Imaginary-time correlations in time-sliced stochastic series expansion

Adam Iaizzi; Anders W. Sandvik; Ryan Flynn; Sibin Yang; Ying-Jer Kao

REVIEW 2 major objections 4 minor 76 references

Imaginary-time correlations in time-sliced stochastic series expansion

T0 review · 2 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Time-sliced stochastic series expansion computes imaginary-time correlation functions of diagonal and off-diagonal operators at well-defined imaginary-time points, with no discretization error and only minimal overhead on top of existing…

desk verdict Solid SSE methods paper: the O∈H estimator is clean and benchmarked, but the O∉H importance-sampling estimator has an unproven technical assumption that deserves scrutiny before publication. read the letter →

arxiv 2608.13477 v1 pith:ZMMMNKUN submitted 2026-08-13 cond-mat.str-el

Ryan Flynn , Adam Iaizzi , Sibin Yang , Ying-Jer Kao , Anders W. Sandvik This is my paper · ORCID

classification cond-mat.str-el PACS 05.10.Ln75.40.Mg

keywords quantumMonteCarlostochasticseriesexpansionimaginary-timecorrelationfunctionsdirected-loopupdatestime-slicedSSEanalyticcontinuationtransverse-fieldIsingmodel

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Time-sliced stochastic series expansion (SSE) splits the quantum operator string into $m$ slices of width $\Delta\tau = \beta/m$, turning each slice boundary into a well-defined imaginary-time coordinate. On this grid the paper derives exact estimators for all three operator classes: diagonal correlators from propagated states at the boundaries; correlators of operators that appear in the Hamiltonian from counting the operators adjacent to a boundary, with estimator $n_l n_{l'}/\Delta\tau^2$; and correlators of operators not in the Hamiltonian from a histogram of slice-boundary crossings accumulated while directed loops propagate. The estimators require no change to the loop or cluster updates themselves and add only minimal computational overhead. Benchmarks on the one-dimensional transverse-field Ising chain and the XXZ chain agree with exact diagonalization within statistical error, and $L=128$ runs extract scaling dimensions and a Luttinger parameter close to known values. If correct, the method removes the long-standing restriction of SSE dynamics to diagonal structure factors or SU(2)-symmetric models, opening transverse correlators to analytic continuation.

What carries the argument

The central object is the sliced operator string: writing $Z = \mathrm{Tr}[(e^{-\Delta\tau H})^m]$ and expanding each of the $m$ slices independently converts every slice boundary into an exact imaginary-time point $\tau_k = k\Delta\tau$. Two identities carry the argument. An operator-absorption reindexing, applied within a single slice, turns an insertion of $O \in H$ at a boundary into a factor $n_l/\Delta\tau$, the number of non-identity operators in slice $l$ divided by the slice width, which yields the counting estimator $f(C) = n_l n_{l'}/\Delta\tau^2$ for a pair of boundary-adjacent operators. For $O \notin H$, the workhorse is the correspondence between directed-loop propagation and sampling of the extended two-defect configuration space $C'$: the loop's head and tail are the $S^+$ and $S^-$ insertions, and the correlator histogram is accumulated from the head's crossings of slice boundaries during the update, with the factor $n_x(l_0)\,n_{\mathrm{legs}}(C)$ compensating for the importance-sampling choice of the starting leg.

What would settle it

Run the importance-sampling $O \notin H$ estimator on a small XXZ chain in a field ($L=10$, $h=0.5$, as in Fig. 9) side by side with the boundary-start estimator and exact diagonalization, and compare the normalized $\tau$-resolved histograms at high statistics: a systematic deviation between the two estimators or from exact diagonalization would refute the claim. A more direct check is to verify that the standard directed-loop transition probabilities leave the extended-state weights $W(C')$ invariant; if they do not, the trial probabilities in the open-defect sector are wrong and the histogram is biased.

Watch

Extended reading notes

Core claim

The paper claims that an explicit discretization of the SSE operator string into $m$ independent slices, with the partition function written as $\mathrm{Tr}[(e^{-\Delta\tau H})^m]$, makes imaginary-time correlation functions measurable on the exact grid $\tau_k = k\Delta\tau$ with no discretization error. For operators $O \in H$, it derives a counting estimator: absorbing $O$ at a slice boundary reindexes that slice's expansion and leaves a factor $n_l/\Delta\tau$, where $n_l$ is the number of non-identity operators in slice $l$, so the two-point correlator at nonzero separation is $f(C) = n_l n_{l'}/\Delta\tau^2$ (Eq. 33), with a separate expression at equal time (Eq. 34). For operators $O \notin H$, it treats the directed loop as sampling an extended configuration space $C'$ in which $S^+$ and $S^-$ defects sit at slice boundaries, and measures the correlator as a histogram of the loop's boundary crossings, weighted by $n_x(l_0)\,n_{\mathrm{legs}}(C)$ in the importance-sampling variant to undo the bias of the starting-leg choice. Both estimator families are validated against exact diagonalization on small transverse-field Ising and XXZ systems and demonstrated at $L=128$, where the extracted exponents agree with expected values and the SU(2)-symmetric relation $2G_{zz} = G_{\pm}$ holds within error bars.

Load-bearing premise

The load-bearing premise is that, in the extended two-defect configurations sampled for operators outside the Hamiltonian, the standard directed-loop transition probabilities remain valid and loop closures of type (b) — where the loop rejoins its starting vertex through an exit leg — can be discarded from the measurement histogram without bias; the paper asserts this but gives no proof that the sampling weights stay correct.

Editorial extensions

If this is right

Transverse ($S^x$, $S^\pm$-type) imaginary-time correlators become accessible in SSE quantum Monte Carlo for models without SU(2) symmetry, such as anisotropic XXZ chains with external fields, lifting a restriction that previously limited SSE dynamics largely to diagonal structure factors.
The $\tau$-resolved correlators feed directly into numerical analytic continuation and thus yield real-frequency transverse spectral functions from the same simulation ensemble that produces static observables.
The overhead stays near the cost of the simulation itself: with the FFT the diagonal and $O \in H$ estimators scale as $O(mL^d \log mL^d)$, while the $O \notin H$ estimator scales as $O(mL^d)$, matching the SSE updates.
A time-slice averaged variant of the $O \in H$ estimator trades a controlled $O(\Delta\tau^2)$ bias (larger, $O(\Delta\tau)$, at equal time) for an order-of-magnitude variance reduction, which the benchmarks show is advantageous for extracting decay exponents.
The construction transfers to any sign-problem-free model with similar operator structure, including bosonic correlators such as $\langle c^\dagger(r,\tau)c(0,0)\rangle$.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The unbiasedness claim for the $O \notin H$ estimator rests on the type-(b) closure exclusion, which the paper asserts without proof; a direct high-statistics comparison of the importance-sampling and boundary-start estimators on a small XXZ system would settle this empirically, since each is currently validated only against exact diagonalization.
Because the boundary-valued and time-slice averaged $O \in H$ estimators differ by a known $O(\Delta\tau^2)$ term, their difference at fixed $\Delta\tau$ is a ready-made convergence diagnostic for deciding when a data set is clean enough to feed analytic continuation.
Read as a worm-type algorithm on the slice-boundary grid, the $O \notin H$ construction should carry over to sign-problem-free boson and fermion Hamiltonians, giving two-point Green's functions $\langle c^\dagger(r,\tau)c(0,0)\rangle$ without the single-hole specialization of recent $t$-$J$-type studies.
The histogram is unnormalized and its absolute scale is fixed by pinning a known value such as $G_{\pm}(0,0)=1/2$; applications where the equal-time value is not independently known would need a separate normalization scheme, which the paper leaves open.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The new thing here is a practical way to get imaginary-time correlators in SSE for off-diagonal operators, not just diagonal ones. The operator-counting estimator for O∈H (Sec. IV) is straightforward, well derived, and matches ED to statistical accuracy on the TFIM. The time-slice-averaged variant with its O(Δτ²) bias is a nice bias–variance tradeoff, and the variance reduction they show is real. That part of the paper is solid and likely to be widely used.

The O∉H estimator (Sec. V) is more delicate and is where I have the main reservation. The histogram-based directed-loop construction is clever, and the XXZ benchmarks against ED at the 0.1% level are genuinely impressive. But the unbiasedness rests on two claims that are asserted rather than proven: first, that type-(b) loop closures can be excluded with no bias, and second, that the compensating factors n_x(l0)n_legs(C) fully account for the configuration-dependent starting-leg selection. The first claim in particular looks like it could break detailed balance if the excluded closures are part of the standard directed-loop transition matrix. The authors should either provide a detailed-balance argument or demonstrate the equivalence more carefully. Small-chain ED agreement is good evidence, but it is not conclusive, especially for a methods paper.

Other soft spots are minor: no code or data is released, which hinders exact reproduction; the Luttinger parameter extraction is 4% off the Bethe value, which is fine but worth a comment; and the paper leans on prior time-slicing work for the diagonal case, as it should. The citation pattern is appropriate and the positioning against worm-type approaches is fair.

Overall, this is a useful contribution that removes a real bottleneck for SSE-based analytic continuation. It deserves a serious referee, but not a desk reject. The referee should ask for a proof or more extensive validation of the O∉H sampling assumption—perhaps tests on larger chains or in regimes with frequent bounces where a bias would be more visible. If that is resolved, the paper is publishable as is.

Referee Report

2 major / 4 minor

Summary. The paper develops a time-sliced variant of stochastic series expansion (SSE) in which the imaginary-time axis is divided into m slices of width Δτ, and constructs estimators for three classes of imaginary-time correlation functions: diagonal operators (Sec. III), off-diagonal operators that appear in the Hamiltonian, O∈H (Sec. IV), and off-diagonal operators not in the Hamiltonian, O∉H (Sec. V). For O∈H, the boundary-valued estimator counts operators adjacent to slice boundaries (Eq. 33) with no discretization error, and an alternative time-slice-averaged estimator (Eqs. 36-39) trades a derived O(Δτ²) bias (Eq. 43) for an order-of-magnitude variance reduction. For O∉H, the paper proposes sampling an extended configuration space C' during directed-loop updates and accumulating histogram contributions from loop crossings of slice boundaries (Sec. V.A, Eqs. 50-51). The method is benchmarked against exact diagonalization for the 1D TFIM and XXZ chain, with additional larger-scale results extracting scaling dimensions and a Luttinger parameter.

Significance. If the method is unbiased as claimed, it removes a long-standing restriction of SSE-based dynamical calculations by making transverse off-diagonal imaginary-time correlators accessible with only modest overhead beyond standard loop updates. The paper's strengths include explicit, parameter-free estimator formulas (the only free parameter is the slice width Δτ), a clean derivation of the O(Δτ²) bias for the time-slice-averaged O∈H estimator, careful ED benchmarks that agree to few percent or better, and large-scale demonstrations of efficiency. The main weakness is that the central claim for the O∉H estimator rests on an assertion about excluding type-(b) loop closures without a detailed-balance proof, and the O∈H algorithm as written contains a potential double-counting ambiguity. These issues are fixable but are load-bearing for the paper's advertised generality.

major comments (2)

[Sec. V.A (Fig. 5, Eq. (47))] The claim that loop closures of type (b) can be excluded entirely with no bias is asserted without proof. The directed-loop transition probabilities of Ref. [10] were derived for the closed configuration space C, whereas the estimator samples the extended space C' whose weights in Eq. (47) depend on the S+/S− insertion points. The compensation factors in Eqs. (51a)-(51b) account only for the probability of selecting the starting leg and boundary; they do not establish that the path probabilities in C' satisfy detailed balance or that continuing the loop until a type-(a) closure occurs preserves the stationary distribution. Because this assumption is load-bearing for the unbiasedness of G±(r,τ) for O∉H, please provide a derivation (or a rigorous argument) that the standard vertex probabilities, together with the modified closure rule, sample C' with the correct weights. Small-system ED benchmarks are not by themselves sufficient to rule out a systematic bias that could grow with system size or parameter range.
[Sec. IV (Eq. (33), pseudocode after Fig. 3)] The boundary-valued estimator is implemented with an 'if ... else if ...' rule that records at most one adjacent operator per time-slice boundary. However, if both the last non-identity operator of slice k−1 and the first non-identity operator of slice k are the same operator O(r), the insertion argument leading to Eq. (33) admits two distinct contributions (O inserted as H_{k−1,n_{k−1}} or as H_{k,1}), so the correct estimator should sum n_{k−1} and n_k rather than select one of them. The comment that this 'avoids double counting' appears to conflate two distinct insertion positions. Please clarify whether the implementation sums both cases; if the pseudocode is literal, the O∈H estimator is biased and the ED agreement in Fig. 6 would need to be re-examined.

minor comments (4)

[Sec. IV, Eqs. (25)-(28)] The derivation of ⟨H_k⟩ carries a minus sign from (−β)^n that is inconsistent with the convention H=−∑_b H_b with positive H_b adopted in Eq. (4). Equation (29) and the later statement that H_k=hσ^x (requiring division by h²) indicate the positive convention; please make the signs consistent throughout.
[Sec. V.A, Eq. (51)] The statement that n_legs(C) is proportional to n_H, followed by a suggestion to normalize by 4mM as 'the total number of legs including identities', is confusing; please define n_legs(C) unambiguously and clarify whether the directed-loop graph includes identity vertices.
[Fig. 10 (left)] The extracted Luttinger parameter K=0.722 is 4% below the Bethe-ansatz value 3/4; please report the fit range, the statistical error of K, and whether the deviation is expected from finite-size or finite-Δτ effects.
[Abstract and Sec. I] The abstract's statement that the method provides correlation functions 'with no discretization error' is qualified for the time-slice-averaged O∈H estimator in Sec. IV.A, which has an O(Δτ²) bias; consider adding a sentence noting that the no-discretization-error property applies to the boundary-valued estimators.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the estimators are derived from the SSE operator string and validated against external exact-diagonalization and Bethe-ansatz benchmarks; the Sec. V.A type-(b) closure exclusion is a correctness gap, not a circular step.

full rationale

The O∈H estimator (Eq. 33) is derived by operator absorption/reindexing within the time-sliced SSE partition function, and the diagonal estimator (Eq. 20) is a direct eigenvalue product; neither fits any parameter to the target correlation functions. The O∉H estimator (Eqs. 46-49) samples an extended configuration space whose weights are the correlation-function numerator, which is the standard worm/directed-loop construction explicitly credited to Refs. [10,61,62]; this is a sampling identity, not a self-referential derivation. The overall normalization is pinned by the fixed spin identity G±(0,0)=1/2, not by a fit, so the τ-dependent shape remains an independent prediction. Central results are checked externally: exact diagonalization for L=10 TFIM and XXZ chains (Figs. 6,8,9), known scaling dimensions Δσ=1/8 and Δε=1 (Fig. 7), Bethe-ansatz Luttinger parameter K=3/4 (Fig. 10), and SU(2) consistency 2Gzz=G± at Δ=1. Self-citations to prior time-slicing and directed-loop work (Refs. [9,10,16,60,67]) are load-bearing algorithmic ingredients but are standard, externally used methods, not uniqueness imports. I flag one non-circular correctness concern: Sec. V.A (near Fig. 5) states, 'These closures can also be excluded entirely with no bias introduced (simply continue to extend the loop until a closure of type (a) is encountered),' without a detailed-balance proof in the extended space C'; if this assertion is wrong, the histogram estimator would be biased. That is an omitted proof, but it is not circularity, because it does not reduce the prediction to a fitted input or to the paper's own assumptions.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The method introduces no new physical entities. The central estimators rest on standard SSE machinery plus two unproven technical assumptions specific to the O∉H importance-sampling estimator (extended-space detailed balance and the type-(b) closure exclusion). Δτ is a user-chosen resolution parameter that governs a controlled bias in the averaged estimator.

free parameters (1)

Time-slice width Δτ
User-chosen discretization of imaginary time. The boundary-valued estimators are exact for any Δτ (no discretization error); the time-slice-averaged estimator has a bias O(Δτ^2) and O(Δτ) at τ=0, so Δτ controls the bias-variance tradeoff. Not fitted to data, but a hand-chosen parameter affecting reported accuracy.

assumptions (4)

domain assumption The standard SSE directed-loop transition probabilities satisfy detailed balance in the physical configuration space C and remain valid for sampling the extended configuration space C' with inserted S± defects.
Invoked in Sec. V when defining the O∉H estimator; no derivation is given for the extended-space detailed balance, which is critical to the unbiasedness of the estimator.
ad hoc to paper Excluding loop closures of type (b) (exit-leg reconnections) from the histogram accumulation introduces no bias in the correlation function.
Stated in Sec. V.A (near Fig. 5) without proof; this is a specific claim of this paper and not a standard background result.
standard math The time-sliced partition function (Eq. 10) with cutoff M is equivalent to the non-sliced SSE representation with no approximation when M is large enough.
Standard SSE truncation argument, cited to Refs. [9,10]; the slicing does not change the distribution.
standard math For the time-slice-averaged estimator, the insertion times Δs_i are uniformly distributed and independent, so the average separation is exact while the variance contributes O(Δτ^2) bias.
Used in the bias estimate Eq. (43); assumes G has continuous second derivative on the relevant interval.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Imaginary-time correlations in time-sliced stochastic series expansion." pith.science (2026). https://pith.science/paper/ZMMMNKUN

@misc{pith2026260813477,
  author       = {Pith},
  title        = {Pith review of: Imaginary-time correlations in time-sliced stochastic series expansion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZMMMNKUN}},
  note         = {Machine review of arXiv:2608.13477}
}

read the original abstract

Combined with numerical analytic continuation techniques, quantum Monte Carlo (QMC) methods enable the extraction of real-frequency dynamical properties from imaginary-time correlation functions. However, the efficient computation of imaginary-time correlation functions by QMC simulations can (depending on the particular model used) be challenging, particularly for operators that are off-diagonal in the computational basis. In this work, we present an efficient and general algorithm within the stochastic series expansion (SSE) framework for evaluating imaginary-time correlation functions of both diagonal and off-diagonal operators. The algorithm builds on a discrete imaginary-time slicing of the SSE operator string, which provides correlation functions on a grid of well-defined imaginary-time points with no discretization error. For off-diagonal operators, we derive estimators that integrate directly into the existing SSE directed-loop or cluster updating schemes, introducing only minimal computational overhead. We benchmark the method on the one-dimensional transverse-field Ising model (sampling with cluster updates) and XXZ spin chain (using directed-loop sampling), demonstrating excellent agreement (with only statistical errors) with exact diagonalization of small systems. We also study larger systems to demonstrate efficiency.

Figures

Figures reproduced from arXiv: 2608.13477 by the authors.

**Figure 3.** FIG. 3. Illustration of the off-diagonal estimator for the cor [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

**Figure 1.** The propagation of these defects (a “head” and [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗

Figures from the paper (6 more)

**Figure 5.** Figure 5: FIG. 5. In the directed loops algorithm [10], there are two [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]

**Figure 6.** Figure 6: FIG. 6. A comparison of the (left) same-site diagonal correlation function [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]

**Figure 7.** Figure 7: FIG. 7. Same-site imaginary-time correlation functions for the [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]

**Figure 8.** Figure 8: FIG. 8. A comparison of same-site correlation functions [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]

**Figure 9.** Figure 9: FIG. 9. Imaginary-time same-site correlation functions (left) [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]

**Figure 10.** Figure 10: FIG. 10. Same-site imaginary-time correlation functions for the [PITH_FULL_IMAGE:figures/full_fig_p016_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

76 extracted references · 57 canonical work pages

[10]

O. F. Sylju ˚ asen and A. W. Sandvik, Quantum Monte Carlo with directed loops, Physical Review E66, 046701 (2002)

work page 2002
[1]

A. W. Sandvik, Computational Studies of Quantum Spin Systems, inAIP Conference Proceedings, Vol. 1297 (American Institute of Physics, 2010) pp. 135–338

work page 2010
[2]

Gubernatis, N

J. Gubernatis, N. Kawashima, and P. Werner,Quan- tum Monte Carlo Methods: Algorithms for Lattice Models (Cambridge University Press, 2016)

work page 2016
[3]

A. W. Sandvik and J. Kurkij¨ arvi, Quantum Monte Carlo simulation method for spin systems, Phys. Rev. B43, 5950 (1991)

1991
[4]

Sandvik, A generalization of Handscomb’s quantum Monte Carlo scheme — Application to the 1-D Hubbard model, Journal of Physics A: Mathematical and General25, 3667 (1992)

Anders W. Sandvik, A generalization of Handscomb’s quantum Monte Carlo scheme — Application to the 1-D Hubbard model, Journal of Physics A: Mathematical and General25, 3667 (1992)

work page 1992
[5]

A. W. Sandvik, Finite-size scaling of the ground-state pa- rameters of the two-dimensional Heisenberg model, Phys. Rev. B56, 11678 (1997)

1997
[6]

E. Y. Loh, J. E. Gubernatis, R. T. Scalettar, S. R. White, D. J. Scalapino, and R. L. Sugar, Sign problem in the nu- merical simulation of many-electron systems, Phys. Rev. B41, 9301 (1990)

1990
[7]

Henelius and A

P. Henelius and A. W. Sandvik, Sign problem in Monte Carlo simulations of frustrated quantum spin systems, Phys. Rev. B62, 1102 (2000)

2000

Show all 76 references

[8]

Troyer and U.-J

M. Troyer and U.-J. Wiese, Computational Complex- ity and Fundamental Limitations to Fermionic Quantum Monte Carlo Simulations, Phys. Rev. Lett.94, 170201 (2005)

2005
[9]

A. W. Sandvik, Stochastic series expansion method with operator-loop update, Physical Review B59, R14157 (1999)

1999
[11]

O. F. Sylju ˚ asen, Directed loop updates for quantum lat- tice models, Phys. Rev. E67, 046701 (2003)

2003
[12]

R. G. Melko and R. K. Kaul, Scaling in the Fan of an Un- conventional Quantum Critical Point, Phys. Rev. Lett. 100, 017203 (2008)

2008
[13]

Desai and S

N. Desai and S. Pujari, Resummation-based quantum Monte Carlo for quantum paramagnetic phases, Phys. Rev. B104, L060406 (2021)

2021
[14]

Takahashi, S

J. Takahashi, S. Slezak, and E. Crosson, Rapidly mixing loop representation quantum Monte Carlo for Heisenberg models on star-like bipartite graphs (2024), arXiv:2411.01452 [quant-ph]

2024 arXiv
[15]

Dao, Y.-C

L. Dao, Y.-C. Wang, and H. Shao, Deterministic Loop Stochastic Series Expansion Algorithm for Quantum Spin Models in Magnetic Fields (2026), arXiv:2604.04635 [cond-mat.str-el]

2026 arXiv
[16]

A. W. Sandvik, Stochastic series expansion method for quantum Ising models with arbitrary interactions, Phys. Rev. E68, 056701 (2003)

2003
[17]

A. W. Sandvik, S. Daul, R. R. P. Singh, and D. J. Scalapino, Striped Phase in a QuantumXYModel with Ring Exchange, Phys. Rev. Lett.89, 247201 (2002)

2002
[18]

R. G. Melko and A. W. Sandvik, Stochastic series expan- sion algorithm for theS= 1/2XYmodel with four-site ring exchange, Phys. Rev. E72, 026702 (2005)

2005
[19]

Biswas, G

S. Biswas, G. Rakala, and K. Damle, Quantum cluster algorithm for frustrated Ising models in a transverse field, Phys. Rev. B93, 235103 (2016)

2016
[20]

Z. Yan, Y. Wu, C. Liu, O. F. Sylju ˚ asen, J. Lou, and Y. Chen, Sweeping cluster algorithm for quantum spin systems with strong geometric restrictions, Phys. Rev. B 99, 165135 (2019)

2019
[21]

Patil, Quantum Monte Carlo simulations in the re- stricted Hilbert space of Rydberg atom arrays, SciPost Phys.20, 022 (2026)

P. Patil, Quantum Monte Carlo simulations in the re- stricted Hilbert space of Rydberg atom arrays, SciPost Phys.20, 022 (2026)

2026
[22]

R. H. Swendsen and J.-S. Wang, Nonuniversal critical dynamics in Monte Carlo simulations, Phys. Rev. Lett. 58, 86 (1987)

1987
[23]

Luijten and H

E. Luijten and H. W. Bl¨ ote, Monte carlo method for spin models with long-range interactions, International Jour- nal of Modern Physics C (IJMPC)6, 359 (1995)

1995
[24]

H. G. Evertz, G. Lana, and M. Marcu, Cluster algorithm for vertex models, Phys. Rev. Lett.70, 875 (1993)

1993
[25]

H. G. Evertz, The loop algorithm, Advances in Physics52, 1 (2003), https://doi.org/10.1080/0001873021000049195

2003 doi
[26]

A. W. Sandvik, Evidence for Deconfined Quantum Crit- icality in a Two-Dimensional Heisenberg Model with Four-Spin Interactions, Phys. Rev. Lett.98, 227202 (2007)

2007
[27]

J. Lou, A. W. Sandvik, and N. Kawashima, Anti- ferromagnetic to valence-bond-solid transitions in two- dimensional SU(N) Heisenberg models with multispin in- teractions, Phys. Rev. B80, 180414(R) (2009)

2009
[28]

R. K. Kaul, Marshall-positive SU(N) quantum spin sys- tems and classical loop models: A practical strategy to design sign-problem-free spin Hamiltonians, Phys. Rev. B91, 054413 (2015)

2015
[29]

Takahashi and A

J. Takahashi and A. W. Sandvik, Valence-bond solids, vestigial order, and emergent SO(5) symmetry in a two- dimensional quantum magnet, Phys. Rev. Res.2, 033459 (2020)

2020
[30]

Takahashi, H

J. Takahashi, H. Shao, B. Zhao, W. Guo, and A. W. Sandvik, SO(5) multicriticality in two-dimensional quan- tum magnets (2024), arXiv:2405.06607 [cond-mat.str-el]

2024 arXiv
[31]

Kundu, N

S. Kundu, N. Desai, and K. Damle, Competition between Neel, Haldane nematic, plaquette valence bond solid, and (π, π) valence bond solid phases in SU(N) analogs of S= 1 square-lattice antiferromagnets, Phys. Rev. B109, 195146 (2024)

2024
[32]

Sengupta, L

P. Sengupta, L. P. Pryadko, F. Alet, M. Troyer, and G. Schmid, Supersolids versus Phase Separation in Two- Dimensional Lattice Bosons, Phys. Rev. Lett.94, 207202 (2005)

2005
[33]

Schaffer, A

R. Schaffer, A. A. Burkov, and R. G. Melko, Superfluid phases of lattice bosons with ring-exchange interaction, Phys. Rev. B80, 014503 (2009)

2009
[34]

Pippan, H

P. Pippan, H. G. Evertz, and M. Hohenadler, Excitation spectra of strongly correlated lattice bosons and polari- tons, Phys. Rev. A80, 033612 (2009)

2009
[35]

Y. Wang, W. Guo, and A. W. Sandvik, Anomalous Quan- 18 tum Glass of Bosons in a Random Potential in Two Di- mensions, Phys. Rev. Lett.114, 105303 (2015)

2015
[36]

Merali, I

E. Merali, I. J. S. De Vlugt, and R. G. Melko, Stochas- tic series expansion quantum Monte Carlo for Rydberg arrays, SciPost Phys. Core7, 016 (2024)

2024
[37]

Weber, Valence bond order in a honeycomb antiferro- magnet coupled to quantum phonons, Phys

M. Weber, Valence bond order in a honeycomb antiferro- magnet coupled to quantum phonons, Phys. Rev. B103, L041105 (2021)

2021
[38]

G¨ otz, F

A. G¨ otz, F. F. Assaad, and N. C. Costa, Tuning the order of a deconfined quantum critical point (2024), arXiv:2412.17215 [cond-mat.str-el]

2024 arXiv
[39]

Wang, Y.-H

L. Wang, Y.-H. Liu, and M. Troyer, Stochastic series ex- pansion simulation of thet−vmodel, Phys. Rev. B93, 155117 (2016)

2016
[40]

Li and H

Z.-X. Li and H. Yao, Sign-Problem-Free Fermionic Quan- tum Monte Carlo: Developments and Applications, An- nual Review of Condensed Matter Physics10, 337 (2019)

2019
[41]

F. Alet, K. Damle, and S. Pujari, Sign-Problem-Free Monte Carlo Simulation of Certain Frustrated Quantum Magnets, Phys. Rev. Lett.117, 197203 (2016)

2016
[42]

Wessel, I

S. Wessel, I. Niesen, J. Stapmanns, B. Normand, F. Mila, P. Corboz, and A. Honecker, Thermodynamic properties of the Shastry-Sutherland model from quantum Monte Carlo simulations, Phys. Rev. B98, 174432 (2018)

2018
[43]

T. Hong, M. Matsumoto, Y. Qiu, W. Chen, T. R. Gen- tile, S. Watson, F. F. Awwadi, M. M. Turnbull, S. E. Dissanayake, H. Agrawal, R. Toft-Petersen, B. Klemke, K. Coester, K. P. Schmidt, and D. A. Tennant, Higgs amplitude mode in a two-dimensional quantum antiferro- magnet near ...

2017
[44]

A. Jain, M. Krautloher, J. Porras, G. H. Ryu, D. P. Chen, D. L. Abernathy, J. T. Park, A. Ivanov, J. Chaloupka, G. Khaliullin, B. Keimer, and B. J. Kim, Higgs mode and its decay in a two-dimensional antiferromagnet, Nature Physics13, 633 (2017)

2017
[45]

Gazit, D

S. Gazit, D. Podolsky, A. Auerbach, and D. P. Arovas, Dynamics and conductivity near quantum criticality, Phys. Rev. B88, 235108 (2013)

2013
[46]

Witczak-Krempa, E

W. Witczak-Krempa, E. S. Sørensen, and S. Sachdev, The dynamics of quantum criticality revealed by quan- tum Monte Carlo and holography, Nat. Phys.10, 361 (2014)

2014
[47]

E. Katz, S. Sachdev, E. S. Sørensen, and W. Witczak- Krempa, Conformal field theories at nonzero tempera- ture: Operator product expansions, Monte Carlo, and holography, Phys. Rev. B90, 245109 (2014)

2014
[48]

Jarrell and J

M. Jarrell and J. E. Gubernatis, Bayesian inference and the analytic continuation of imaginary-time quantum Monte Carlo data, Physics Reports269, 133 (1996)

1996
[49]

R. N. Silver, D. S. Sivia, and J. E. Gubernatis, Maximum- entropy method for analytic continuation of quantum Monte Carlo data, Phys. Rev. B41, 2380 (1990)

1990
[50]

Jarrell, The maximum entropy method: Analytic con- tinuation of qmc data, inCorrelated Electrons: From Models to Materials Modeling and Simulation, Vol

M. Jarrell, The maximum entropy method: Analytic con- tinuation of qmc data, inCorrelated Electrons: From Models to Materials Modeling and Simulation, Vol. 2, edited by E. Pavarini, E. Koch, F. Anders, and M. Jarrell (Verlag des Forschungszentrum J¨ ulich, J¨ ulich, Germany, 2012)

2012
[51]

A. W. Sandvik, Stochastic method for analytic continu- ation of quantum Monte Carlo data, Physical Review B 57, 10287 (1998)

1998
[52]

K. S. D. Beach, Identifying the maximum entropy method as a special limit of stochastic analytic continu- ation, arXiv preprint cond-mat/0403055 (2004)

2004 arXiv
[53]

A. W. Sandvik, Constrained sampling method for ana- lytic continuation, Physical Review E94, 063308 (2016)

2016
[54]

H. Shao, H. Guo, and A. W. Sandvik, Almost linear de- pendence of the optimized stochastic analytic continua- tion on the imaginary-time data, Physical Review B95, 134425 (2017)

2017
[55]

Loh¨ ofer, T

M. Loh¨ ofer, T. Coletta, D. G. Joshi, F. F. Assaad, M. Vo- jta, S. Wessel, and F. Mila, Dynamical structure factors and excitation modes of the bilayer Heisenberg model, Phys. Rev. B92, 245137 (2015)

2015
[56]

H. Shao, Y. Q. Qin, S. Capponi, S. Chesi, Z. Y. Meng, and A. W. Sandvik, Nearly Deconfined Spinon Excita- tions in the Square-Lattice Spin-1/2 Heisenberg Antifer- romagnet, Phys. Rev. X7, 041072 (2017)

2017
[57]

Y. Q. Qin, B. Normand, A. W. Sandvik, and Z. Y. Meng, Amplitude Mode in Three-Dimensional Dimerized Anti- ferromagnets, Phys. Rev. Lett.118, 147207 (2017)

2017
[58]

Ma, G.-Y

N. Ma, G.-Y. Sun, Y.-Z. You, C. Xu, A. Vishwanath, A. W. Sandvik, and Z. Y. Meng, Dynamical signature of fractionalization at a deconfined quantum critical point, Phys. Rev. B98, 174421 (2018)

2018
[59]

S. Yang, G. Schumm, B. Zhao, and A. W. Sandvik, Single-hole spectral functions in 1D quantum magnets with different ground states (2025), arXiv:2511.20447 [cond-mat.str-el]

2025 arXiv
[60]

Sandvik, Stochastic series expansion meth- ods, inMany-Body Methods for Real Materials, Vol

Anders W. Sandvik, Stochastic series expansion meth- ods, inMany-Body Methods for Real Materials, Vol. 9, edited by E. Pavarini, E. Koch, and S. Zhang (Verlag des Forschungszentrum J¨ ulich, J¨ ulich, Germany, 2019)

2019
[61]

N. V. Prokof’ev, B. V. Svistunov, and I. S. Tupitsyn, Exact quantum Monte Carlo process for the statistics of discrete systems, JETP Lett.64, 911 (1996)

1996
[62]

Dorneich and M

A. Dorneich and M. Troyer, Accessing the dynamics of large quantum systems using the stochastic series expan- sion, Physical Review E64, 066701 (2001)

2001
[63]

Laflorencie, H

N. Laflorencie, H. Rieger, A. W. Sandvik, and P. Henelius, Crossover effects in the random-exchange spin- 1 2 antiferromagnetic chain, Phys. Rev. B70, 054430 (2004)

2004
[64]

O. F. Sylju ˚ asen, Numerical evidence for unstable magnons at high fields in the Heisenberg antiferromag- net on the square lattice, Phys. Rev. B78, 180413(R) (2008)

2008
[65]

Grossjohann and W

S. Grossjohann and W. Brenig, Spin dynamics of the antiferromagnetic spin- 1 2 chain at finite magnetic fields and intermediate temperatures, Phys. Rev. B79, 094409 (2009)

2009
[66]

Rahnavard and W

Y. Rahnavard and W. Brenig, Spin dynamics of the anisotropic spin-1 antiferromagnetic chain at finite mag- netic fields, Phys. Rev. B91, 054405 (2015)

2015
[67]

H. Shao, W. Guo, and A. W. Sandvik, Emergent topolog- ical excitations in a two-dimensional quantum spin sys- tem, Physical Review B91, 094426 (2015)

2015
[68]

N. V. Prokof’ev, B. V. Svistunov, and I. S. Tupitsyn, Exact, complete, and universal continuous-time world- line Monte Carlo approach to the statistics of discrete quantum systems, J. Exp. Theor. Phys.87, 310 (1998)

1998
[69]

Prokof’ev, B

N. Prokof’ev, B. Svistunov, and I. Tupitsyn, “Worm” algorithm in quantum Monte Carlo simulations, Physics Letters A238, 253 (1998)

1998
[70]

A. S. Mishchenko, N. V. Prokof’ev, and B. V. Svistunov, Single-hole spectral function and spin-charge separation in thet-Jmodel, Phys. Rev. B64, 033101 (2001). 19

2001
[71]

Yang and A

S. Yang and A. W. Sandvik, Spinons and Spin-Charge Separation at the Deconfined Quantum Critical Point (2025), arXiv:2512.02962 [cond-mat.str-el]

2025
[72]

Louis and C

K. Louis and C. Gros, Stochastic cluster series expansion for quantum spin systems, Phys. Rev. B70, 100410(R) (2004)

2004
[73]

S. Yang, G. Schumm, and A. W. Sandvik, Dynamic struc- ture factor of a spin-1/2 Heisenberg chain with long-range interactions, Phys. Rev. B111, 224404 (2025)

2025
[74]

Giamarchi,Quantum Physics in One Dimension(Ox- ford University Press, 2003)

T. Giamarchi,Quantum Physics in One Dimension(Ox- ford University Press, 2003)

2003
[75]

Z. Wang, Z. Liu, B. B. Mao, Y. Li, and S. Zhang, Ad- dressing general measurements in quantum Monte Carlo, Nature Communications17, 679 (2026)

2026
[76]

Z. Wang, Z. Wang, B.-B. Mao, and Z. Yan, Generalized Reduced-Density-Matrix Quantum Monte Carlo Gives Access to More (2026), arXiv:2603.10948 [cond-mat.str- el]

2026

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [10]

O. F. Sylju ˚ asen and A. W. Sandvik, Quantum Monte Carlo with directed loops, Physical Review E66, 046701 (2002)

work page 2002

[2] [1]

A. W. Sandvik, Computational Studies of Quantum Spin Systems, inAIP Conference Proceedings, Vol. 1297 (American Institute of Physics, 2010) pp. 135–338

work page 2010

[3] [2]

Gubernatis, N

J. Gubernatis, N. Kawashima, and P. Werner,Quan- tum Monte Carlo Methods: Algorithms for Lattice Models (Cambridge University Press, 2016)

work page 2016

[4] [3]

A. W. Sandvik and J. Kurkij¨ arvi, Quantum Monte Carlo simulation method for spin systems, Phys. Rev. B43, 5950 (1991)

1991

[5] [4]

Sandvik, A generalization of Handscomb’s quantum Monte Carlo scheme — Application to the 1-D Hubbard model, Journal of Physics A: Mathematical and General25, 3667 (1992)

Anders W. Sandvik, A generalization of Handscomb’s quantum Monte Carlo scheme — Application to the 1-D Hubbard model, Journal of Physics A: Mathematical and General25, 3667 (1992)

work page 1992

[6] [5]

A. W. Sandvik, Finite-size scaling of the ground-state pa- rameters of the two-dimensional Heisenberg model, Phys. Rev. B56, 11678 (1997)

1997

[7] [6]

E. Y. Loh, J. E. Gubernatis, R. T. Scalettar, S. R. White, D. J. Scalapino, and R. L. Sugar, Sign problem in the nu- merical simulation of many-electron systems, Phys. Rev. B41, 9301 (1990)

1990

[8] [7]

Henelius and A

P. Henelius and A. W. Sandvik, Sign problem in Monte Carlo simulations of frustrated quantum spin systems, Phys. Rev. B62, 1102 (2000)

2000