture, the chaos of the edge naturally settles into the order of the bulk.

A Simple Analogy

Imagine a kaleidoscope.

• When you look through the edge of the lens, the patterns are sharp, jagged, and wild (The Airy/Edge).
• When you look through the center, the patterns are smooth, repeating, and symmetrical (The Sine/Bulk).

This paper is the instruction manual that explains exactly how the glass pieces inside the kaleidoscope shift to turn those jagged edges into smooth circles. It proves that the transition isn't a mystery; it's a predictable, mathematical dance.

Technical Summary

1. Problem Statement

The paper addresses a fundamental question in Random Matrix Theory (RMT): the relationship between the local eigenvalue statistics at the soft edge and the bulk of $\beta$-ensembles.

• The Soft Edge: Characterized by the Stochastic Airy Operator ($H_\beta$), a random Sturm-Liouville operator (Schrödinger type) defined on $L^2(0, \infty)$ with a potential containing white noise. Its eigenvalues converge to the Airy$_\beta$ point process.
• The Bulk: Characterized by the Stochastic Sine Operator, a random Dirac operator defined on $(0,1)$. Its eigenvalues converge to the Sine$_\beta$ point process.
• The Gap: While the convergence of the point processes (eigenvalue distributions) from the soft edge to the bulk is known (specifically, $2\sqrt{E}(\text{Airy}_\beta + E) \to \text{Sine}_\beta$ as $E \to \infty$), there was no rigorous proof of this transition at the operator level. The two operators belong to different classes (Sturm-Liouville vs. Dirac), making direct comparison difficult. The authors aim to unify these frameworks and prove the operator-level convergence.

2. Methodology

The authors employ the theory of Canonical Systems as a unifying mathematical framework.

A. Canonical Systems Framework

A canonical system is a first-order differential equation of the form:
$$J u' = -z H u$$
where $J = \begin{pmatrix} 0 & -1 \\ 1 & 0 \end{pmatrix}$, $z \in \mathbb{C}$, and $H(t)$ is a $2 \times 2$ coefficient matrix.

• Unification: Both the Stochastic Airy and Stochastic Sine operators can be represented as canonical systems.
  • The Sine operator corresponds to a canonical system with an invertible coefficient matrix (Dirac type).
  • The Airy operator corresponds to a canonical system with a degenerate (non-invertible) coefficient matrix (Sturm-Liouville type).
• Convergence Topology: The authors define convergence using the vague topology of the coefficient matrices $H(t)$, treating them as matrix-valued measures.

B. Key Technical Steps

1. Time Change and Scaling:
   The Airy system is defined on $(0, \infty)$ while the Sine system is on $(0, 1)$. The authors introduce a specific time-change bijection $\eta_E: [0, 1+\epsilon_E) \to [0, \infty)$ that maps the Airy domain to the Sine domain. This transformation is designed to counteract the natural decay of Airy solutions, converting "Airy-like" oscillations into "Sine-like" oscillations.

2. Polar Coordinates Decomposition:
   The solutions to the Airy equation are analyzed using polar coordinates $(\rho, \xi)$. The coefficient matrix of the time-changed Airy system is decomposed into two terms:

   • A slowly varying term involving the difference of radial and phase coordinates ($\Delta \rho, \Delta \xi$).
   • A highly oscillatory term involving the sum of coordinates ($\Sigma \rho, \Sigma \xi$).
     The authors prove that the oscillatory term vanishes in the vague limit due to rapid averaging, while the slowly varying term converges to the coefficient matrix of the Sine system.

3. Coupling of Brownian Motions:
   A critical innovation is the construction of a coupling between the real Brownian motion driving the Airy operator and the complex Brownian motion driving the Sine operator (via hyperbolic Brownian motion).

   • They construct a probability space where the driving noise of the Airy system, when transformed, becomes pathwise close to the driving noise of the Sine system.
   • This allows for pathwise convergence arguments rather than just distributional convergence, which is essential for controlling the spectral measures.

4. Asymptotic Analysis:
   The proof relies on deriving precise asymptotics for the solutions of the stochastic Airy equation as $t \to -\infty$ (which corresponds to the high-energy limit $E \to \infty$). They show that the phase of the solution becomes uniformly distributed on the unit circle, and the radial part behaves like a geometric Brownian motion.

3. Key Contributions and Results

Theorem 1: Vague Convergence of Coefficient Matrices

For a diverging sequence $E_n \to \infty$, the time-changed and scaled coefficient matrix of the Airy system, $\tilde{H}_{\beta, E_n}$, converges vaguely in probability to the coefficient matrix of the Sine system, $\tilde{R}_\beta$.

• This establishes the convergence of the operators in the sense of their defining differential equations.

Corollary 1.1: Convergence of Transfer Matrices

The transfer matrices (solutions to the canonical system starting from the identity) of the Airy system converge compactly to those of the Sine system. This implies the convergence of the solutions themselves.

Theorem 2: Spectral Convergence (Weyl-Titchmarsh Functions)

The Weyl-Titchmarsh functions (which encode the spectral data) of the Airy system converge compactly to those of the Sine system.

• Significance: Since the Weyl-Titchmarsh function is the Stieltjes transform of the spectral measure, this proves that the spectral measures of the shifted Airy operator converge vaguely to the spectral measure of the Sine operator.
• Point Process Convergence: Because the spectral weights (masses at eigenvalues) are shown to be independent of the eigenvalues and follow a specific Gamma distribution, the vague convergence of spectral measures implies the convergence of the eigenvalue point processes (recovering the known result that $2\sqrt{E}(\text{Airy}_\beta + E) \to \text{Sine}_\beta$).

Theorem 3: Asymptotics of Airy Solutions

The paper provides new asymptotic formulas for solutions to $H_\beta f = 0$ as $t \to -\infty$. Specifically, it characterizes the behavior of the solution in terms of a radial process $X(t)$ and a phase process $\xi_\beta(t)$, showing that the phase becomes uniformly distributed modulo $2\pi$.

4. Significance and Impact

• Unified Framework: The paper successfully demonstrates that Canonical Systems provide a single, robust framework to describe all local scaling limits of $\beta$-ensembles (Hard edge, Soft edge, and Bulk), regardless of whether the underlying operator is Schrödinger or Dirac type.
• Operator-Level Transition: It moves beyond the convergence of eigenvalue distributions to prove the convergence of the operators themselves. This is a stronger result that explains why the point processes converge.
• Methodological Innovation: The construction of a coupling between the driving Brownian motions of two distinct operator classes is a novel technique that may be applicable to other problems in random operator theory.
• Spectral Weights: The authors provide an intrinsic proof (via the canonical system framework) that the spectral weights of the Airy operator are independent Gamma random variables, a property previously deduced via comparison with finite matrix models.

5. Conclusion

Painchaud and Paquette have rigorously bridged the gap between the soft edge and the bulk of $\beta$-ensembles. By embedding both the Stochastic Airy and Stochastic Sine operators into the theory of canonical systems and utilizing a sophisticated coupling of Brownian motions, they proved that the soft edge operator converges to the bulk operator in the high-energy limit. This result unifies the spectral theory of random matrices and provides a deeper, operator-theoretic understanding of universality in RMT.