Pith. sign in

REVIEW 3 major objections 4 minor 41 references

One-Shot Simulation of Static Disorder in Quantum Dynamics with Equilibrium Initial State via Matrix Product State Sampling

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper shows that static-disorder-averaged equilibrium correlation functions can be computed from a single matrix product state simulation, with benchmarks on the Holstein model matching direct sampling.

desk verdict Genuinely useful extension of auxiliary-DoF static-disorder averaging to thermal equilibrium initial states, with solid benchmarks and one addressable gap in per-branch truncation error control. read the letter →

arxiv 2506.07120 v1 pith:Q55C6VFC submitted 2025-06-08 physics.chem-ph

Zhao Zhang , Jiajun Ren , Wei-Hai Fang This is my paper
classification physics.chem-ph
keywords staticdisordermatrixproductstateauxiliarydegreesoffreedomthermalequilibriumtimecorrelationfunctionHolsteinmodeldiscretevariablerepresentationtime-dependentDMRG
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

Static disorder in molecular aggregates is normally handled by running thousands of independent quantum dynamics simulations, one per disorder sample. This paper claims that the same disorder-averaged quantity can be obtained from a single one-shot simulation if the disorder variables are built into a matrix product state as auxiliary coordinates. The key formula, Eq. (25), expresses the disorder-averaged equilibrium time correlation function as a sample average of ratios of overlaps between three states extracted from that one MPS. Benchmarks on the Holstein model's dipole-dipole correlation function at 300 K match the conventional direct-sampling result while cutting wall time by roughly two orders of magnitude and allowing more than a million samples. If the construction holds, the reusable ingredient is the ratio-sampling identity, not any particular Hamiltonian.

What carries the argument

The carrying object is the auxiliary-coordinate MPS $|\Psi(t)\rangle = \int ds \sqrt{\rho(s)}\,|\psi(s,t)\rangle |s\rangle$, where $s$ are static disorder parameters represented in coordinate space and $\rho(s)$ is their probability distribution. The new mechanism is the three-state construction of Eqs. (22)-(24) followed by the per-sample ratio estimator of Eq. (25), which handles disorder-dependent thermal initial states by reweighting each extracted branch by its own squared norm. Supporting machinery: a sine discrete variable representation of the auxiliary coordinates, which gives a predictable recurrence time $T \approx 2\pi/\Delta s$; and caching of tensor contractions that are shared across samples, so the sampling overhead stays small.

What would settle it

One decisive test: for an exactly solvable disorder-dependent thermal model, compare the MPS-sampling estimator of Eq. (25) with the exact branch-weighted average while tracking the per-branch norms $\langle \Psi_\beta(s_k)|\Psi_\beta(s_k)\rangle$ against the analytic partition function $Z(s_k)$. If the branch norms drift from $Z(s_k)$ at fixed bond dimension even though the global MPS appears converged, the per-branch ratio estimate is biased and the one-shot claim fails in that regime.

Watch

Extended reading notes

Core claim

The paper establishes a concrete identity for static-disorder-averaged equilibrium correlation functions. Starting from the purified infinite-temperature state $|\psi_{\beta=0}\rangle$, it builds one MPS over physical and auxiliary disorder degrees of freedom: $|\Psi_\beta\rangle = e^{-\beta \hat H(\hat s)/2}|\psi_{\beta=0}\rangle|\chi\rangle$, then $|\Psi_L\rangle = e^{-i\hat H(\hat s)t}|\Psi_\beta\rangle$, and $|\Psi_R\rangle = e^{-i\hat H(\hat s)t}\hat B(\hat s)|\Psi_\beta\rangle$. Sampling disorder configurations $s_k$ and restricting each state to that configuration gives Eq. (25), $\langle \hat A(t)\hat B\rangle = \frac{1}{N_\mathrm{samp}}\sum_{s_k} \frac{\langle \Psi_L(s_k)|\hat A(s_k)|\Psi_R(s_k)\rangle}{\langle \Psi_\beta(s_k)|\Psi_\beta(s_k)\rangle}$. The denominator is the disorder-dependent partition function for that sample, so the ratio is the correct disorder-specific contribution; the MPS compression is what makes all these branches available at once. The paper validates the identity on the Holstein model under Gaussian site-energy disorder, for several disorder strengths and system sizes, by comparison with direct sampling.

Load-bearing premise

The construction rests on the assumption that the single compressed MPS, after global truncation, represents every disorder branch accurately enough that none of the per-branch ratios in Eq. (25) is biased; the paper demonstrates $M$-convergence for one Holstein benchmark but proves no error bound linking global truncation error to per-branch bias.

Editorial extensions

If this is right

  • One simulation replaces thousands of independent realizations, with a speedup over direct sampling of roughly two orders of magnitude for the tested systems.
  • The bond dimension needed by the one-shot MPS stays close to that of a single realization for moderate disorder, rising to $M=64$ when the disorder strength approaches the electronic bandwidth.
  • Sample sizes can be pushed beyond $10^6$ with modest added cost, reducing statistical noise.
  • The sine DVR grid gives a predictable recurrence time $T \approx 2\pi/\Delta s$, letting a user choose the grid size for the desired time window.
  • The method is presented as a general tool for equilibrium time correlation functions in system-bath models with static disorder, beyond the Holstein example.

Reading between the lines

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

  • Beyond the paper: the same three-state ratio could be reused for nonlinear response functions, since only forward and backward branches are needed; the paper itself demonstrates only the two-time dipole correlation function.
  • Beyond the paper: tracking per-branch norms $\langle \Psi_\beta(s_k)|\Psi_\beta(s_k)\rangle$ against analytic partition functions in exactly solvable limits would turn the unproved truncation assumption into a testable diagnostic.
  • Beyond the paper: because all samples are drawn from one compressed MPS, sample fluctuations are not statistically independent; whether the effective noise decreases faster than $1/\sqrt{N_\mathrm{samp}}$ or requires correlated-error analysis is left open.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. This manuscript extends the auxiliary-degree-of-freedom (aux-DoF) matrix product state (MPS) method to treat static disorder in quantum dynamics when the initial state is a disorder-dependent thermal equilibrium state. The authors introduce an auxiliary wavefunction |χ⟩ = ∫ ds √ρ(s)|s⟩, define three MPSs — a thermal state |Ψβ⟩, a forward-evolved state |ΨL⟩, and a backward-propagated |ΨR⟩ — and then extract, for each sampled disorder configuration s_k, conditional states from the single global MPS. The static-disorder-averaged two-time correlation function is evaluated as the sample average of the ratio ⟨ΨL(s_k)|Â(s_k)|ΨR(s_k)⟩ / ⟨Ψβ(s_k)|Ψβ(s_k)⟩ [Eq. (25)]. The method is benchmarked on the dipole-dipole time correlation function of the Holstein model at 300 K for two, four, eight, and sixteen sites, with comparisons to direct sampling, studies of DVR basis convergence and recurrence artifacts, bond-dimension convergence, and wall-time scaling. The central claim is that the method computes disorder-averaged equilibrium correlation functions using one time evolution, with only a moderate increase in MPS bond dimension and a large reduction in computational cost.

Significance. If the central claim holds, this is a practically valuable extension of the aux-DoF approach of Gelin, Velardo, and Borrelli [J. Chem. Phys. 155, 134102 (2021)] to the important case of disorder-dependent thermal equilibrium initial states, which previously required many independent realizations. The formal derivation from the definition of the disorder average and the thermal purification is clean and, for untruncated states, exact. The numerical benchmarks are encouraging: the MPS-sampling results agree with direct sampling across disorder strengths, the required bond dimensions grow only moderately for the tested parameters, and the wall-time comparison in Fig. 5 indicates a substantial speedup. The paper also provides a useful practical discussion of DVR choices, including the prediction of Poincaré recurrence times for SineDVR grids. The main gap is that the error analysis is incomplete: the central estimator is validated by M-convergence of the disorder-averaged correlation function in one 4-site benchmark, but no per-branch truncation check and no statistical error bars are reported. These additions would make the accuracy claim fully supported.

major comments (3)
  1. [Section II.C, Eqs. (22)-(25)] The central claim that Eq. (25) accurately captures each disorder contribution rests on the accuracy of the conditional states |Ψβ(s_k)⟩, |ΨL(s_k)⟩, and |ΨR(s_k)⟩ extracted from a globally truncated MPS after TD-DMRG evolution. The paper provides M-convergence only for the disorder-averaged C(t) in one 4-site benchmark (Fig. 3(d-f)). Global truncation minimizes a global norm error; it does not directly control per-branch relative errors, and the ratio in Eq. (25) is not protected against branch-direction errors by cancellation. Since the method aims to replace many independent simulations with one simulation, please add a per-branch check: for several sampled s_k, compare the extracted conditional states and the ratio in Eq. (25) against independent direct-sampling calculations at increasing M, or provide an error bound that relates the global truncation error to the error in the ratio. Without this, the general accuracy claim is not fully supported, although the presented Holstein results are encouraging.
  2. [Section III, Figs. 3 and 4] The comparison between MPS-sampling and direct sampling is presented as single curves without statistical uncertainties. For a method whose abstract claims reduced statistical errors, the quantitative statement requires error bars: report standard errors of the mean (or bootstrap intervals) for both estimators and confirm that the deviations in Figs. 3(d-f) and 4(a) are within those uncertainties. This is particularly important because increasing Nsamp reduces Monte Carlo noise but does not reduce a possible systematic truncation bias.
  3. [Section II.C, Eq. (20)] Equation (20) is formally ambiguous: it places a fraction of operator-valued expressions inside ⟨χ|·|χ⟩. The intended meaning must be the pointwise-in-s ratio, i.e., ∫ds ρ(s) N(s)/D(s), which can be written as ⟨χ|D(ŝ)^{-1}N(ŝ)|χ⟩. Please define this notation explicitly, since Eq. (21) correctly warns against the alternative (a ratio of expectation values).
minor comments (4)
  1. [Section II.C, Eq. (17)] The word "fullfilling" should be "fulfilling".
  2. [Section II.C, text after Eq. (25)] The phrase "according to the discretised distribution function ρ(s)" should specify how the DVR quadrature weights enter the sampling probabilities and how |χ⟩ is normalized after discretization.
  3. [Section III, Fig. 2] The recurrence-time estimate T = 2π/Δs is derived for equally spaced SineDVR grids; the text and caption should state explicitly that this estimate does not apply to SHODVR, where no vertical dashed line is shown.
  4. [Abstract and Conclusion] The term "one-shot" should be qualified: the time evolution is a single simulation, but Eq. (25) still requires sample-by-sample contractions of conditional states. The wall-time comparisons in Fig. 5 are for a fixed Nsamp = 3000, and the text should make clear that the claimed speedup is for that sample size.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: Eq. (25) follows algebraically from the stated definitions and is benchmarked against independent direct sampling.

full rationale

The central estimator, Eq. (25), is derived from the disorder-averaged thermal TCF, Eq. (20), by inserting the auxiliary wavefunction |χ⟩ = ∫ ds √ρ(s)|s⟩ of Eq. (9), projecting the global MPS states of Eqs. (22)-(24) onto a disorder configuration s_k, and canceling the common factor √ρ(s_k) in the numerator and denominator. This is exact algebra from stated definitions, not a fitted or self-referential reduction. The numerical approximations used in practice (TD-DMRG truncation, finite DVR grids, and finite Nsamp) are validated against direct sampling in Figs. 2-4. The main weakness noted in the paper—uncontrolled per-branch truncation error in individual disorder realizations—is a numerical accuracy concern, not a circularity: nothing in Eq. (25) is defined in terms of the target correlation function, and no parameter is fitted to that target. Self-citations (Refs. 10, 25, 36, 37, 39) support background methodology and tools but are not load-bearing for the derivation of Eq. (25). The recurrence-time estimate T ≈ 2π/Δs is derived from the grid spacing, not from the computed TCF. Therefore no circular step is present.

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

The method rests on standard purification, the encoding of ρ(s) in an auxiliary wavefunction, and the unproven numerical assumption that a single truncated MPS faithfully represents all disorder branches. Model parameters in the Holstein benchmarks are standard values, not fitted to any experiment. Convergence parameters M, Nb, grid truncation, and Nsamp are chosen by hand and are listed as free parameters.

free parameters (4)
  • SineDVR grid truncation width = ±5σ
    The continuous Gaussian disorder is truncated to [-5σ, 5σ], chosen by hand; the authors report a truncation error of 5.74e-7. This is a numerical approximation parameter, not a physical constant.
  • DVR grid points per auxiliary DoF Nb = 21 for σ=50 meV; 31 for σ=100 and 200 meV
    Basis size chosen by convergence tests; sparse or poorly placed grids produce artificial Poincare recurrences in the TCF.
  • MPS bond dimension M = 16 to 64 depending on disorder strength and system size
    Truncation parameter chosen so that TCFs converge; it directly controls the claimed moderate cost increase.
  • Sample count Nsamp = 3e3 to 1e6
    Number of disorder configurations drawn for Eq. 25; larger values reduce statistical error at modest cost, which underlies the claim that statistical noise can be reduced cheaply.
assumptions (4)
  • domain assumption Representation of the disorder distribution by an auxiliary wavefunction |χ⟩ = ∫ ds sqrt(ρ(s)) |s⟩.
    Used in Eqs. 9-13. Exact in the continuum limit for any normalized ρ, but becomes approximate with finite DVR grids.
  • standard math The purification identity Tr_Q |ψβ=0⟩⟨ψβ=0| = I/d^N and imaginary-time evolution generate the thermal equilibrium state.
    Used in Eqs. 17-18; this is the standard thermofield/purification construction.
  • domain assumption Static disorder parameters appear as c-numbers in H, A, and B for each realization, so diagonalizing the auxiliary coordinates gives well-defined per-configuration ratios.
    Used in Eqs. 10-12 and 20. Valid for the Holstein model considered, but limits generalization if observables were not diagonal in the disorder parameters.
  • ad hoc to paper The MPS representing the superposition over disorder branches stays compressible, and per-branch ratios extracted from the truncated MPS are accurate.
    Sections II B/C and Fig. 3. No error bound is given, only numerical evidence for moderate disorder below the electronic bandwidth.
invented entities (1)
  • Auxiliary disorder coordinates |s⟩ with amplitude sqrt(ρ(s))
    purpose: Encode the static disorder distribution into a single quantum state so one MPS time evolution replaces thousands of disorder realizations.
    Computational construction in Eq. 9, not a physical observable or particle, and it has no falsifiable handle outside the simulation. It extends Gelin et al.'s auxiliary DoF idea to coordinate space.

how reviews work

0 comments
Cite this review

Pith. "Pith review of One-Shot Simulation of Static Disorder in Quantum Dynamics with Equilibrium Initial State via Matrix Product State Sampling." pith.science (2026). https://pith.science/paper/Q55C6VFC

@misc{pith2026250607120,
  author       = {Pith},
  title        = {Pith review of: One-Shot Simulation of Static Disorder in Quantum Dynamics with Equilibrium Initial State via Matrix Product State Sampling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Q55C6VFC}},
  note         = {Machine review of arXiv:2506.07120}
}
read the original abstract

Static disorder plays a crucial role in the electronic dynamics and spectroscopy of complex molecular systems. Traditionally, obtaining observables averaged over static disorder requires thousands of realizations via direct sampling of the disorder distribution, leading to high computational costs. In this work, we extend the auxiliary degree-of-freedom based matrix product state (MPS) method to handle system-bath correlated thermal equilibrium initial states. We validate the effectiveness of the extended method by computing the dipole-dipole time correlation function of the Holstein model relevant to the emission spectrum of molecular aggregates. Our results show that the method accurately captures static disorder effects using a one-shot quantum dynamical simulation, with only a moderate increase in MPS bond dimension, thereby significantly reducing computational cost. Moreover, it enables the generation of a much larger number of samples than the conventional direct sampling method at negligible additional cost, thus reducing statistical errors. This method provides a broadly useful tool for calculating equilibrium time correlation functions in system-bath coupled models with static disorder.

Figures

Figures reproduced from arXiv: 2506.07120 by the authors.

Figure 1
Figure 1. FIG. 1. Contraction diagram of [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Time correlation function [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. (d-f) compare the MPS-sampling method with the direct sampling method (same 10 [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Time correlation function [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. Wall time for computing the dipole-dipole time correlation function of the Holstein model [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

41 extracted references · 41 canonical work pages

  1. [1]

    author author J. M. \ Moix , author M. Khasin , \ and\ author J. Cao ,\ @noop journal journal New J. Phys. \ volume 15 ,\ pages 085010 ( year 2013 ) NoStop

    work page 2013

  2. [2]

    Zhong , author Y

    author author X. Zhong , author Y. Zhao , \ and\ author J. Cao ,\ @noop journal journal New J. Phys. \ volume 16 ,\ pages 045009 ( year 2014 ) NoStop

    work page 2014

  3. [3]

    author author N. J. \ Hestand \ and\ author F. C. \ Spano ,\ @noop journal journal Chem. Rev. \ volume 118 ,\ pages 7069 ( year 2018 ) NoStop

    work page 2018

  4. [4]

    Peng , author Q

    author author X. Peng , author Q. Li , \ and\ author Z. Shuai ,\ @noop journal journal Nanoscale \ volume 13 ,\ pages 3252 ( year 2021 ) NoStop

    work page 2021

  5. [5]

    Brey \ and\ author I

    author author D. Brey \ and\ author I. Burghardt ,\ @noop journal journal J. Phys. Chem. Lett. \ volume 15 ,\ pages 1836 ( year 2024 ) NoStop

    work page 2024

  6. [6]

    Zheng , author J

    author author J. Zheng , author J. Peng , author Y. Xie , author Y. Long , author X. Ning , \ and\ author Z. Lan ,\ @noop journal journal Phys. Chem. Chem. Phys. \ volume 22 ,\ pages 18192 ( year 2020 ) NoStop

    work page 2020

  7. [7]

    Zheng , author N

    author author Z. Zheng , author N. R. \ Tummala , author T. Wang , author V. Coropceanu , \ and\ author J.-L. \ Br \'e das ,\ @noop journal journal Adv. Energy Mater. \ volume 9 ,\ pages 1803926 ( year 2019 ) NoStop

    work page 2019

  8. [8]

    author author M. L. \ Chaillet , author F. Lengauer , author J. Adolphs , author F. Muh , author A. S. \ Fokas , author D. J. \ Cole , author A. W. \ Chin , \ and\ author T. Renger ,\ @noop journal journal J. Phys. Chem. Lett. \ volume 11 ,\ pages 10306 ( year 2020 ) NoStop

    work page 2020

Show all 41 references
  1. [9]

    Li , author C

    author author T. Li , author C. Huang , author S. Bai , \ and\ author Q. Shi ,\ @noop journal journal J. Chem. Phys. \ volume 162 ,\ pages 094105 ( year 2025 ) NoStop

    2025

  2. [10]

    Zhang , author Y

    author author Z. Zhang , author Y. Wang , author X. Zheng , author J. Ren , author Z. Shuai , \ and\ author W. Fang ,\ @noop journal journal ChemRxiv,DOI:10.26434/chemrxiv-2025-mk06f \ ( year 2025 ) NoStop

    2025 doi

  3. [11]

    B \"a ssler ,\ @noop journal journal Phys

    author author H. B \"a ssler ,\ @noop journal journal Phys. Status Solidi B \ volume 175 ,\ pages 15 ( year 1993 ) NoStop

    1993

  4. [12]

    Jang , author S

    author author S. Jang , author S. E. \ Dempster , \ and\ author R. J. \ Silbey ,\ @noop journal journal J. Phys. Chem. B \ volume 105 ,\ pages 6655 ( year 2001 ) NoStop

    2001

  5. [13]

    Makri ,\ @noop journal journal J

    author author N. Makri ,\ @noop journal journal J. Phys. Chem. Lett. \ volume 15 ,\ pages 1462 ( year 2024 ) NoStop

    2024

  6. [14]

    author author M. F. \ Gelin , author A. Velardo , \ and\ author R. Borrelli ,\ @noop journal journal J. Chem. Phys. \ volume 155 ,\ pages 134102 ( year 2021 ) NoStop

    2021

  7. [15]

    Schollw \"o ck ,\ @noop journal journal Ann

    author author U. Schollw \"o ck ,\ @noop journal journal Ann. Phys. \ volume 326 ,\ pages 96 ( year 2011 ) NoStop

    2011

  8. [16]

    Sheng , author T

    author author Z. Sheng , author T. Jiang , author W. Li , \ and\ author Z. Shuai ,\ @noop journal journal J. Chem. Theory Comput. \ volume 20 ,\ pages 6470 ( year 2024 ) NoStop

    2024

  9. [17]

    author author A. Nitzan ,\ @noop title Chemical dynamics in condensed phases: relaxation, transfer, and reactions in condensed molecular systems \ ( publisher Oxford university press ,\ year 2024 ) NoStop

    2024

  10. [18]

    Light , author I

    author author J. Light , author I. Hamilton , \ and\ author J. Lill ,\ @noop journal journal J. Chem. Phys. \ volume 82 ,\ pages 1400 ( year 1985 ) NoStop

    1985

  11. [19]

    author author J. C. \ Light \ and\ author T. Carrington Jr ,\ @noop journal journal Adv. Chem. Phys. \ volume 114 ,\ pages 263 ( year 2000 ) NoStop

    2000

  12. [20]

    Vidal ,\ @noop journal journal Phys

    author author G. Vidal ,\ @noop journal journal Phys. Rev. Lett. \ volume 93 ,\ pages 040502 ( year 2004 ) NoStop

    2004

  13. [21]

    author author A. J. \ Daley , author C. Kollath , author U. Schollw \"o ck , \ and\ author G. Vidal ,\ @noop journal journal J. Stat. Mech. Theory Exp. \ volume 2004 ,\ pages P04005 ( year 2004 ) NoStop

    2004

  14. [22]

    author author A. E. \ Feiguin \ and\ author S. R. \ White ,\ @noop journal journal Phys. Rev. B \ volume 72 ,\ pages 020404 ( year 2005 a ) NoStop

    2005

  15. [23]

    Ronca , author Z

    author author E. Ronca , author Z. Li , author C. A. \ Jimenez-Hoyos , \ and\ author G. K.-L. \ Chan ,\ @noop journal journal J. Chem. Theory Comput. \ volume 13 ,\ pages 5560 ( year 2017 ) NoStop

    2017

  16. [24]

    Ma , author Z

    author author H. Ma , author Z. Luo , \ and\ author Y. Yao ,\ @noop journal journal Mol. Phys. \ volume 116 ,\ pages 854 ( year 2018 ) NoStop

    2018

  17. [25]

    Ren , author W

    author author J. Ren , author W. Li , author T. Jiang , author Y. Wang , \ and\ author Z. Shuai ,\ @noop journal journal Wiley Interdiscip Rev Comput Mol Sci \ volume 12 ,\ pages e1614 ( year 2022 ) NoStop

    2022

  18. [26]

    Ma , author U

    author author H. Ma , author U. Schollw \"o ck , \ and\ author Z. Shuai ,\ @noop title Density Matrix Renormalization Group (DMRG)-Based Approaches in Computational Chemistry \ ( publisher Elsevier ,\ year 2022 ) NoStop

    2022

  19. [27]

    Borrelli \ and\ author M

    author author R. Borrelli \ and\ author M. F. \ Gelin ,\ @noop journal journal Wiley Interdiscip Rev Comput Mol Sci \ volume 11 ,\ pages e1539 ( year 2021 ) NoStop

    2021

  20. [28]

    Wang \ and\ author M

    author author H. Wang \ and\ author M. Thoss ,\ @noop journal journal J. Chem. Phys. \ volume 119 ,\ pages 1289 ( year 2003 ) NoStop

    2003

  21. [29]

    Manthe ,\ @noop journal journal J

    author author U. Manthe ,\ @noop journal journal J. Chem. Phys. \ volume 128 ,\ pages 164116 ( year 2008 ) NoStop

    2008

  22. [30]

    Vendrell \ and\ author H.-D

    author author O. Vendrell \ and\ author H.-D. \ Meyer ,\ @noop journal journal J. Chem. Phys. \ volume 134 ,\ pages 044135 ( year 2011 ) NoStop

    2011

  23. [31]

    Wang ,\ @noop journal journal J

    author author H. Wang ,\ @noop journal journal J. Phys. Chem. A \ volume 119 ,\ pages 7951 ( year 2015 ) NoStop

    2015

  24. [32]

    author author M. H. \ Beck , author A. J \"a ckle , author G. A. \ Worth , \ and\ author H.-D. \ Meyer ,\ @noop journal journal Phys. Rep. \ volume 324 ,\ pages 1 ( year 2000 ) NoStop

    2000

  25. [33]

    author author A. E. \ Feiguin \ and\ author S. R. \ White ,\ @noop journal journal Phys. Rev. B \ volume 72 ,\ pages 220401 ( year 2005 b ) NoStop

    2005

  26. [34]

    Verstraete , author J

    author author F. Verstraete , author J. J. \ Garcia-Ripoll , \ and\ author J. I. \ Cirac ,\ @noop journal journal Phys. Rev. Lett. \ volume 93 ,\ pages 207204 ( year 2004 ) NoStop

    2004

  27. [35]

    Holstein ,\ @noop journal journal Ann

    author author T. Holstein ,\ @noop journal journal Ann. Phys. \ volume 8 ,\ pages 343 ( year 1959 ) NoStop

    1959

  28. [36]

    Li , author J

    author author W. Li , author J. Ren , \ and\ author Z. Shuai ,\ @noop journal journal Nat. Commun. \ volume 12 ,\ pages 4260 ( year 2021 ) NoStop

    2021

  29. [37]

    Shuai , author Q

    author author Z. Shuai , author Q. Sun , author J. Ren , author T. Jiang , \ and\ author W. Li ,\ @noop journal journal Aggregate \ volume 6 ,\ pages e70013 ( year 2025 ) NoStop

    2025

  30. [38]

    @noop title Renormalizer,https://github.com/shuaigroup/renormalizer , \ NoStop

  31. [39]

    Li , author J

    author author W. Li , author J. Ren , \ and\ author Z. Shuai ,\ @noop journal journal J. Chem. Phys. \ volume 152 ,\ pages 024127 ( year 2020 ) NoStop

    2020

  32. [40]

    Haegeman , author C

    author author J. Haegeman , author C. Lubich , author I. Oseledets , author B. Vandereycken , \ and\ author F. Verstraete ,\ @noop journal journal Phys. Rev. B \ volume 94 ,\ pages 165116 ( year 2016 ) NoStop

    2016

  33. [41]

    author author P. W. \ Anderson ,\ @noop journal journal Phys. Rev. \ volume 109 ,\ pages 1492 ( year 1958 ) NoStop

    1958

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.