Continuum limit of gauged tensor network states

This paper demonstrates that the continuum limit of specific gauged tensor networks is well-defined, yielding a new class of states suitable for the non-perturbative study of gauge theories directly in the continuum.

The Gist

Imagine the universe is built on a set of invisible, strict rules called gauge theories. These rules dictate how particles interact and ensure that the laws of physics stay consistent no matter how you look at them. Think of these rules like a massive, complex puzzle where every piece must fit perfectly with its neighbors. If you try to force a piece in the wrong way, the whole picture breaks.

For a long time, scientists have studied these puzzles using a "pixelated" approach, like a video game grid. They break space into tiny squares (a lattice) and solve the rules square by square. A recent breakthrough showed that a specific type of digital puzzle piece, called a Tensor Network, is perfect for solving these grid-based puzzles while strictly obeying the rules.

However, the real universe isn't made of pixels; it's smooth and continuous, like a flowing river. The big challenge has been: How do we take these perfect grid-based puzzle solutions and turn them into smooth, continuous river-like solutions without breaking the rules?

This paper, "Continuum limit of gauged tensor network states," by Gertian Roose and Erez Zohar, proposes a new way to do exactly that.

The Core Idea: From Grids to Smooth Rivers

The authors introduce a new mathematical tool they call gauged continuous tensor networks. Here is how they built it, using simple analogies:

1. The "Virtual" Shadow World
Imagine you are trying to describe a complex 3D object (the real universe) using a 2D shadow (the math). In their method, there is a "virtual" layer of invisible fields that act like a shadow puppet show. The real particles (matter) and the force fields (like electricity or magnetism) interact with these invisible shadows. The magic is that the shadows are set up in a way that forces the real world to obey the strict gauge rules automatically. You don't have to check the rules manually; the structure of the shadow ensures the rules are never broken.

2. Smoothing the Grid
Previously, scientists could only make these "shadow" networks work on a grid (like graph paper). This paper shows how to stretch that grid until the lines disappear, creating a smooth, continuous surface.

• The Analogy: Think of a digital image made of square pixels. If you zoom out enough, the jagged edges of the pixels disappear, and you see a smooth curve. The authors figured out the specific mathematical "zoom" that turns their grid-based puzzle pieces into a smooth, continuous shape that still obeys the universe's strict rules.

3. The "Gauss Law" Safety Net
In physics, there is a rule called the Gauss Law (part of the gauge theory) that acts like a safety net. It says that the total amount of "charge" entering a room must equal the total amount leaving, or the room must be empty.

• The authors prove that their new smooth, continuous shapes always respect this safety net. No matter how they tweak the math, the "charge" never gets lost or created out of thin air. This is crucial because it means their method describes physically possible states of the universe.

How They Check the Work

The paper also discusses how to actually use these new shapes to calculate things, like the energy of a system or how particles interact.

• The "Recipe" (Generating Functionals): To get answers, they use a mathematical "recipe" called a generating functional. Think of this as a master list of ingredients. If you want to know how two particles interact, you just tweak the recipe slightly and see how the result changes.
• The "Folding" Trick: Calculating these recipes in 3D (or 4D with time) is incredibly hard, like trying to solve a Rubik's cube while juggling. The authors propose a method to "fold" the problem down. They show that you can reduce the complex 3D calculation into a simpler 2D problem, and then even simpler 1D problems, until it becomes something manageable.
• The "Truncation" Safety Valve: In the real world, calculations can sometimes go wild and produce infinite numbers (divergences). The authors note that by limiting the size of their "virtual shadow" (a process called truncation), they naturally stop these infinities from happening, keeping the math clean and finite.

What This Means (According to the Paper)

The paper claims three main things:

1. Existence: They have successfully defined what these smooth, rule-abiding states look like mathematically.
2. Connection: They proved that these smooth states are the natural "continuum limit" of the grid-based states scientists already use. In other words, if you take the grid-based puzzle and make the squares infinitely small, you get exactly what they described.
3. Universality: Because the grid-based versions are known to be the most general way to describe these rules on a computer, the authors suspect their new smooth versions are the most general way to describe these rules in the real, continuous universe.

Summary

In short, this paper builds a bridge between the digital, pixelated way we currently simulate the universe and the smooth, continuous reality we observe. They created a new type of mathematical "puzzle piece" that flows smoothly like a river but is constructed so strictly that it can never break the fundamental laws of physics. This provides a new toolkit for scientists to study the universe's most complex interactions without getting stuck in the limitations of a pixelated grid.

Technical Summary

Technical Summary: Continuum Limit of Gauged Tensor Network States

Problem Statement
Gauge theories are fundamental to modern physics, describing interactions in the Standard Model and higher-form symmetries. However, their non-perturbative study is challenging due to rich phase diagrams and the non-local nature of gauge-invariant observables (e.g., Wilson loops). While lattice gauge theories have successfully utilized gauged projected entangled pair states (gPEPS) to provide a manifestly gauge-invariant framework, a rigorous continuum limit for these states has not been fully established. Existing continuum tensor network states, such as continuous matrix product states (CMPS) and continuous projected entangled pair states (CPEPS), describe matter fields but lack the explicit incorporation of local gauge invariance required for gauge theories. This work addresses the gap by constructing a well-defined continuum limit for gauged tensor networks, leading to a new class of states termed gauged continuous projected entangled pair states (gCPEPS).

Methodology
The authors propose a construction of gCPEPS that is explicitly Lorentz invariant and satisfies the Gauss law constraints. The methodology proceeds through several key steps:

1. Variational Definition: The state is defined on a spatial manifold $M$ with boundary $\partial M$ using a path integral over auxiliary complex scalar fields $\chi$ (virtual fields) coupled to physical matter fields $\phi$ and gauge fields $A$. The state is given by:
   $$|\psi_{B,V,J}\rangle = \int \mathcal{D}^2\chi \, B[\chi_{\partial M}] e^{-\int_M d^D x \, \hat{\mathcal{L}}[\hat{A}, \chi, \hat{a}^\dagger, \hat{b}^\dagger]} |0\rangle |s_E\rangle$$
   where $\hat{\mathcal{L}}$ is a generating Lagrangian containing a covariant derivative $\hat{D}_\mu$ acting on the virtual fields, and $|s_E\rangle$ is the electric singlet state. The boundary functional $B$ ensures the state is a group singlet.

2. Gauge Invariance Verification: The authors explicitly demonstrate that the state satisfies the Gauss law $\hat{G}^a(x)|\psi\rangle = 0$ if the generating Lagrangian and boundary functional are singlets under the gauge group $G$. This is shown by verifying the local symmetry transformation properties of the creation/annihilation operators and the gauge fields.

3. Operator Representation: To facilitate computation, the authors introduce an operator formalism where virtual degrees of freedom are represented as operators acting on a virtual Hilbert space. This involves defining a transfer matrix $\hat{T}_i$ and a virtual Hamiltonian $\hat{H}$ that couples the virtual fields to the physical gauge fields. In one dimension, this reduces to a gauged continuous matrix product state (gCMPS) with a truncated virtual Hilbert space, naturally regulating UV divergences.

4. Computation of Observables: Two methods are proposed for computing generating functionals and expectation values:

   • Perturbative Expansion: For states close to Gaussian (in the coupling $g$), the generating functional is expanded perturbatively. The authors outline how to renormalize these expansions using UV cutoffs or counterterms, particularly in dimensions $D \leq 3$.
   • Transfer Matrix Formalism: For non-Gaussian cases, a non-perturbative method is developed. This involves mapping the $D$-dimensional generating functional to the optimization of a $(D-1)$-dimensional eigenvalue problem for an effective operator. By iteratively reducing dimensions, the problem eventually maps to a 1D CMPS, which can be regulated via virtual Hilbert space truncation.

5. Derivation from Discrete gPEPS: The paper establishes that gCPEPS are indeed the continuum limit of discrete gauged PEPS. Starting with gPEPS defined via second-quantized operators on lattice links, the authors introduce local gauge symmetry by coupling to link variables. By taking the lattice spacing $a \to 0$ and rescaling fields appropriately, they recover the continuum gCPEPS definition, confirming that the proposed states are the natural continuum generalization of the most general gauge-invariant lattice states.

Key Contributions and Results

• Definition of gCPEPS: The paper defines a new class of manifestly gauge-invariant states in the continuum that generalize CPEPS to include local gauge symmetry.
• Proof of Continuum Limit: It is demonstrated that gCPEPS arise as the continuum limit of sequences of discrete gPEPS. Given recent proofs that gPEPS are the most general gauge-invariant states on the lattice, the authors argue that gCPEPS represent the most general gauge-invariant states in the continuum.
• Computational Framework: The authors provide two distinct frameworks for computing expectation values: a perturbative expansion suitable for weak coupling and a transfer matrix formalism that reduces the dimensionality of the problem, allowing for non-perturbative treatment.
• Dimensional Reduction and Regularization: The work highlights that the transfer matrix approach allows for the iterative reduction of the problem to lower dimensions. In 1D, the truncation of the virtual Hilbert space serves as a natural regulator for UV divergences, a feature that persists in higher dimensions through dimensional reduction without breaking gauge symmetry.
• Extension to Fermions: The formalism is noted to be extensible to fermionic matter by treating virtual degrees of freedom as Grassmann variables or by working directly in the operator formalism.

Significance and Claims
The paper claims to provide a rigorous framework for studying gauge theories directly in the continuum using tensor network methods. By establishing that gCPEPS are the continuum limit of gPEPS, the authors suggest these states are the most general gauge-invariant states in the continuum. The significance lies in offering a non-perturbative numerical ansatz that inherently respects gauge invariance, potentially bypassing the difficulties associated with gauge fixing and the non-local nature of Wilson loops in traditional perturbation theory.

The authors modestly note that while the existence of the continuum limit is proven, a rigorous proof that these are the most general states is left for future work. Additionally, the paper suggests future directions, including the investigation of integrable models within this framework, the study of fusion categories of gauged continuous tensor networks, and the application of these methods to gauged matrix product operators to study dualities in the continuum. The work does not claim immediate experimental applications but positions itself as a theoretical tool for the non-perturbative analysis of strongly interacting gauge theories.