Higher-Order Structure of Hamiltonian Truncation Effective Theory

This paper advances the Hamiltonian truncation effective theory for two-dimensional $\lambda\phi^4$ by deriving all-order local matching corrections through diagram resummation and computing next-to-next-to-local non-local corrections at $\mathcal{O}(E_{\rm max}^{-4})$, thereby demonstrating the necessity of an increasingly rich operator basis to systematically mitigate truncation artifacts.

The Gist

Imagine you are trying to predict the weather for a specific city. To do this perfectly, you would need to know the position, speed, and interaction of every single air molecule in the entire atmosphere, from the ground up to the edge of space.

If you tried to calculate this on a computer, your machine would melt instantly. There are simply too many variables.

The Problem: The "Too Much Data" Dilemma
In physics, scientists face a similar problem when studying the fundamental building blocks of the universe (Quantum Field Theory). They want to know how particles interact, but the math involves an infinite number of possible energy states. To make the math solvable, physicists use a method called Hamiltonian Truncation.

Think of this like taking a high-resolution photo and shrinking it to a low-resolution thumbnail. You throw away all the "high-energy" details (the tiny pixels) and only keep the "low-energy" ones (the big shapes). You can then run your calculations on this simplified version.

The Flaw: The "Blurry" Thumbnail
The problem is that by throwing away the high-energy details, you introduce errors. It's like trying to guess the shape of a mountain just by looking at a blurry, low-res photo; you might miss the sharp peaks or the deep valleys. In physics, this means your predictions for particle masses or energy levels are slightly wrong.

Usually, scientists fix this by making the "thumbnail" higher resolution (keeping more data), but that requires supercomputers that are too expensive or slow.

The Solution: The "Smart Correction" Kit (HTET)
This paper introduces a clever new toolkit called Hamiltonian Truncation Effective Theory (HTET). Instead of just throwing away the high-energy data, this method asks: "Can we mathematically guess what the missing data would have done, and add a 'correction sticker' to our low-res model to fix the errors?"

The authors developed two major upgrades to this toolkit:

1. The "Infinite Loop" Fix (Resummation)

Imagine you are trying to fix a leaky roof. You could patch one hole, then another, then another. But what if the holes are connected in a complex pattern where fixing one affects the others?

In the old method, physicists would calculate these "patches" one by one (Order 1, Order 2, Order 3...). The authors realized that for certain types of errors, these patches follow a repeating pattern, like a loop in a video game. Instead of calculating the loop 100 times, they found a mathematical shortcut (a "resummation") that calculates the entire infinite loop in one go.

• The Analogy: Instead of counting every single grain of sand on a beach to estimate the weight, they found a formula that tells them the total weight based on the beach's shape. This makes the "correction stickers" much more accurate and efficient.

2. The "Long-Distance" Fix (Non-Local Corrections)

Sometimes, the error isn't just a small patch; it's a ripple that travels across the whole system. In physics, this is called a "non-local" effect.

The authors realized that to get truly precise results, they couldn't just look at the "blurry photo" (the truncated model). They had to do the heavy math in the "real world" (infinite volume) first to figure out exactly how the ripples behave, and then apply those rules to the blurry photo.

• The Analogy: Imagine trying to fix a wobbly table.
  • Old Way: You just put a piece of paper under the short leg. It helps a little.
  • New Way: You first study how the floor is sloped in the entire house (the infinite volume), calculate exactly how the table should sit, and then cut a custom, perfectly shaped wedge for the leg. This ensures the table is stable even if the floor is uneven.

What Did They Find?

The team tested these new "correction stickers" on a simulated universe (a 2D version of a theory called $\lambda\phi^4$).

• The Surprise: They found that just adding the "Infinite Loop" fix (the first upgrade) didn't always make things better immediately. Sometimes, fixing one part of the error made another part look worse, like over-correcting a steering wheel.
• The Real Win: The magic happened when they combined the "Infinite Loop" fix with the "Long-Distance" fix. When they added the complex, non-local corrections, the results became incredibly stable. The predictions stopped fluctuating wildly and settled into a smooth, accurate line, even when using a "low-resolution" computer model.

The Bottom Line

This paper is a guide on how to get high-precision physics results without needing a supercomputer.

By using these new mathematical "correction stickers," scientists can now simulate complex quantum systems with much less computing power. It's like learning how to bake a perfect cake using a simple hand mixer instead of a $10,000 industrial oven, simply because you learned the secret recipe for adjusting the ingredients.

This is a huge step forward for understanding the universe, potentially helping us simulate things that are currently impossible to calculate, like the behavior of matter in the early universe or inside black holes.

Technical Summary

1. Problem Statement

Hamiltonian Truncation (HT) is a non-perturbative method used to study Quantum Field Theories (QFTs), particularly in regimes where standard perturbation theory fails (e.g., strong coupling, real-time dynamics). The method involves restricting the Hilbert space to states with energy below a UV cutoff $E_{\text{max}}$ and diagonalizing the truncated Hamiltonian.

However, HT suffers from two main limitations:

1. Computational Cost: The size of the truncated basis grows exponentially with $E_{\text{max}}$, making high-precision calculations expensive.
2. Truncation Artifacts: Simply discarding high-energy states introduces errors that scale as power laws in $1/E_{\text{max}}$. While the "raw" truncation ($H_{\text{raw}} = PHP$) converges, the convergence is slow.

Hamiltonian Truncation Effective Theory (HTET) addresses this by treating the cutoff $E_{\text{max}}$ as an effective field theory scale. It integrates out high-energy states ($E > E_{\text{max}}$) to generate an effective Hamiltonian ($H_{\text{eff}}$) acting on the truncated subspace, containing corrective terms organized in an expansion of $1/E_{\text{max}}$.

The specific gap this paper addresses: Previous HTET work (Ref. [31, 32]) established leading-order (LO) local corrections and next-to-leading-order (NLO) non-local corrections. This paper aims to:

• Improve the convergence of local corrections by deriving all-order resummations.
• Extend the non-local sector to Next-to-Next-to-Leading Order (NNLO), specifically at $O(E_{\text{max}}^{-4})$, and resolve technical ambiguities regarding distributional structures in finite volume.

2. Methodology

The authors apply the HTET framework to the two-dimensional $\lambda\phi^4$ theory defined on a circle of circumference $L=2\pi R$.

A. All-Order Resummation of Local Terms

• Approach: Instead of calculating corrections order-by-order in the coupling $\lambda$, the authors identify infinite classes of Feynman diagrams that share a fixed topology (e.g., specific loop structures contributing to the mass and quartic coupling).
• Technique: Within the local approximation (where external energies are negligible compared to $E_{\text{max}}$), these infinite series of diagrams are summed analytically.
• Result: This yields compact, closed-form expressions for the matching corrections to the mass ($\delta \tilde{m}^2$) and the quartic coupling ($\delta \tilde{\lambda}$) that capture dominant high-energy effects to all orders in perturbation theory.

B. Derivation of NNLO Non-Local Corrections

• Order: The authors target corrections scaling as $O(E_{\text{max}}^{-4})$. This includes sub-leading components of $O(V^2)$ non-local terms and genuine $O(V^3)$ local effects.
• Diagrammatic Classification: They classify diagrams with 2 and 4 external legs into distinct topologies.
• Continuum-First Matching: A critical methodological shift is the continuum-first approach.
  • Problem: In finite volume, the spectrum is discrete. Derivatives of the Heaviside step function (arising from the cutoff constraint) lead to ambiguous numerical prescriptions (e.g., how to handle $\delta(2\omega_k - E_{\text{max}})$ when $2\omega_k$ never exactly equals $E_{\text{max}}$).
  • Solution: The matching coefficients are computed in infinite volume (continuous momentum space) where distributions are well-defined. Only after determining the coefficients is the spatial direction compactified to implement the operators on the finite-dimensional Hilbert space.
• Operator Basis: The expansion reveals that NNLO corrections require an enriched operator basis. Unlike LO/NLO, NNLO terms depend irreducibly on individual external frequencies ($\omega_i$), necessitating higher-derivative operators (e.g., $H_0^2 : \phi^4 : H_0^2$) and terms involving the sum of squared frequencies ($\Omega_4 = \sum \omega_i^2$).

3. Key Contributions

1. Compact All-Order Expressions: Derivation of resummed formulas for mass and coupling corrections (Eqs. 20 and 27) that sum infinite diagrammatic classes, providing a more efficient alternative to perturbative truncation.
2. NNLO Non-Local Framework: A systematic derivation of $O(E_{\text{max}}^{-4})$ non-local corrections, including the identification of necessary higher-derivative operators and the mapping of energy factors to commutators with the free Hamiltonian ($H_0$).
3. Resolution of Finite-Volume Ambiguities: The demonstration that matching coefficients involving derivatives of step functions must be calculated in the continuum to avoid numerical artifacts, followed by a consistent re-compactification.
4. Vacuum Sector Resummation: Showing that in the vacuum sector (zero external legs), the entire tower of non-local $1/E_{\text{max}}$ corrections resums exactly into an operator-valued function of the free Hamiltonian: $B_0(E_{\text{max}} - H_0)$.

4. Results and Numerical Analysis

The authors tested their improved Hamiltonians on the first spectral gap ($\Delta E_1$) and higher gaps ($\Delta E_5$) for various cutoffs $E_{\text{max}}$.

• Resummed Local vs. LO: Surprisingly, the resummed local prescription did not show a systematic improvement over the standard LO (fixed-order) result in the tested range.
  • Reason: The resummation captures only a specific subset of diagrams (fixed topology) within the local approximation. At the same order in $1/E_{\text{max}}$, non-local terms and other diagram topologies become numerically comparable. Resumming only the local subset can lead to "over-correction" and slower convergence.
• NNLO Non-Local Performance: The inclusion of NNLO non-local corrections (Table I) provided the most significant improvement.
  • Compared to Raw, LO, and NLO, the NNLO results exhibited a markedly flatter dependence on $E_{\text{max}}$.
  • While NNLO curves showed more variation at very low cutoffs (especially at strong coupling), they converged rapidly to a stable plateau at intermediate $E_{\text{max}}$, becoming indistinguishable from NLO results at higher cutoffs but with better stability at lower ones.
• Finite Volume Effects: Numerical comparisons confirmed that increasing the spatial volume ($L$) suppresses residual finite-volume effects, allowing the discrete spectrum to converge smoothly to the continuum limit.

5. Significance

• Feasibility of Systematic HTET: The paper confirms that a systematic HTET expansion beyond the leading local approximation is feasible and necessary for high precision.
• Operator Basis Complexity: It highlights that improving precision requires not just higher-order coefficients, but an increasingly rich operator basis that includes higher-derivative structures to account for the irreducible dependence on external frequencies.
• Methodological Standard: The "continuum-first" matching procedure is established as a robust standard for handling distributional coefficients in truncated theories, avoiding the pitfalls of discrete sampling of derivatives.
• Future Outlook: The work provides the theoretical and numerical foundation for extending HTET to higher orders and other QFTs (including higher dimensions), aiming to achieve high-precision results at moderate computational costs.

In summary, the paper demonstrates that while all-order resummation of local terms alone is insufficient due to the dominance of non-local effects at the same order, the systematic inclusion of NNLO non-local corrections significantly enhances the convergence and reliability of Hamiltonian Truncation calculations.