The Roaming Bethe Roots: An Effective Bethe Ansatz Beyond Integrability
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.
Kai: We're looking at the paper: "The Roaming Bethe Roots".
Mira: Here's the gist of the paper. We propose an effective Bethe ansatz (EBA) for solving quantum many-body systems near an integrable point by renormalizing Bethe roots to account for integrability-breaking interactions.
Kai: The paper introduces the effective Bethe ansatz (EBA) as a method to approximate the exact eigenstates of non-integrable models by retaining the functional form of the Bethe wave function while renormalizing the Bethe roots to account for integrability-breaking interactions.
Mira: This approach determines these effective roots by minimizing physically motivated cost functions, resulting in off-shell Bethe states that serve as approximate eigenstates.
Kai: The efficacy of the method is assessed using physical observables such as the energy eigenvalue, state fidelity, and bipartite entanglement entropy.
Title and authors: Kai: Let's start with the title and who wrote this, "The Roaming Bethe Roots: An Effective Bethe Ansatz Beyond Integrability." It suggests something moving or adapting the core concept of the bethe ansatz.
Mira: The authors are Wenlong Zhao, Yunfeng Jiang, Rui-Dong Zhu, and Peng Huanwu Center for Fundamental Theory folks. They bring different angles to this problem—Zhao and Jiang from Soochow University, Zhu from Jiangsu Key Laboratory at Soochow University, and Peng is from the Hefei Center for Fundamental Theory.
Lev: It's interesting that we have authors coming in from physics institutes and a fundamental theory center; that's a good mix for this kind of problem. I wonder if the theoretical setup they use is robust enough to handle real-world noise later on.
Kai: Yeah, the core idea is making these effective roots by minimizing some physically motivated cost functions. That’s the mechanism they’re using to bridge the gap between a perfect integrable system and a messy, non-integrable one.
Mira: It's about taking that trial wave function idea—the original spirit of the ansatz—and making it work for models like the XXZ chain when they aren't perfectly symmetric.
Lev: So, when we look at this paper, what I immediately think about is how much overhead this approximation adds. If we’re going to use this on real hardware, optimizing these cost functions and finding these effective roots is a significant computational step before we even get to the physics of the state itself.
The paper's summary: Kai: So what's the main takeaway from this paper about what they actually did? They introduced this EBA method to approximate exact eigenstates for non-integrable models by renormalizing those Bethe roots.
Mira: They explicitly distinguish between weak and strong integrability breaking deformations. Weak ones, like the J1-J2 model, where charges are conserved up to second order corrections, degrade the approximation slowly over a wide parameter range.
Lev: That distinction is crucial for us; if it's weak breaking, the approximation might hold for a decent chunk of our parameter space before we hit trouble. If it's strong breaking, things go way downhill fast.
Kai: Right, and they found that the rate of breakdown acts as a diagnostic for how strong the integrability breaking perturbation actually is in the system they are studying.
Mira: They also looked at specific points in their parameter space, like level crossing points or phase transition points, where they saw a sharp decline in fidelity for this EBA method.
Lev: A sharp decline suggests a quantum critical point might be hiding there; so if we use this as a probe for criticality, we need to make sure the method is stable enough to actually see that feature rather than just noise.
The paper's improvements: Kai: Now they talk about how they could improve this EBA approach. They aren't satisfied with just the basic setup; they suggest ways to make the ansatz itself more flexible.
Mira: They explore using a more general R matrix for XXZ type models, and optimizing both those Bethe roots *and* additional parameters, like beta and gamma. That lets them get significantly improved approximations for models where integrability is broken.
Lev: Optimizing those extra parameters means we're dealing with a larger optimization surface now; that’s going to require much more sophisticated numerical solvers than just minimizing one cost function for the roots alone.
Kai: And they also found that by incorporating inhomogeneous parameters into the transfer matrix, which is a bit different from what they started with, the fidelity can go up to one point nine nine for strong integrability breaking cases.
Mira: That's a big jump; it means we can handle long-range interactions better with this method. They also mention that they can use this structure to approximate systems with long-range interactions, which is really useful for modeling things that aren't just nearest neighbors.
Conclusion: Kai: So to wrap up, the EBA approach gives us a tool to characterize integrability breaking by watching how quickly the approximation breaks down as we change system parameters.
Mira: It’s a useful way to probe quantum criticality because you can see those sharp declines in fidelity when you hit specific points in the parameter space.
Lev: From an engineering standpoint, it tells us where we might need to invest more effort in error correction or optimization if we want to use this on actual hardware. The method itself is powerful, but the cost of getting those precise effective roots is high.
Kai: That’s true; the paper shows that for weakly breaking models, you can get good results over a broad range, and for strong breaking ones, it's definitely a diagnostic tool to see where the physics gets really complicated.
Mira: Overall, "The Roaming Bethe Roots: An Effective Bethe Ansatz Beyond Integrability" gives us a new representation of eigenstates by exploiting the algebraic structure inherited from integrability. It opens up avenues for applying this idea to spin-one models next <ref:2604.04627#pg3>.
Lev: I think looking at how these off-shell states behave dynamically, like in quench dynamics or transport, might be a useful next step for building error correcting codes based on this structure.
Wenlong Zhao, Yunfeng Jiang, ★ and Rui-Dong Zhu
School of Physical Science and Technology & Institute for Quantum Science, Soochow University · School of Physics & Shing-Tung Yau Center, Southeast University · Suzhou Municipal Key Laboratory of Quantum-Intelligence Fusion Computing, Suzhou Municipal Key Laboratory of Frontier Material Physics and Devices, Jiangsu Key Laboratory of Frontier Material Physics and Devices at Soochow University · Peng Huanwu Center for Fundamental Theory
cond-mat.stat-mech, hep-th, nlin.SI
Submitted: 2026-04-06
Updated: 2026-08-31
Comments: 7+4 pages
Journal ref: Chinese Phys. Lett. 43 (2026) 100001
DOI: 10.1088/0256-307X/43/10/100001
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 77/100
The gist: We propose an effective Bethe ansatz (EBA) for solving quantum many-body systems near an integrable point by renormalizing Bethe roots to account for integrability-breaking interactions.
Key concepts
- Effective Bethe Ansatz (EBA)
- A method to approximate exact solutions for non-integrable models by modifying the standard Bethe roots. It retains the functional form of the exact wave function but adjusts the roots to account for interactions that break integrability, allowing it to find approximate eigenstates.
- Integrability Breaking Regimes
- The study distinguishes between weak and strong breaking. Weak breaking leads to behavior close to integrable systems (like prethermalization), where EBA is expected to work well. Strong breaking causes generic non-integrable chaos, where the approximation deteriorates rapidly.
- Renormalized Bethe Roots
- The process of adjusting the standard Bethe roots by minimizing physically motivated cost functions. This adjustment accounts for integrability-breaking interactions, effectively moving from exact solutions to approximate 'off-shell' states that are better suited for non-integrable Hamiltonians.
- State Fidelity and Entanglement Entropy
- These physical observables are used to measure the quality of the EBA approximation. High fidelity (close to 1) and specific changes in entanglement entropy help characterize how well the EBA captures the true behavior of a quantum state near an integrable point.
Terminology
Summary
We propose an effective Bethe ansatz (EBA) for solving quantum many-body systems near an integrable point by renormalizing Bethe roots to account for integrability-breaking interactions. p. 0
The paper introduces the effective Bethe ansatz (EBA) as a method to approximate the exact eigenstates of non-integrable models by retaining the functional form of the Bethe wave function while renormalizing the Bethe roots to account for integrability-breaking interactions. p. 0 This approach determines these effective roots by minimizing physically motivated cost functions, resulting in off-shell Bethe states that serve as approximate eigenstates. p. 0 The efficacy of the method is assessed using physical observables such as the energy eigenvalue, state fidelity, and bipartite entanglement entropy. p. 0
Modeling Integrability Breaking Regimes
The study distinguishes between two types of deformations: weak and strong integrability breaking. Weakly integrability-breaking models exhibit approximate conservation laws and display behavior close to integrable systems, such as prethermalization. p. 1 In contrast, strongly integrability-breaking perturbations tend to produce generic non-integrable behavior, including chaotic level statistics and particle decay. p. 1 The authors expect the effective Bethe ansatz to remain accurate for weakly integrability-broken models over a relatively wide parameter window, whereas for strongly-broken integrability the approximation will deteriorate rapidly. p. 1
Optimization of Bethe Roots
The deformed Bethe roots are obtained by treating the off-shell Bethe state as a variational ansatz with parameters, which are then optimized by minimizing a suitable functional. For the ground state, this involves minimizing H0(λ; u j) ≡ ⟨u jH(λ)u j⟩⟨u ju j⟩ to find the rapidities denoted by u¯j. p. 1 For excited states, a modified functional H1(λ; u j) is minimized, where K1 ≫ 1 enforces orthogonality to the ground state. p. 1 This procedure can be iterated to find approximate eigenstates for higher excitations. p. 1
Performance on Concrete Models
The proposal is tested on the periodic Heisenberg XXX spin chain, H0 = HXXX = X L n=1 ⃗σn · ⃗σn+1. For weak integrability-breaking deformation H1 = X L n=1 ⃗σn · ⃗σn+2, the energy error for the ground state is within 0. p. 21% for L=4 to L=10. The fidelity of the ground state remains close to 1 in the region λ < 0. p. 25, with a sudden drop observed at the Majumdar-Ghosh (MG) point λ = 0. p. 25.
Improvements and Generalizations
The authors explore improvements to the ansatz by introducing more tunable parameters, such as a more general R-matrix for XXZ type models. Optimizing both the Bethe roots and these additional parameters can lead to significantly improved approximations, with fidelity reaching unity with an error of order 10−13 in some cases. p. 3 Furthermore, incorporating inhomogeneous parameters into the transfer matrix can increase fidelity F to F ≥ 0. p. 599 for all values of λ in the strong integrability-breaking case. p. 5
Future Directions
The work opens avenues for further development by considering different choices of cost functions and ansatzes for a given non-integrable model. Future research should explore applying this method to spin-1 models, such as those arising from deformations of the integrable Takhtajan–Babujian model. p. 5 A deeper physical understanding of why the off-shell Bethe ansatz breaks down for strong integrability-breaking interactions is desirable. p. 5 p. 5 The algebraic structure of off-shell Bethe states may also be useful for investigating dynamical properties like quench dynamics and quantum transport. p. 5
Algebraic Background
The appendix provides a concise review of the algebraic bethe ansatz (ABA) for the Heisenberg spin chain, starting with the 6-vertex R-matrix. The central object in ABA is the monodromy matrix Ma(u) defined as RaL(u − θL). p. 6.. Ra1(u − θ1). p. 6 The transfer matrix T(u) is obtained by taking the trace over the auxiliary space, and its fundamental property is that [T(u), T(v)] = 0. p. 6 This leads to the Hamiltonian of the XXX chain being derived from the transfer matrix of the homogeneous model. p. 7
Numerical Implementation Details
The numerical algorithm relies on SciPy for exact diagonalization and optimization, and PyTorch for constructing the effective Bethe state and computing energy expectation values. The procedure employs random initialization, parallel attempts, and selection of the best result among all runs to ensure robust convergence. p. 7 p. 7 A highly effective initialization for systems near the integrable point is achieved by numerically solving the BAE for the underlying integrable XXX model. p. 7 p. 7 The program has been checked up to L = 24, and it can run for L = 26. p. 8
Key Findings on Fidelity
The fidelity comparison shows that for the XXZ chain with ∆ = 1.1, the fidelity is F ≈ a + bL1 + cL2 with coefficients (a, b, c) found to be (1. p. 8000389, -1. p. 847e-4, -2e-6). p. 8 For the strong breaking model at λ = 0. p. 21, the fidelity is F ≈ 1 + bL1 + cL2 with coefficients (a, b, c) found to be (1. p. 8041342, -7. p. 8968e-3, -4. p. 847e-4). All fittings achieve R squared > 0. p. 8996. The no-opt curves provide the fidelity from the XXX Bethe states without optimization of Bethe roots. p. 8
Entanglement and Phase Transitions
The entanglement entropy shows an abrupt change at λ = 0.5 for the weak integrability-breaking model, which is also observed in the fidelity curve. p. 4 The distribution of the effective Bethe roots undergoes a sudden change in pattern when crossing this point. p. 2 This suggests that the effective Bethe ansatz may serve as a probe for potential quantum phase transitions, even in relatively short spin chains. p. 2 In the thermodynamic limit, the ground state and 1st excited state merge into a twofold-degenerate ground state in the dimerized phase. p. 2
Entanglement Structure at MG Point
The bipartite entanglement entropy of the state with large overlap with the EBA state varies between 1 and about 2 with an even-odd effect at the MG point. The EBA does not capture such an effect at this special point. p. 3 The ground states at the MG point are twofold-degenerate and dimerized, with specific entanglement entropy values depending on the cut position. p. 2 The fidelity revives to 1 when the magnetic field becomes large enough in the strong integrability-breaking model. p. 4 The fidelity is slightly improved towards 1 at a generic point for the weak integrability-breaking model when λ ≳ 0. p. 44.
Discussions and Outlook
The central idea is that sufficiently close to integrability, the functional form of the Bethe wave function remains valid while the Bethe roots undergo renormalization due to integrability-breaking perturbations. The authors suggest that different choices of cost functions may be suited to different purposes, such as starting from the XX point for EBA analysis. p. 5 A natural extension is applying the method to spin-1 models, such as those arising from deformations of the integrable Takhtajan–Babujian model. p. 5 A deeper physical understanding of why the off-shell Bethe ansatz breaks down for strong integrability-breaking interactions is desirable. The algebraic structure of off-shell Bethe states may prove useful for investigating dynamical properties such as quench dynamics and quantum transport. p. 5
Appendix Review
The appendix provides a concise review of the algebraic bethe ansatz (ABA) for the Heisenberg spin chain. It details the 6-vertex R-matrix and the monodromy matrix Ma(u). p. 6 The transfer matrix T(u) is derived from the trace over the auxiliary space, and its fundamental property is that [T(u), T(v)] = 0. p. 7 The Hamiltonian of the XXX chain is then derived from the transfer matrix of the homogeneous model. p. 7 The Bethe state is defined by rapidities satisfying the BAE. p. 7
Numerical Algorithm Summary
The numerical algorithm uses SciPy for exact diagonalization and optimization, and PyTorch for constructing the effective Bethe state. The procedure employs random initialization, parallel attempts, and selection of the best result among all runs to ensure robust convergence. A highly effective initialization for systems near the integrable point is achieved by numerically solving the BAE for the underlying integrable XXX model. This pipeline successfully produced high-fidelity approximations for systems near the integrable point. p. 8 The core computations rely on vectorized operations and reverse mode automatic differentiation in PyTorch for speed. p. 7
State Quality Metrics
The quality of the approximation is assessed by comparing energy expectation values E(EBA) with exact energies from ED. The fidelity F = ⟨ψexψEBA⟩ 2/(⟨ψEBAψEBA⟩⟨ψexψex⟩) is computed, where a fidelity close to 1 indicates a good approximation. p. 2 The variance Var(H)/∆E squared measures eigenstate quality with ∆E denoting the energy difference between the lowest and next-to-lowest levels in the same sector. p. 2 The coefficients of L 2-term are almost negligible except for the strong breaking model. p. 8 The no-opt curves give the fidelity from the XXX Bethe states without optimization of Bethe roots. p. 8
Conclusion and Outlook
The effective Bethe ansatz provides a useful probe for characterizing the strength of integrability breaking. The idea of approximating exact eigenstates of nonintegrable models by appropriate off-shell Bethe states offers numerous avenues for further development. p. 5 Progress in applying this method to spin-1 models is a natural extension. p. 5 A deeper physical understanding of the breakdown mechanism for strong integrability-breaking interactions is desirable. p. 5 The algebraic structure of off-shell Bethe states may prove useful for investigating dynamical properties like prethermalization and anomalous transport. p. 5 The paper concludes by emphasizing that the effective Bethe ansatz can be remarkably accurate with a suitable refinement of the ansatz. p. 5 The work provides a new representation of eigenstates for nearly integrable models, exploiting the algebraic structure inherited from integrability. p. 0 The paper is 2604. p. 004627.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Effective Bethe ansatz implementation for quantum many-body systems near integrability allows AI to approximate exact eigenstates of non-integrable models by renormalizing Bethe roots determined by minimizing physically motivated cost functions. This enables AI to predict energy eigenvalues, state fidelity, and bipartite entanglement entropy for models with weak integrability breaking.
-
Characterization of integrability breaking strength: The method provides a
useful probe for characterizing the strength of integrability breaking
because therate of breakdown thus serves as a diagnostic for the type of perturbations.
AI systems can use this diagnostic to classify physical interactions in complex quantum simulations. -
Quantum critical point detection: The effective Bethe ansatz can be used to find sharp declines in fidelity or sudden jumps in effective Bethe roots, suggesting it
can also be a useful probe for quantum criticality.
AI can search parameter spaces for these signatures to locate quantum phase transitions. -
Handling degenerate levels: When the target Hamiltonian has degenerate energy levels, the EBA method can be upgraded by using a "superposition of multiple off-shell Bethe states P n cn u (n) j>
as a variational ansatz to
lift the degeneracy.This allows AI to find correct linear combinations that lift degeneracies, such as identifying states like
u¯1 = 0.5⟩ ± iu¯1 = −0.5⟩" in small chains. -
Generalization of the ansatz for improved accuracy: The introduction of a more general 8-vertex R-matrix allows AI to optimize both Bethe roots and parameters β and γ, achieving
significantly improved approximation
for models like the XXZ chain with high fidelity, even when integrability is broken. This means AI can utilize richer algebraic structures inherited from integrability. -
Accounting for long-range interactions: Incorporating inhomogeneous parameters theta j into the transfer matrix allows the EBA to approximate
long-range interacting spin chains,
increasing fidelity F to F ≥ 0.99 in the strong integrability-breaking case and F ≥ 0.98 in weakly breaking models. AI can model and simulate systems with long-range correlations more accurately than with nearest-neighbor approximations.
Sources
- Boosting Nearest-Neighbour to Long-Range Integrable Spin Chains
- Long-Range Deformations for Integrable Spin Chains
- Approximate conservation laws in perturbed integrable lattice models
- Quasi-conserved quantities in the perturbed XXX spin chain
- Finite-size generators for weak integrability breaking perturbations in the Heisenberg chain
- Prethermalization
- Generalized Gibbs ensemble prediction of prethermalization plateaus and their relation to nonthermal steady states in integrable systems
- Prethermalization and universal dynamics in near-integrable quantum systems
- Prethermalization and Thermalization in Isolated Quantum Systems
- Spin transport in a one-dimensional anisotropic Heisenberg model
- Multiple relaxation times in perturbed XXZ chain
- Spin conductivity in almost integrable spin chains
- Scaling of electrical and thermal conductivities in an almost integrable chain
- Spin and thermal conductivity of quantum spin ladders
- Heat conductivity of the Heisenberg spin-1/2 ladder: From weak to strong breaking of integrability
- Non-Equilibrium Dynamics and Weakly Broken Integrability
- Non-integrable Quantum Field Theories as Perturbations of Certain Integrable Models
- Ising field theory in a magnetic field: analytic properties of the free energy
- Bethe ansatz on a quantum computer?
- Estimating Bethe roots with VQE
Related papers
- Emergent long-tail dynamics in driven magnets with dynamical frustration
- Statistical Mechanics of Dynamical System Identification
- Second law of thermodynamics in closed quantum many-body systems
- Chemical potentials from structure factors: II. Charged multi-component mixtures
- Genuine quantum scars in Floquet chaotic many-body systems
- Stochastic Thermodynamics for Autoregressive Generative Models: A Non-Markovian Perspective