High-efficiency quantum Monte Carlo algorithm for extracting entanglement entropy in interacting fermion systems

arXiv:2409.20009 · cond-mat.stat-mech, cond-mat.str-el, quant-ph · Submitted 2024-09-30 · Read on arXiv

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: "High-efficiency quantum Monte Carlo algorithm for extracting entanglement entropy in interacting fermion systems".

Title and authors: Kai: So we're diving into the details of this paper, "High-efficiency quantum Monte Carlo algorithm for extracting entanglement entropy in interacting fermion systems." We talked about how they get hundreds of data points from one run; now let's look at what the actual method entails.

Mira: The authors describe their core technique as an incremental method that’s based on an incremental technique along physical parameters, focusing on smoothly connecting far-away distributions by inserting intermediate distributions to realize importance sampling.

Lev: From my point of view, implementing this incrementally in a real quantum simulation, like using PQMC as the example they gave, is doable in principle because it can be done generally in the same spirit as DQMC.

Kai: But they benchmarked this on the two dimensional square lattice Hubbard model to show how it works practically and to demonstrate the high efficiency compared to existing methods.

Mira: They also point out that when you are simulating a realistic scenario, like reweighting between parameter points f and f0, if those points are not close enough, you need those intermediate points inserted to split the reweighting process.

Lev: That means the success of the reweighting depends on how close those two parameter points are; if they aren't, you have to use that incremental insertion strategy.

Kai: They also show a sketch map of the lattice structure and their choice of entangled region in figure one, which gives us a visual idea of what they are measuring.

Mira: And when looking at the results, they show hundreds of entanglement entropy values across the parameter space for the Hubbard model with L equal to four and U equal to eight for Grover's method versus their new algorithm.

Lev: That comparison shows high controllability and validity for the error bar in a single calculation, which is what we need when you’re trying to build something that actually runs on real hardware.

Kai: So, this algorithm seems to be designed specifically to handle the complexity of sampling across these parameter spaces efficiently.

Mira: It seems like their main goal here was establishing a robust way to obtain high-quality entanglement entropy data for interacting fermions in a single simulation run, which addresses the heavy computational cost issue we have with these systems.

The paper's summary: Kai: Now let's look at what the paper summarizes—essentially, they are saying this method allows them to scan the entanglement entropy for exploring its relation against the Hamiltonian parameters across a wide range.

Mira: They show that in the absence of interaction, when U is zero, the system behaves like a metal with a Fermi surface whose scaling behavior is dominated by the leading term L ln L.

Lev: That leads into their findings around adding positive U, where the system turns into an insulator and they see that coefficient A in equation four vanishes, meaning an area law emerges.

Kai: And then they focus on the critical point around Gross-Neveu criticality, quantifying that critical behavior by identifying the correlation length exponent.

Mira: The results show that this scaling behavior is determined by the correlation length exponent, which they found to be similar to what you see in an O N phase transition.

Lev: That ties back into their work on the two dimensional square lattice Hubbard model, where they used these methods to identify features like Fermi surfaces and Goldstone modes through entanglement entropy scaling.

Kai: They also apply this idea to the pi-flux Hubbard model, where the associated phase transition is at a finite interaction strength U with Gross-Neveu universality class.

Mira: So, in summary, they are using these methods to reveal intrinsic physics in these interacting fermion systems, finding things like the Fermi surface when U goes to zero and Goldstone modes when U is large.

Lev: And they even designed an entanglement entropy derivative algorithm that aims for high precision by assigning partial derivatives of the weight function as observables.

Kai: It seems like the overall implication is that this method provides a ubiquitous tool for exploring the physics in these interacting fermion systems, linking different physical behaviors together through entanglement.

The paper's improvements: Mira: When we look at how they improve their approach, they focus on making sure the error in calculating the derivative using partial derivatives of Z with respect to a parameter f is smaller than ten to the minus four compared to direct calculation methods <ref:2409.20009#pg1>.

Kai: That high precision for the derivative is impressive because it means you can get much tighter constraints on how entanglement entropy responds to changes in the Hamiltonian parameters.

Lev: If you're trying to run this on real hardware, that level of precision is what makes a difference when you are pushing the limits of what you can measure accurately.

Mira: They also detail how they handle the free fermion limit, showing that while far from it, the system size-dependence of both observables in their new algorithm increases faster than others with system size.

Lev: That suggests good scaling behavior in that limit, which is important for understanding the ground state wavefunction structure when you are dealing with multi-choice filling at the Fermi surface.

Kai: They found that by adding a specific trial wavefunction correction, like a Perries phase factor on each hopping amplitude, they can adapt the half-filling condition to all even system sizes.

Mira: But they also noted a limitation: the goodness of fit gets worse for U between one and four because of inadequate projection at small U, but that problem disappears at large U when the gap becomes larger.

Lev: So, they’ve identified a specific regime where their method struggles, which is important information for anyone trying to deploy this on real hardware.

Kai: And looking further into the free fermion limit, they describe it using the Widom-Sobolev equation to describe the scaling behavior of ground state entanglement entropy.

Conclusion: Mira: To wrap up, this paper on "High-efficiency quantum Monte Carlo algorithm for extracting entanglement entropy in interacting fermion systems" shows that we can get a lot of entanglement entropy data in just one simulation run.

Kai: This capability lets us explore the phase diagram of the two dimensional square lattice Hubbard model by scanning the entanglement entropy behavior and revealing features like the Fermi surface and Goldstone modes.

Lev: From my perspective, they’ve also given us a high-precision derivative calculation that’s better than direct methods for understanding how entanglement entropy responds to changes in the Hamiltonian parameters.

Mira: And they've quantified the scaling behavior around Gross-Neveu criticality by using the correlation length exponent, which we found to be similar to an O N phase transition, which is a really useful connection.

Kai: So this method offers us a way to probe the intrinsic physics in these interacting fermion systems, whether it’s looking at the Fermi surface when U goes to zero or Goldstone modes when U is large.

Lev: And they also applied it to the pi-flux Hubbard model, which supports the conclusion that O N phase transitions apply to Gross-Neveu criticality.

Mira: Overall, this paper provides an efficient tool for exploring these systems and gives us a solid framework for understanding the physics through entanglement in condensed matter.

Kai: We should keep an eye on how this algorithm scales as they push it further into even larger systems, because that’s where the real potential lies.

Lev: Yeah, if we can get the computational complexity down to something manageable on actual hardware, then this becomes a very powerful tool for studying these complex many-body problems.

Weilun Jiang, *Gaopei Pan, *Zhe Wang, *Bin-Bin Mao, Heng Shen†

State Key Laboratory of Quantum Optics Technologies and Devices, Institute of Opto-Electronics, Shanxi University · Collaborative Innovation Center of Extreme Optics, Shanxi University Collaborative Innovation Center of Extreme Optics ct.qmat at Universitat Wuerzburg · Department of Physics, School of Science and Research Center for Industries of the Future at Westlake University · Institute of Natural Sciences, Westlake Institute for Advanced Study

cond-mat.stat-mech, cond-mat.str-el, quant-ph

Submitted: 2024-09-30

Updated: 2025-05-14

Comments: Main text: 7 pages, 4 figures. Supplementary Material: 7 pages, 7 figures

Journal ref: Communications Physics 9, 258 (2026)

DOI: 10.1038/s42005-026-02794-2

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 77/100

The gist: The paper proposes a high-efficiency fermionic quantum Monte Carlo algorithm designed to extract entanglement entropy in interacting fermion systems with low computational cost The key idea is to…

Key concepts

Entanglement Entropy (EE)
A measure quantifying how entangled a quantum system is. In this paper, it is used to probe different phases in interacting fermion systems, such as identifying the Fermi surface in metals or Goldstone modes near critical points.
Incremental Technique
The core algorithm uses an incremental approach to smoothly connect two distant states by inserting intermediate distributions. This allows the importance sampling to be realized, enabling the calculation of EE values at various points in physical parameter space efficiently.
Scaling Behavior
This refers to how a physical quantity, like entanglement entropy, changes as a system size increases or as interaction strength ($U$) is varied. Analyzing this scaling helps classify phases and identify critical points, such as the Gross-Neveu criticality.
Correlation Length Exponent ($ u$)
A critical exponent that describes how quickly correlations decay near a phase transition. The paper uses $ u$ to quantitatively characterize the critical behavior of entanglement entropy in interacting fermion systems.

Terminology

Summary

The paper proposes a high-efficiency fermionic quantum Monte Carlo algorithm designed to extract entanglement entropy in interacting fermion systems with low computational cost The key idea is to design specific update flows along a meaningful path in the physical parameter space This method offers the possibility for exploring the entanglement entropy in large-scale interacting-fermion systems

Algorithm Development

The core of the proposed fermionic quantum Monte Carlo algorithm is based on an incremental technique along physical parameters The key spirit is smoothly connecting two far-away distributions by inserting several intermediate distributions, thereby the importance sampling can be realized This technique allows the intermediate products of the incremental process to turn into EE values of different parameter points In a realistic simulation, the result of the reweighting is good only if the two parameter points f and f0 are close enough, i.e., the ratio is closed to 1 Otherwise, several intermediate points are inserted to split the reweighting process The algorithm benchmarked by calculating scaling behavior in a two-dimensional square lattice Hubbard model

Benchmark and Validation

The algorithm was applied to the two-dimensional square lattice Hubbard model as a benchmark The method generates hundreds of EE values in the parameter space, demonstrating high efficiency with respect to existing algorithms The transition between the Fermi surface and Goldstone modes is identified through the scaling behavior of EE The results show that critical behavior can be quantitatively characterized by the correlation length exponent Accuracy was verified by comparing results with other methods, showing consistency for small system size L = 4

Scaling Behavior Analysis

The study explores the scaling behavior of EE in the square lattice Hubbard model as a function of interaction strength U In the absence of U, the model behaves as a metal with a Fermi surface (FS), whose scaling behavior is dominated by the characteristic leading term LlnL When adding positive U, the system turns into an insulator and the coefficient A in Eq.(4) vanishes, showing an area law The critical behavior around Gross-Neveu criticality is quantified by the correlation length exponent ν

Derivative Calculation and Outlook

The algorithm is also designed to calculate the EE derivative at high precision using the expression ∂ ln Z/∂f = 1/Z ∂Z/∂f The error in calculating the derivative using this method is smaller than 10-4 compared to direct derivatives The paper concludes that this method offers a ubiquitous tool for exploring the physics in interacting fermion systems

Supplemental Material Details

The sketch map of the lattice structure and choice of the entangled region is provided in Fig. S1 The algorithm flow diagram shows two identical memory space for parameter sets fn and fn-1 The complexity analysis shows that the total computation complexity is O(βN3) The incremental method can potentially reach exponential magnitude of the increase of algorithm efficiency under certain conditions The projection length β is adjusted to be large enough to reach the ground state as close as possible for various choice of trial wavefunction The scaling behavior of ground state EE in the free fermion system is described by the Widom-Sobolev equation The final coefficient of A in Eq. (5) is 0.5 for theoretical result

Ground State Wavefunction at Free Fermion Limit

The ground state of the free fermion limit has degeneracy due to multi-choice for filling at the FS A simple perturbation theory was used to find a non-degenerate ground state wavefunction The choice of trial wavefunction varies with the compile environment, and these 20 eigenstates have different EEs The TBC is applied by adding Perries phase factor on each hopping amplitude as Eq.(S7) This kind of TBC leads to the half-filling condition, which adapts to all even system size The leading term coefficient of EE could emerge close to the thermodynamic limit when the TBC is fixed for all system sizes The goodness-of-fit becomes worse at U ∼ [1, 4] for both two scaling function forms shown in (d) because of inadequate projection at the small U region This problem disappears at large U since the gap becomes larger and the projection is adequate to reach the true ground state The leading term coefficient of EE could emerge close to its thermodynamic limit value 0.5 when TBC is applied, as shown in Fig. p. 3 S6 The result of the integral in Eq. p. 5(S9) is 8π

Widom-Sobolev Equation for Free Fermion Limit

The scaling behavior of ground state EE in free fermion system has been studied, where the pioneer work was concluded as Widom conjecture The crucial discovery is that in presence of the FS, the leading term of EE scales as L(d−1)lnL In two dimensional system, each point on the FS owns a chiral model contributing to the lnL term The n-order Renyi entropy has S(n)M ∼ n + 1/24n L(d−1)lnL (2π)(d−1) Z ∂omega/Z ∂Γ nx · np dSxdSp, where the result of the integral is 8π The final coefficient of A in Eq. (5) is 0.5 for theoretical result

Summary and Outlook

We report an efficient fermionic QMC algorithm with amount of EE data in once simulation to fix the difficulty of the heavy computational cost of the EE calculation in large-scale interacting fermion systems Our algorithm provides the opportunity to scan EE for exploring its relation against the Hamiltonian parameters The intrinsic physics in square lattice Hubbard model has been revealed via EE by our method, such as the FS in U → 0 limit and Goldstone modes in large U limit Furthermore, we apply it to the π-flux Hubbard model where the associated phase transition is at finite U with Gross-Neveu universality class Numerically, we find the scaling behavior of EE seems to be determined by the correlation length exponent, similar to the O(N) phase transition We also design an EE derivative algorithm to obtain high-precision derivative data Given that highly entangled matter plays an essential role in condensed matter and statistic physics, our method offers an ubiquitous tool for exploring the physics in the interacting fermion systems

Acknowledgment

We thank the helpful discussion with Wei Zhu, Yao Zhou, Peng Ye, Zi Hong Liu and Xiaofan Luo This work is supported by National Natural Science Foundation of China (Project No. 12404275 and 12222409) Z. p. 4W. and Z.Y. acknowledge the China Postdoctoral Science Foundation under Grants No. p. 42024M752898, the Scientific Research Project (No. p. 4WU2024B027) and the start-up funding of the Westlake University G. p. 4P. acknowledges the Wurzburg-Dresden Cluster ¨ of Excellence on Complexity and Topology in Quantum Matter - ct. p. 4qmat (EXC 2147, Project No. p. 4 390858490) H. p. 4 S. acknowledges the Royal Society Newton International Fellowship Alumni follow-on funding (AL201024) of UK The authors thank the high-performance computing center of Westlake University and the Beijng PARATERA Tech Co. p. 4,Ltd. for providing HPC resources

References

[1] R. Horodecki, P. p. 4 Horodecki, M. p. 4 Horodecki, and K. p. 4 Horodecki, Quantum entanglement, Rev. p. 4 Mod. Phys. 81, 865 (2009)

[2] L. Amico, R. Fazio, A. Osterloh, and V. p. 4 Vedral, Entanglement in many-body systems, Rev. p. 4 Mod. Phys. 80, 517 (2008)

[3] N. Laflorencie, Quantum entanglement in condensed matter systems, Physics Reports 646, 1 (2016), quantum entanglement in condensed matter systems

[4] B. Zeng, X. Chen, D.-L. Zhou, and X. p. 4-G. Wen, Quantum information meets quantum matter (Springer, 2019)

[5] H. Casini and M. p. 4 p. 4 Huerta, Universal terms for the entanglement entropy in 2+1 dimensions, Nuclear Physics B 764, 183 (2007)

[6] H. Casini and M. Huerta, Positivity, entanglement entropy, and minimal surfaces, Journal of High Energy Physics 2012, 87 (2012)

[7] P. Calabrese and A. p. 4 Lefevre, Entanglement spectrum in onedimensional systems, Phys. p. 4 Rev. A 78, 032329 (2008)

[8] E. Fradkin and J. p. 4 E. Moore, Entanglement entropy of 2d conformal quantum critical points: Hearing the shape of a quantum drum, Phys. p. 4 Rev. Lett. 97, 050404 (2006)

[9] Z. Nussinov and G. p. 4 p. 4 Ortiz, Sufficient symmetry conditions for Topological Quantum Order, Proc. p. 4 Nat. Acad. Sci. 106, 16944 (2009)

[10] Z. Nussinov and G. Ortiz, A symmetry principle for topological quantum order, Annals Phys. p. 4 324, 977 (2009)

[11] W. Ji and X.-G. Wen, Noninvertible anomalies and mappingclass-group transformation of anomalous partition functions, Phys. p. 4 Rev. Research 1, 033054 (2019)

[12] W. Ji and X.-G. Wen, Categorical symmetry and noninvertible anomaly in symmetry-breaking and topological phase transitions, Phys. p. 4 Rev. Research 2, 033417 (2020)

[13] L. Kong, T. Lan, X.-G. Wen, Z.-H. Zhang, and H. p. 4 Zheng, Algebraic higher symmetry and categorical symmetry: A holographic and entanglement view of symmetry, Phys. p. 4 Rev. Research 2, 043086 (2020)

[14] X.-C. Wu, W. Ji, and C. Xu, Categorical symmetries at criticality, Journal of Statistical Mechanics: Theory and Experiment 2021, 073101 (2021)

[56] Y. D. Liao, Controllable incremental algorithm for entanglement entropy in quantum monte carlo simulations, arXiv eprints (2023), arXiv:2307. p. 510602

[57] G. Pan, Y. Da Liao, W. p. 5 Jiang, J. D’Emidio, Y. p. 2 Qi, and Z. Y. Meng, Stable computation of entanglement entropy for twodimensional interacting fermion systems, Phys. p. 5 Rev. B 108, L081123 (2023)

[58] X. Zhou, Z. Y. Meng, Y. Qi, and Y. Da Liao, Incremental swap operator for entanglement entropy: Application for exponential observables in quantum monte carlo simulation, Phys. p. 5 Rev. B 109, 165106 (2024)

[59] Y.-M. Ding, J.-S. Sun, N. Ma, G. Pan, C. Cheng, and Z. p. 5 Yan Reweight-annealing method for calculating the value of partition function via quantum monte carlo arXiv e-prints (2024), arXiv:2403. p. 508642

[60] S. R. White, D. J. Scalapino, R. p. 5 L. Sugar, E. Y. Loh, J. E Gubernatis, and R T Scalettar Numerical study of the twodimensional hubbard model Phys Rev B 40, 506 (1989)

[61] S G and K S., E Auxiliary field monte-carlo for quantum many-body ground states, Annals of Physics 168, 1 (1986)

[70] Z. Deng, L. Liu, W. Guo, and H. p. 5 Lin Improved scaling of the entanglement entropy of quantum antiferromagnetic heisenberg systems Physical Review B 108, 125144 (2023)

[71] F Parisen Toldin, M Hohenadler, F F Assaad, and I F Herbut Fermionic quantum criticality in honeycomb and π-flux hubbard models: Finite-size scaling of renormalization-groupinvariant observables from quantum monte carlo Phys. Rev. B 91, 165108 (2015)

[72] Y Otsuka, S Yunoki, and S Sorella Universal quantum criticality in the metal-insulator transition of two-dimensional interacting dirac electrons Phys. Rev. X 6, 011029 (2016)

[73] J Helmes, L E Hayward Sierens, A Chandran, W WitczakKrempa and R G Melko Universal corner entanglement of dirac fermions and gapless bosons from the continuum to the lattice Phys. Rev. B 94, 125142 (2016)

[74] M A Metlitski, C A Fuertes, and S Sachdev Entanglement entropy in the o(n) model Phys. Rev. B 80, 115122 (2009)

[75] J Helmes and S Wessel Entanglement entropy scaling in the bilayer heisenberg spin system Phys. Rev. B 89, 245120 (2014)

[76] Z Wang, Z Deng, Z Wang, Y.-M. Ding, W Guo, and Z Yan Probing phase transition and underlying symmetry breaking via entanglement entropy scanning arXiv preprint arXiv:2409. p. 609942

[78] F Parisen Toldin, M Hohenadler, F F Assaad, and I F Herbut Fermionic quantum criticality in honeycomb and π-flux hubbard models: Finite-size scaling of renormalization-groupinvariant observables from quantum monte carlo Phys. Rev. B 91, 165108 (2015)

[79] B Swingle, Entanglement entropy and the fermi surface Phys. Rev. Lett. 105, 050502 (2010)

[86] Q.-C Tang and W Zhu, Critical scaling behaviors of entanglement spectra Chinese Physics Letters 37, 010301 (2020)

[87] X.-C Wu, C.-M. Jian, and C Xu Universal Features of HigherForm Symmetries at Phase Transitions SciPost Phys. p. 4 p. 4 11, 33 (2021)

[94] Z Nussinov and G Ortiz Sufficient symmetry conditions for Topological Quantum Order Proc. Nat. Acad. Sci. 106, 16944 (2009)

[95] Z Nussinov and G Ortiz A symmetry principle for topological quantum order Annals Phys. 324, 977 (2009)

[16] Q.-C Tang and W Zhu, Critical scaling behaviors of entanglement spectra Chinese Physics Letters 37, 010301 (2020)

[17] X.-C Wu, C.-M. Jian, and C Xu Universal Features of HigherForm Symmetries at Phase Transitions SciPost Phys. 11, 33 (2021)

[18] G Vidal, J. I. Latorre, E Rico, and A Kitaev Entanglement in quantum critical phenomena Phys. p. 4 Rev. Lett. 90, 227902 (2003)

[19] P Calabrese and J Cardy Entanglement entropy and quantum field theory Journal of Statistical Mechanics: Theory and Experiment 2004, P06002 (2004)

[21] M B Hastings, I Gonzalez, A B Kallin, and R G Melko Measuring renyi entanglement entropy in quantum monte carlo simulations Phys. Rev. Lett. 104, 157201 (2010)

[23] T Grover Entanglement of interacting fermions in quantum monte carlo calculations Phys. Rev. Lett. 111, 130402 (2013)

[54] J D’Emidio, Universal features of entanglement entropy in the honeycomb hubbard model Phys. Rev. Lett. 132, 076502 (2024)

[55] X Zhang, G Pan, B.-B Chen, K Sun, and Z. p. 5 Y Meng Integral algorithm of exponential observables for interacting fermions in quantum monte carlo simulations Phys. p. 5 Rev. B 109, 205147 (2024)

[63] J D’Emidio Entanglement entropy from nonequilibrium work Phys. Rev. Lett.

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

  1. Incremental Quantum Monte Carlo algorithm for fermionic systems can be used to explore entanglement entropy in large-scale interacting fermion systems by calculating hundreds of entanglement entropy values in parameter space from a single simulation run.

  2. The improved AI system can quantify the scaling behavior of entanglement entropy and its derivative at Gross-Neveu criticality by determining the correlation length exponent, which is crucial for classifying phase transitions.

  3. The algorithm can provide high precision data for entanglement entropy derivatives by assigning partial derivatives of the weight function to be observables, leading to smaller error bars compared to direct calculation methods.

  4. The system can efficiently explore the phase diagram of the two-dimensional square lattice Hubbard model by scanning entanglement entropy behavior in parameter space, revealing features like Fermi surface and Goldstone modes.

  5. The improved system can quantitatively characterize the scaling behavior of entanglement entropy in various phases, fitting results to equations like S2M to extract universal coefficients A and b.

  6. The algorithm can be utilized to explore the physics of scaling behavior around Gross-Neveu criticality in the pi-flux Hubbard model, supporting the conclusion that O(N) phase transitions apply to Gross-Neveu criticality.

Sources

Related papers