Matrix Bootstrap Approximation without Positivity Constraint

This paper proposes a positivity-constraint-free bootstrap approximation method for the Hermitian one-matrix model that numerically determines a self-consistent eigenvalue distribution and moments by minimizing least-squares errors, successfully reproducing exact Euclidean solutions and perturbative Minkowski results without encountering a sign problem.

The Gist

Imagine you are trying to understand the behavior of a massive, chaotic crowd of people (representing the "matrix" in physics). In the world of quantum physics, this crowd is made of invisible particles that interact in incredibly complex ways. Scientists want to predict how this crowd moves, but the math is so difficult that even supercomputers struggle to solve it, especially when the rules of the game change from "calm and orderly" (Euclidean) to "wild and oscillating" (Minkowski).

This paper introduces a new, clever way to solve these problems without getting stuck in the usual mathematical traps. Here is the breakdown using simple analogies.

1. The Problem: The "Sign Problem" and the "Positivity Wall"

In the past, scientists used two main tools to study these crowds:

• Monte Carlo Simulations: Imagine trying to predict the crowd's movement by taking a million random snapshots. This works great when the crowd is calm (Euclidean). But when the crowd gets wild (Minkowski), the snapshots start flipping between positive and negative values so fast that they cancel each other out. It's like trying to hear a whisper in a room where everyone is shouting "Yes!" and "No!" at the exact same time. This is called the Sign Problem, and it makes the calculation impossible.
• The Bootstrap Method: This is a smarter approach. Instead of taking random snapshots, you assume the crowd follows certain rules (like "you can't have negative people"). You use these rules to narrow down the possibilities until only one answer remains. However, this method relies on a rule called Positivity (everything must be positive). In the wild, Minkowski world, things aren't always positive. The positivity wall collapses, and the method fails.

2. The Solution: The "Eigenvalue Distribution" Map

The author, Reishi Maeta, proposes a new way to do the Bootstrap. Instead of relying on the "Positivity" rule, they focus on the Eigenvalue Distribution.

The Analogy:
Imagine the crowd isn't just a blur, but a specific shape of people standing on a line.

• The Old Way: You tried to guess the shape by checking if the total number of people was positive.
• The New Way: You assume there is a specific shape (a distribution map) that the crowd forms. You don't need to know the map perfectly at first. Instead, you guess a rough shape (like a polynomial curve) and then check if it fits the "Loop Equations."

What are Loop Equations?
Think of these as the laws of physics for the crowd. If you know how the crowd behaves at a small scale (like how many people are in a tiny circle), the laws of physics tell you exactly how they must behave in a larger circle. It's like a domino effect: if you know the first domino falls, you know the rest will follow a specific pattern.

3. How the New Method Works

The author's method is like a self-correcting puzzle:

1. The Guess: You start with a blank canvas and draw a rough curve (a polynomial) to represent the crowd's shape. You also guess a few key numbers (moments) that describe the crowd.
2. The Check: You run these guesses through the "Loop Equations" (the laws of physics). The equations tell you what the crowd should look like if your guess was right.
3. The Correction: You compare your drawn curve with what the laws of physics demanded. If they don't match, you tweak your curve and the numbers slightly.
4. The Loop: You repeat this until your curve and the laws of physics agree perfectly.

Why is this special?

• No Positivity Needed: Because you are just matching shapes and numbers, you don't need the "everything must be positive" rule. This allows the method to work in the wild, Minkowski world where things can be complex and negative.
• No Sign Problem: Since you aren't summing up millions of random positive/negative snapshots, the "shouting Yes/No" problem disappears.
• High Accuracy: The author tested this on known problems (Euclidean models) and found it matched the exact answers almost perfectly. Then, they tried it on the wild Minkowski models, where no one knew the answer, and it successfully reproduced the expected theoretical results.

4. The "One-Cut" Assumption

To make the math work for the wild Minkowski models, the author had to make one big assumption: that the crowd's shape is a single, continuous line (a "one-cut" structure) rather than a scattered mess.

The Metaphor:
Imagine the crowd is standing on a bridge. In the calm world, they stand on a straight, solid bridge. In the wild world, the bridge might be tilted or curved. The author assumes the bridge is still one single, connected piece, just tilted at a specific angle.

• The Result: This assumption worked beautifully for small changes in the physics.
• The Caveat: If the physics gets too wild (large coupling constants), the bridge might break or split into two pieces. The author admits that for extreme cases, this assumption might need to be re-evaluated, but for now, it's a powerful tool.

Summary

Reishi Maeta has built a new mathematical microscope.

• Old Microscopes broke when looking at the "wild" side of the universe because they required everything to be positive.
• This New Microscope ignores the positivity rule. Instead, it looks for a consistent shape that fits the laws of physics (Loop Equations).
• The Result: It can now see clearly into the "Minkowski" realm (real-time dynamics, like how the universe actually evolves) without getting lost in the noise of the Sign Problem.

This is a significant step forward because it opens the door to simulating complex quantum systems (like the IKKT matrix model, which tries to describe the birth of spacetime) that were previously impossible to calculate.

Technical Summary

Here is a detailed technical summary of the paper "Matrix Bootstrap Approximation without Positivity Constraint" by Reishi Maeta.

1. Problem Statement

The paper addresses the challenge of performing numerical computations on large-$N$ Minkowski-type matrix models.

• The Sign Problem: Standard Monte Carlo methods fail for Minkowski models (where the path integral weight is $e^{iS}$) due to the rapidly oscillating phase factor, which destroys the probabilistic interpretation required for importance sampling.
• Limitations of Existing Bootstrap Methods: The Matrix Bootstrap approach, which uses semidefinite programming (SDP) based on positivity constraints (e.g., $\langle \text{Tr}(M^\dagger M) \rangle \geq 0$), is highly effective for Euclidean models ($e^{-S}$). However, in Minkowski models, the weight is complex, causing positivity constraints to break down. Consequently, standard bootstrap methods cannot derive meaningful inequalities or bounds for Minkowski systems.
• The Gap: There is currently no robust numerical method to analyze large-$N$ Minkowski matrix models (such as the IKKT model or Minkowski-reduced gauge theories) that avoids the sign problem and does not rely on positivity.

2. Methodology

The author proposes a Bootstrap Approximation method based on Eigenvalue Distributions that bypasses positivity constraints entirely. The core idea is to enforce self-consistency between the eigenvalue distribution and the loop equations.

A. Theoretical Framework

1. Eigenvalue Distribution Ansatz: Instead of using positivity, the method assumes the existence of an eigenvalue distribution $\rho(\lambda)$ (or $\rho(z)$ in the complex plane for Minkowski cases) that generates the moments $w_n = \langle \text{tr} \phi^n \rangle$.
   $$w_n = \int \lambda^n \rho(\lambda) d\lambda$$
2. Polynomial Approximation: The unknown distribution $\rho(\lambda)$ is approximated by a finite-degree polynomial $\rho^{(P)}(\lambda) = \sum c_m \lambda^m$.
3. Self-Consistency via Loop Equations: The moments $w_n$ must satisfy the loop equations (Schwinger-Dyson equations). In the large-$N$ limit, these equations reduce higher-order moments to combinations of lower-order moments (specifically $w_1$ and $w_2$).
4. Optimization Strategy:
   • The method treats the polynomial coefficients ($c_m$), the support boundaries ($a, b$), and the low-order moments ($w_1, w_2$) as unknown variables.
   • It minimizes an objective function using the least-squares method:
     $$F = \sum_{n=0}^{\Lambda} r_n |w_n - w_n^{(P)}|^2$$
     where $w_n$ are derived from loop equations and $w_n^{(P)}$ are integrals of the polynomial ansatz.
   • By minimizing the difference, the method finds a self-consistent pair of $\rho(\lambda)$ and $w_n$ that satisfies the loop equations without requiring positivity.

B. Extension to Minkowski Models

To apply this to Minkowski models ($Z = \int d\phi e^{iS}$), the author introduces specific theoretical extensions:

• Master Field Interpretation: Utilizing the large-$N$ factorization and the "master field" conjecture, the eigenvalue distribution is reinterpreted as a density matrix $\hat{\rho}$ in a Hilbert space.
• Complex Support: Since the matrix $\phi$ is no longer Hermitian in the Minkowski weight, eigenvalues can be complex. The distribution is defined on a contour $\Gamma$ in the complex plane.
• Single-Cut Assumption: A crucial working hypothesis is that the resolvent $R(z)$ has a single cut structure. The eigenvalues are assumed to lie on a line segment $\Gamma = \{ e^{i\theta}\lambda \mid \lambda \in [-a, a] \}$, where $\theta$ is a phase angle determined numerically.
• Formulation: The moments are computed as $w_n = \int_\Gamma z^n \rho_M(z) dz$. The optimization simultaneously determines the polynomial coefficients, the support $a$, and the phase angle $\theta$.

3. Key Contributions

1. Positivity-Free Bootstrap: The paper introduces the first numerical bootstrap framework for matrix models that does not rely on positivity constraints, thereby opening the door to Minkowski-type models.
2. Resolution of the Sign Problem: By formulating the problem as a least-squares minimization of absolute differences ($|w_n - w_n^{(P)}|$), the method avoids the oscillatory integrals that plague Monte Carlo simulations.
3. Formal Extension to Minkowski Space: The author provides a rigorous mathematical framework (via the master field and density matrix interpretation) to define eigenvalue distributions for non-Hermitian Minkowski systems.
4. Consistency Checks: The paper develops robust consistency checks, including:
   • Residual Analysis: Checking if the solution satisfies loop equations beyond the training cutoff ($\Lambda$).
   • Positivity as a Diagnostic: Even though positivity is not used in the derivation, the resulting solution is tested against positivity constraints (Hankel matrix positivity) to verify its physical validity in Euclidean cases.

4. Numerical Results

The method was tested on the Hermitian one-matrix model with action $S = N \text{Tr}(\frac{1}{2}\phi^2 - \frac{g}{4}\phi^4)$.

• Euclidean Case ($e^{-S}$):

  • The method reproduced exact analytical solutions with very high accuracy for $g < 1/12$ (the critical coupling).
  • Endpoint-Fixed vs. Free: An "endpoint-fixed" ansatz (enforcing $\rho(\pm a)=0$) significantly improved accuracy near the boundaries compared to the free-endpoint approach.
  • Failure at Criticality: For $g > 1/12$, the optimization failed to converge to a consistent solution, correctly identifying the phase transition where the single-cut solution breaks down.

• Minkowski Case ($e^{iS}$):

  • The method successfully reproduced perturbative results for small coupling constants ($|g| \ll 1$).
  • The numerical results for the moments $w_2$ and the phase angle $\theta$ matched the formal perturbative expansion $w_2 = i - 2g - 9ig^2 + \dots$ with high precision.
  • The reconstructed complex eigenvalue distribution $\rho_M(z)$ matched the formal solution derived from the single-cut ansatz.
  • No Phase Transition Observed: Unlike the Euclidean case, no critical coupling $g_c$ was observed where the solution broke down, suggesting the Minkowski model may not have a phase transition in the same sense, or that the single-cut assumption holds over a wider range (though the author notes this requires further investigation).

5. Significance and Future Outlook

• Bridging Analytical and Numerical Gaps: This work provides a powerful tool to study large-$N$ limits of real-time dynamics in quantum field theories, which are currently inaccessible to Monte Carlo simulations.
• Applicability to String Theory: The method is directly applicable to the IKKT matrix model and BFSS matrix model, which are non-perturbative definitions of superstring theory and M-theory. These models are inherently Minkowski and large-$N$.
• Future Directions: The author proposes extending this method to multi-matrix models (like IKKT). The challenge lies in the increased number of unknowns; the author suggests moving from eigenvalue distributions to directly approximating master fields and density matrices as functions of two variables to manage the complexity.

In conclusion, Maeta's paper presents a paradigm shift in matrix model analysis, replacing the restrictive positivity constraints with a self-consistent eigenvalue distribution approach, thereby enabling the numerical study of Minkowski spacetime emergence and real-time quantum dynamics.