pith. sign in

arxiv: 2303.13403 · v1 · submitted 2023-03-23 · ❄️ cond-mat.str-el · cond-mat.stat-mech

Boosting quantum Monte Carlo and alleviating sign problem by Gutzwiller projection

Wei-Xuan Chang , Zi-Xiang Li This is my paper

Pith reviewed 2026-05-24 09:56 UTC · model grok-4.3

classification ❄️ cond-mat.str-el cond-mat.stat-mech
keywords quantum Monte CarloGutzwiller projectionsign problemprojective QMCvariational Monte Carlofermionic systemsstrongly correlated electronsHubbard model
0
0 comments X

The pith

Gutzwiller projection trial wave functions accelerate convergence and reduce the sign problem when used inside projective determinant quantum Monte Carlo.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a hybrid method called Gutzwiller projection QMC that takes a minimum-energy wavefunction obtained from variational Monte Carlo with Gutzwiller projection and inserts it as the trial state for the projective determinant QMC projector. Numerical tests show this choice makes observables converge much faster than with conventional trial states, cutting the required simulation length. The same construction also raises the average sign substantially, especially in parameter regions where the sign problem is otherwise prohibitive. The authors conclude that the approach opens a practical route to more efficient zero-temperature simulations of interacting fermions.

Core claim

By replacing the usual trial wavefunction with a Gutzwiller-projected variational state of minimum energy, the projective determinant QMC algorithm reaches accurate expectation values after far fewer steps and exhibits a markedly larger average sign in the Hubbard model and related systems.

What carries the argument

The Gutzwiller-projected Slater determinant obtained from variational Monte Carlo, used as the guiding trial state inside the projective determinant QMC projector.

If this is right

  • Shorter projection lengths suffice to reach a given statistical error, lowering overall CPU cost.
  • Parameter regions previously inaccessible because of severe sign cancellations become simulable.
  • Zero-temperature properties of strongly correlated fermions can be extracted with reduced resources.
  • The same trial-state construction can be reused across multiple observables in a single run.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The method may combine with other variational ansatzes that further suppress the sign problem.
  • It could enable studies of larger clusters or lower temperatures where standard projective QMC fails.
  • Similar projection ideas might transfer to auxiliary-field or diagrammatic Monte Carlo frameworks.

Load-bearing premise

The minimum-energy Gutzwiller-projected wavefunction remains sufficiently unbiased and sign-problem-friendly when inserted into the projective QMC evolution without introducing offsetting errors.

What would settle it

A side-by-side run on the doped Hubbard model comparing the number of steps to convergence and the average sign obtained with versus without the Gutzwiller trial state.

Figures

Figures reproduced from arXiv: 2303.13403 by Wei-Xuan Chang, Zi-Xiang Li.

Figure 1
Figure 1. Figure 1: FIG. 1. The contour plot of energy versus variational param [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. The results of simulation on spinful Hubbard model at half filling for Gutzwiller QMC and conventional PQMC with [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. The results of observable in spinless t-V model at [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. The results of sign problem in spinless t-V model at half filling for Gutzwiller QMC and conventional PQMC with [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
read the original abstract

Here we develop a new scheme of projective quantum Monte-Carlo (QMC) simulation combining unbiased zero-temperature (projective) determinant QMC and variational Monte-Carlo based on Gutzwiller projection wave function, dubbed as ``Gutzwiller projection QMC''. The numerical results demonstrate that employment of Gutzwiller projection trial wave function with minimum energy strongly speed up the convergence of computational results, thus tremendously reducing computational time in the simulation. More remarkably, we present an example that sign problem is enormously alleviated in the Gutzwiller projection QMC, especially in the regime where sign problem is severe. Hence, we believe that Gutzwiller projection QMC paves a new route to improving the efficiency, and alleviating sign problem in QMC simulation on interacting fermionic systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes 'Gutzwiller projection QMC', a hybrid scheme that inserts a variationally optimized Gutzwiller-projected trial wavefunction (obtained via VMC) as the initial state for projective determinant QMC. The central claims are that this choice accelerates imaginary-time convergence (thereby reducing required projection length and computational cost) and substantially alleviates the fermion sign problem relative to conventional Slater-determinant trials, with the largest gains reported in regimes where the sign problem is severe.

Significance. If the numerical demonstrations are robust and the improvements are shown to be general rather than model-specific, the method could meaningfully extend the reach of projective DQMC for interacting fermions by lowering the projection-time overhead and mitigating sign cancellations. The approach combines two established techniques in a straightforward way, but its value hinges on whether the variational energy minimum also optimizes the overlap and phase properties inside the auxiliary-field integral.

major comments (2)
  1. [numerical results and discussion sections] The central claim that the minimum-energy Gutzwiller trial simultaneously shortens the required projection length and improves the average sign rests on an unproven assumption that variational optimality for energy also optimizes the dynamical quantities relevant to the projector; no explicit test (e.g., comparison against another trial state with comparable variational energy but different Gutzwiller parameter) is provided to rule out model-specific coincidence or post-selection.
  2. [abstract and numerical results] Abstract and numerical demonstrations assert 'enormously alleviated' sign problem and 'strongly speed up' convergence, yet the provided text supplies no quantitative metrics (average sign values, projection lengths, system sizes, error bars, or direct comparisons to standard Slater trials) that would allow assessment of effect size or exclusion of fitting artifacts.
minor comments (2)
  1. [method section] Notation for the Gutzwiller projector and the precise form of the trial wavefunction should be defined explicitly with equations before the numerical section.
  2. [figures] Figure captions should include the precise Hamiltonian parameters, lattice sizes, and number of samples used for each data point.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading of our manuscript and the constructive comments. We address each major point below and indicate where revisions will be made.

read point-by-point responses

  1. Referee: [numerical results and discussion sections] The central claim that the minimum-energy Gutzwiller trial simultaneously shortens the required projection length and improves the average sign rests on an unproven assumption that variational optimality for energy also optimizes the dynamical quantities relevant to the projector; no explicit test (e.g., comparison against another trial state with comparable variational energy but different Gutzwiller parameter) is provided to rule out model-specific coincidence or post-selection.

    Authors: The referee is correct that we have not included an explicit comparison of the energy-minimized Gutzwiller parameter against a different parameter yielding comparable variational energy. Such a test would help exclude coincidence. Our choice follows the standard variational procedure, and the reported improvements appear consistently across the studied models and parameter regimes. We will add this comparison (using a non-optimal Gutzwiller factor with similar energy) to the numerical results section of the revised manuscript. revision: yes

  2. Referee: [abstract and numerical results] Abstract and numerical demonstrations assert 'enormously alleviated' sign problem and 'strongly speed up' convergence, yet the provided text supplies no quantitative metrics (average sign values, projection lengths, system sizes, error bars, or direct comparisons to standard Slater trials) that would allow assessment of effect size or exclusion of fitting artifacts.

    Authors: Quantitative metrics are shown in the figures of the numerical results section, which report average sign values, the projection lengths needed to reach convergence within statistical error bars, the system sizes, and direct comparisons against standard Slater-determinant trials. To make these values more immediately accessible from the text, we will insert explicit numerical examples and error estimates into the abstract and results sections of the revised manuscript. revision: yes

Circularity Check

0 steps flagged

No circularity: hybrid method reports empirical gains from independent variational optimization and projective QMC

full rationale

The paper describes a numerical combination of variational Monte Carlo (Gutzwiller projection to minimum energy) with projective determinant QMC. Reported speed-up in convergence and sign-problem relief are presented as outcomes of simulations on specific models, not as quantities that reduce by construction to the variational parameters or to a self-citation. The variational step selects the trial state; the projective step is an independent imaginary-time evolution whose convergence properties are measured separately. No load-bearing derivation equates a 'prediction' to its own input fit, and no uniqueness theorem or ansatz is smuggled via self-citation. The approach is therefore self-contained against external simulation benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract-only review yields no explicit free parameters, invented entities, or additional axioms beyond the standard domain assumption that a variational Gutzwiller state can serve as an effective trial wavefunction for projective QMC.

axioms (1)
  • domain assumption A minimum-energy Gutzwiller-projected wavefunction constitutes a useful trial state for projective determinant QMC
    The entire performance claim rests on this premise being true for the models studied.

pith-pipeline@v0.9.0 · 5661 in / 1310 out tokens · 22087 ms · 2026-05-24T09:56:43.099934+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

41 extracted references · 41 canonical work pages

  1. [1]

    We focus on the half filling of the model

    (4) where ci is the annihilation operator of fermion on sitei, t is the NN hopping amplitude, andV > 0 denotes the den- sity repulsive interaction between NN sites. We focus on the half filling of the model. The quantum phase diagram of the model at half filling has been extensively investi- gated in recent years, which features a quantum phase transition f...

    work page

  2. [2]

    D. J. Scalapino and R. L. Sugar, Phys. Rev. Lett. 46, 519 (1981)

    work page 1981

  3. [3]

    J. E. Hirsch, D. J. Scalapino, R. L. Sugar, and R. Blankenbecler, Phys. Rev. Lett. 47, 1628 (1981)

    work page 1981

  4. [4]

    S. R. White, D. J. Scalapino, R. L. Sugar, and N. E. Bickers, Phys. Rev. Lett. 63, 1523 (1989)

    work page 1989

  5. [5]

    Zhang, J

    S. Zhang, J. Carlson, and J. E. Gubernatis, Phys. Rev. Lett. 74, 3652 (1995)

    work page 1995

  6. [6]

    E. Gull, A. J. Millis, A. I. Lichtenstein, A. N. Rubtsov, M. Troyer, and P. Werner, Rev. Mod. Phys. 83, 349 (2011)

    work page 2011

  7. [7]

    F. F. Assaad and H. G. Evertz, Computational Many- Particle Physics , page 277-356, springer (2008)

    work page 2008

  8. [8]

    Troyer and U.-J

    M. Troyer and U.-J. Wiese, Phys. Rev. Lett. 94, 170201 (2005)

    work page 2005

  9. [9]

    Li and H

    Z.-X. Li and H. Yao, Annual Review of Condensed Mat- ter Physics 10, 337 (2019)

    work page 2019

  10. [10]

    E. Y. Loh, J. E. Gubernatis, R. T. Scalettar, S. R. White, D. J. Scalapino, and R. L. Sugar, Phys. Rev. B 41, 9301 (1990)

    work page 1990

  11. [11]

    Mondaini, S

    R. Mondaini, S. Tarat, and R. T. Scalettar, Science 375, 418 (2022)

    work page 2022

  12. [12]

    Li, Y.-F

    Z.-X. Li, Y.-F. Jiang, and H. Yao, Phys. Rev. B 91, 241117 (2015)

    work page 2015

  13. [13]

    Li, Y.-F

    Z.-X. Li, Y.-F. Jiang, and H. Yao, Phys. Rev. Lett. 117, 267002 (2016)

    work page 2016

  14. [14]

    C. J. Umrigar, J. Toulouse, C. Filippi, S. Sorella, and R. G. Hennig, Phys. Rev. Lett. 98, 110201 (2007)

    work page 2007

  15. [15]

    Hangleiter, I

    D. Hangleiter, I. Roth, D. Nagaj, and J. Eisert, Science Advances 6, eabb8341 (2020)

    work page 2020

  16. [16]

    Wan, S.-X

    Z.-Q. Wan, S.-X. Zhang, and H. Yao, arXiv preprint arXiv:2010.01141 (2020)

    work page arXiv 2010

  17. [17]

    Zhang, G

    X. Zhang, G. Pan, X. Y. Xu, and Z. Y. Meng, Phys. Rev. B 106, 035121 (2022)

    work page 2022

  18. [18]

    Ouyang and X

    Y. Ouyang and X. Y. Xu, Phys. Rev. B 104, L241104 (2021)

    work page 2021

  19. [19]

    Wang, Y.-H

    L. Wang, Y.-H. Liu, M. Iazzi, M. Troyer, and G. Harcos, Phys. Rev. Lett. 115, 250601 (2015)

    work page 2015

  20. [20]

    Z. C. Wei, C. Wu, Y. Li, S. Zhang, and T. Xiang, Phys. Rev. Lett. 116, 250601 (2016)

    work page 2016

  21. [21]

    Levy and B

    R. Levy and B. K. Clark, Phys. Rev. Lett. 126, 216401 (2021)

    work page 2021

  22. [22]

    Y.-Y. He, H. Shi, and S. Zhang, Phys. Rev. Lett. 123, 136402 (2019)

    work page 2019

  23. [23]

    Cohen-Stead, O

    B. Cohen-Stead, O. Bradley, C. Miles, G. Batrouni, R. Scalettar, and K. Barros, Phys. Rev. E 105, 065302 (2022)

    work page 2022

  24. [24]

    L. Wang, M. Iazzi, P. Corboz, and M. Troyer, Phys. Rev. B 91, 235151 (2015)

    work page 2015

  25. [25]

    E. Berg, S. Lederer, Y. Schattner, and S. Trebst, Annual Review of Condensed Matter Physics 10, 63 (2019)

    work page 2019

  26. [26]

    E. Berg, M. A. Metlitski, and S. Sachdev, Science 338, 1606 (2012)

    work page 2012

  27. [27]

    X. Y. Xu, Z. H. Liu, G. Pan, Y. Qi, K. Sun, and Z. Y. Meng, Journal of Physics: Condensed Matter 31, 463001 (2019)

    work page 2019

  28. [28]

    Fradkin, S

    E. Fradkin, S. A. Kivelson, and J. M. Tranquada, Rev. Mod. Phys. 87, 457 (2015)

    work page 2015

  29. [29]

    M. C. Gutzwiller, Phys. Rev. Lett. 10, 159 (1963)

    work page 1963

  30. [30]

    T. M. Rice and K. Ueda, Phys. Rev. B 34, 6420 (1986)

    work page 1986

  31. [31]

    D. S. Rokhsar and B. G. Kotliar, Phys. Rev. B 44, 10328 (1991)

    work page 1991

  32. [32]

    Jin, H.-H

    H.-K. Jin, H.-H. Tu, and Y. Zhou, Phys. Rev. B 104, L020409 (2021)

    work page 2021

  33. [33]

    D. P. Arovas, E. Berg, S. A. Kivelson, and S. Raghu, Annual Review of Condensed Matter Physics 13, 239 (2022), https://doi.org/10.1146/annurev-conmatphys- 031620-102024

    work page doi:10.1146/annurev-conmatphys- 2022

  34. [34]

    Otsuka, S

    Y. Otsuka, S. Yunoki, and S. Sorella, Phys. Rev. X 6, 011029 (2016)

    work page 2016

  35. [35]

    Janssen and I

    L. Janssen and I. F. Herbut, Phys. Rev. B 89, 205403 (2014)

    work page 2014

  36. [36]

    Ladovrechis, S

    K. Ladovrechis, S. Ray, T. Meng, and L. Janssen, Phys. Rev. B 107, 035151 (2023)

    work page 2023

  37. [37]

    N. Zerf, L. N. Mihaila, P. Marquard, I. F. Herbut, and M. M. Scherer, Phys. Rev. D 96, 096010 (2017)

    work page 2017

  38. [38]

    Parisen Toldin, M

    F. Parisen Toldin, M. Hohenadler, F. F. Assaad, and I. F. Herbut, Phys. Rev. B 91, 165108 (2015)

    work page 2015

  39. [39]

    Li, Y.-F

    Z.-X. Li, Y.-F. Jiang, and H. Yao, New Journal of Physics 17, 085003 (2015)

    work page 2015

  40. [40]

    L. Wang, P. Corboz, and M. Troyer, New Journal of Physics 16, 103008 (2014)

    work page 2014

  41. [41]

    Li, Z.-Q

    Z.-X. Li, Z.-Q. Wan, and H. Yao, arXiv preprint arXiv:2211.00663 (2022)

    work page arXiv 2022