Pith. sign in

REVIEW 2 major objections 4 minor 1 cited by

On the Discretization Error of the Discrete Generalized Quantum Master Equation

T0 review · 2 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The transfer tensor method's discrete memory kernel is a consistent discretization of the continuous Nakajima-Zwanzig kernel, with a definite and correctable initial-time offset.

desk verdict A useful, honest response to Makri's critique with plausible numerics, but the proof sketch leans on a scaling assumption that does not follow from Young; referee with a request for the fuller derivation. read the letter →

arxiv 2507.19323 v1 pith:MRNPI4TG submitted 2025-07-25 quant-ph physics.chem-ph

Ruojing Peng , Lachlan P. Lindoy , Joonho Lee This is my paper · ORCID
classification quant-phphysics.chem-ph
keywords transfertensormethodNakajima-Zwanzigquantummasterequationgeneralizedmemorykerneldiscretizationerrornon-Markovianopendynamicsspin-bosonmodelinitial-timecorrection
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

The paper claims that the transfer tensor method (TTM), a numerical recipe for non-Markovian open quantum dynamics, is not an ad hoc discretization but a consistent discrete version of the exact Nakajima-Zwanzig quantum master equation. Specifically, the discrete-time memory kernel $\mathbb{K}_N$ extracted by TTM agrees with the continuous-time memory kernel $\mathcal{K}(N\Delta t)$ up to a term of order $\Delta t$ for $N>0$, and at the initial time $N=0$ it equals a definite combination of the continuous kernel and the system Liouvillian, not a spurious artifact. If correct, this resolves a recent criticism that TTM contains an overcounted convolution at $t=0$, and it means memory kernels reconstructed from exact short-time dynamics converge to the true continuous kernel as $\Delta t\to 0$. The paper also shows that an alternative midpoint scheme is viable but not uniformly more accurate, depending on whether one wants kernels from dynamics or dynamics from kernels.

What carries the argument

The load-bearing object is the discrete-time memory kernel $\mathbb{K}_m$ defined by the TTM recursion $U_{N+1} = L U_N + \Delta t^2 \sum_{m=0}^{N} \mathbb{K}_m U_{N-m}$, with $L = I - i\Delta t \mathcal{L}_s$ and $U_N$ the exact system propagator at $N\Delta t$. The argument compares this recursion against a Taylor expansion of $U_{N+1}$ at $t = N\Delta t$ in which the continuous Nakajima-Zwanzig equation's first, second, and third derivatives are discretized by right Riemann sums and the trapezoidal rule. Matching like terms produces the identities (10) and (18); the machinery therefore consists of a side-by-side expansion of the exact propagator and the discrete recursion, with the convolution kernel $\mathcal{F}(t) = \{\mathcal{K}(t), -i\mathcal{L}_s\} + \int_0^t \mathcal{K}(\tau)\mathcal{K}(t-\tau)\,d\tau$ absorbing the cross terms that control the $O(\Delta t)$ coefficient.

What would settle it

Compute exact propagators $U_N$ and continuous kernels $\mathcal{K}_N$ for a model with an algebraically decaying, long-memory kernel (e.g. a sub-Ohmic spin-boson model at low temperature), extract $\mathbb{K}_N$ by the TTM recursion, and test whether $\|\mathbb{K}_N - \mathcal{K}_N - (\Delta t/2)\mathcal{F}_N\|$ is $O(\Delta t^2)$ for $N>0$ and whether $\|\mathbb{K}_0 - \frac{1}{2}[(-i\mathcal{L}_s)^2 + \mathcal{K}_0]\|$ is $O(\Delta t)$. Growth faster than $O(\Delta t^2)$ as $\Delta t$ shrinks, or a fan of errors that does not converge at $N=0$, would falsify the claimed relation.

Watch

Extended reading notes

Core claim

The central discovery is the pair of error identities linking the discrete TTM kernel $\mathbb{K}_N$ to the continuous Nakajima-Zwanzig kernel $\mathcal{K}(N\Delta t)$. For every $N>0$, $\mathbb{K}_N = \mathcal{K}_N + \frac{\Delta t}{2}\mathcal{F}_N + O(\Delta t^2)$, where $\mathcal{F}_N$ is a convolution-plus-anticommutator correction built from $\mathcal{K}$ and the system Liouvillian. At the initial step, $\mathbb{K}_0 = \frac{1}{2}\left((-i\mathcal{L}_s)^2 + \mathcal{K}_0\right) + \frac{\Delta t}{6}\dddot U_0 + O(\Delta t^2)$, so the term previously identified as spurious is actually the correct half-weighting of the continuous kernel at the left endpoint of the trapezoidal rule. Truncating at $O(\Delta t)$ gives the TTM(1) scheme and retaining the $O(\Delta t^2)$ correction gives TTM(2); both treat $N=0$ with the same time-step order as the other points. The claim is validated on the spin-boson model, where reconstructed kernels converge and dynamics propagated from exact kernels are accurate for $\Delta t \le 0.1$.

Load-bearing premise

The error-order claims rest on the assumption that the memory kernel's derivative and the convolution correction stay bounded, in matrix norm, by a constant multiple of the kernel itself; if a system's kernel decays slowly or has sharp features, that scaling may fail and the stated $O(\Delta t^2)$ relation would need re-examination.

Editorial extensions

If this is right

  • Memory kernels extracted by TTM from exact discrete dynamics converge to the continuous Nakajima-Zwanzig kernel as $\Delta t \to 0$, with controlled $O(\Delta t)$ error for TTM(1) and $O(\Delta t^2)$ error for TTM(2) at all time points including $N=0$.
  • The initial-time kernel is not spurious: the simple identification $\mathbb{K}_0 = \mathcal{K}_0$ must be replaced by the half-weighting identity, which removes the apparent overcounting in propagation from $U_0$ to $U_1$.
  • TTM(1) and TTM(2) offer routes to estimate continuous memory kernels from coarse-grid dynamics without solving integral equations on a dense time grid.
  • The midpoint derivative/midpoint integral scheme proposed in the cited criticism is consistent and can give more accurate short-time dynamics from exact kernels, but it is not uniformly more accurate for extracting kernels: its error in $\mathbb{K}_N$ does not decay with $N$ and can exceed the kernel itself.
  • Original TTM propagation that uses only discrete kernels and exact short-time channels carries no additional discretization error beyond the channels themselves and the memory-truncation time.

Reading between the lines

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

  • If the identity holds for general system-bath models, it suggests a cheap consistency check for any data-driven generalized master equation construction: compute $\mathbb{K}_N$ at two small time steps and verify the difference scales as $\Delta t \|\mathcal{F}_N\|$; deviations would signal either insufficiently converged dynamical channels or violation of the kernel-decay assumption.
  • The same Taylor-matching strategy could be applied to other convolution discretizations, predicting their error constants analytically rather than benchmarking them after the fact.
  • The correction term at $N=0$ implies that initialization protocols that discard $\mathbb{K}_0$ entirely are discarding genuine physical information about the kernel's value at zero, not merely a numerical artifact.
  • One can test the claimed $O(\Delta t^2)$ accuracy of TTM(2) beyond spin-boson, for example in discrete-time propagation with a highly non-Markovian, slowly decaying kernel, where the assumption $\|\dot{\mathcal{K}}\|,\|\ddot{\mathcal{K}}\| = O(\|\mathcal{K}\|)$ is most likely to fail.
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.

Referee Report

2 major / 4 minor

Summary. This manuscript analyzes the relationship between the discrete-time transfer tensor method (TTM) memory kernel and the continuous-time Nakajima-Zwanzig (NZ) memory kernel. The central results are Eqs. (10) and (18), which express the TTM kernel at N=0 and N>0 in terms of the continuous kernel, the system Liouvillian, and an auxiliary function F(t), with O(Δt) (TTM(1)) or O(Δt^2) (TTM(2)) corrections. The paper argues that this resolves the 'spurious initial-time term' reported by Makri [28], shows that TTM is a consistent discretization of the NZ-QME, and provides numerical evidence on the spin-boson model that both the extracted kernels and the propagated dynamics converge as Δt→0.

Significance. If the stated error relations are correct, this is an important clarification of an ongoing controversy about the transfer tensor method. The paper provides a resolution of the t=0 correction, a route to convert between discrete and continuous memory kernels with controlled accuracy, and an explicit comparison with the alternative MPD/I scheme. Strengths include reproducible code and data on GitHub, external validation against HEOM, and numerical results consistent with the claimed convergence orders. The main weakness is that the proof is only sketched and depends on an unproved norm-scaling assumption, so the significance of the result is currently not matched by the rigor of the derivation.

major comments (2)
  1. [Section II, after Eq. (5)] The assertion that F(t), dK/dt, d^2K/dt^2, and R(t) are all O(||K(t)||) in Frobenius norm is not established. Young's convolution inequality does not imply pointwise control of F(t) = {K(t), -iL_s} + ∫_0^t K(τ)K(t-τ)dτ by ||K(t)||; for example, a constant kernel K(t)=a gives the convolution term equal to a^2 t, which is not O(a) for large t. The derivative and residual bounds are likewise asserted without proof or a precise reference. Because this scaling is used to convert the summed error terms in Eq. (17) into the local O(Δt^2(D^K_N+D^F_N)) claim in Eq. (18), the central error relation is not proved as written and requires either a proof under stated assumptions or an explicit restriction to kernels for which the scaling holds.
  2. [Section II, Eqs. (17)-(18)] The derivation of Eq. (18) from Eq. (17) is incomplete. The error terms in Eq. (17) are sums over N quadrature points: O(Δt^4 ∑_m (D^K_m+D^F_m)) and (Δt^4/6) ∑_m R_{N-m}U_m. For a fixed physical time t=NΔt, these sums are generically O(Δt^3) under the stated boundedness assumptions, and after division by Δt^2 they contribute O(Δt), not O(Δt^2), to the kernel relation. To obtain the claimed O(Δt^2) error in Eq. (18), the leading O(Δt^3) contributions must cancel or telescope, but no such cancellation is shown. The numerical tests support second-order behavior, but the analytical derivation of Eq. (18) is not self-contained.
minor comments (4)
  1. [Throughout] The notation for the discrete and continuous kernels is easy to confuse: in the main text, K_m denotes both the TTM kernel and the continuous kernel evaluated at mΔt, distinguished only by font. The abstract uses a calligraphic symbol for the continuous kernel; the same distinct notation should be used consistently in the body.
  2. [Eq. (15)] In Eq. (15), the term (1/2)K_N is missing the factor U_0; since U_0=I the expression is correct, but writing (1/2)K_N U_0 explicitly would improve clarity.
  3. [Section III, Fig. 3] The discussion of the FDIO discrepancy with Ref. [28] is noted but not explained. A brief comment on a possible cause (e.g., different treatment of K_0 or normalization) would help the reader assess the source of the difference, especially because the FDIO comparison is used to motivate the TTM correction.
  4. [Fig. 2 and captions] The saturation of the TTM(2) error at approximately 2×Δt_ref is stated but not annotated in the figure; adding a horizontal line or an explicit marker would make the saturation visible. Minor typos also appear: 'ReK01,01 and ReK01,01' in the Fig. 1 caption should refer to the real and imaginary parts, and the acronym 'MPD/I' is written as 'MPI/D' in the Fig. 3 legend.

Circularity Check

1 steps flagged · score 2.0 of 10

Central K_N–K_N relation is derived from first principles; the TTM(2) validation is self-referential because its F_N correction is computed from the target continuous kernel.

  1. self definitional [Section II, Eq. (18) and following TTM(2) paragraph; Section III, Fig. 2]
    "Furthermore, given exact U···_0 and F_n, K_N and K_N can be computed from each other with O(Δt^2) error for all N≥ 0 using Eqs. (10) and (18); we refer to this scheme as TTM(2). Note that the TTM(2) scheme requires accurate computation of U···_0 and F_n. In particular, the integral in Eq. (3) needs to be computed on dense grid, making the TTM(2) scheme much more costly."

    Eq. (18) states K_N = K_N + (Δt/2)F_N + O(Δt^2), while Eq. (3) defines F(t) as a functional of the continuous kernel K(t) itself. Thus in TTM(2) the correction term F_N is not an independent input; it is constructed from the very continuous kernel K_N that the scheme claims to reconstruct from the discrete kernel K_N. Numerically, the paper computes F_N from the reference K_N on a fine grid (Δt_ref = 0.0005) and then subtracts it, so the observed O(Δt^2) agreement verifies the Taylor remainder of Eq. (18) rather than providing an independent reconstruction of K_N from U_N. This makes the TTM(2) validation self-referential, although the analytic relation itself remains a genuine first-principles expansion.

full rationale

The paper's main analytic claim, Eqs. (10) and (18), is obtained by Taylor expansion of the exact system propagator and consistent quadrature of the continuous Nakajima-Zwanzig equation; it is not assumed or fitted. The numerical benchmarks use HEOM, an external and independent simulation method, on the spin-boson model, and the observed convergence of TTM(1) and TTM(2) is consistent with the derived orders. The only noteworthy self-referential element is the TTM(2) scheme: its leading correction F_N is defined through the continuous memory kernel being reconstructed, so the numerical demonstration that TTM(2) reproduces K_N uses the target K_N to build F_N. The paper also leans on the authors' own addendum (Ref. [3]) for the detailed analysis and asserts an O(||K(t)||) scaling that is not fully proved, but these are support and correctness concerns, not reductions of the central derivation to its own inputs. Overall, the derivation is self-contained against external benchmarks, with only a mild self-referential validation of TTM(2).

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The central derivation rests on the Nakajima-Zwanzig equation as the exact continuous-time starting point, on exact discrete propagators in the TTM channels, and on a scaling assumption for F(t) and the kernel derivatives. No free parameters are introduced by the authors: the spin-boson parameters are taken from Ref. [28], and the HEOM truncation parameters are standard numerical convergence controls. No new physical entities are postulated.

assumptions (3)
  • domain assumption The Nakajima-Zwanzig equation (Eq. (1)) is the exact integro-differential equation for the system propagator U(t).
    The entire analysis starts from this equation, which is standard open-quantum-systems theory and is not proved in this paper.
  • domain assumption The discrete-time propagators U_m entering Eq. (6) are exact.
    Section II states 'we assume that {U_m} are exact, which recovers the TTM'. The derived relations between discrete and continuous kernels hold only for exact U_m; the paper does not analyze how errors in U_m propagate into the memory kernels.
  • domain assumption The quantities F(t), the derivatives of K(t), and R(t) are O(||K(t)||) in Frobenius norm.
    Asserted after Eq. (5) with a terse Young's-convolution-inequality justification and references [16,17]. This scaling underlies the O(dt) and O(dt^2) error bounds in Eqs. (10) and (18), but the convolution term K*K in F(t) is not obviously bounded by ||K(t)|| without additional assumptions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of On the Discretization Error of the Discrete Generalized Quantum Master Equation." pith.science (2026). https://pith.science/paper/MRNPI4TG

@misc{pith2026250719323,
  author       = {Pith},
  title        = {Pith review of: On the Discretization Error of the Discrete Generalized Quantum Master Equation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MRNPI4TG}},
  note         = {Machine review of arXiv:2507.19323}
}
abstract

The transfer tensor method (TTM) [Cerrillo and Cao, Phys. Rev. Lett. 2014, 112, 110401] can be considered a discrete-time formulation of the Nakajima-Zwanzig quantum master equation (NZ-QME) for modeling non-Markovian quantum dynamics. A recent paper [Makri, J. Chem. Theory Comput. 2025, 21, 5037] raised concerns regarding the consistency of the TTM discretization, particularly a spurious term at the initial time \( t=0 \). This Communication presents a detailed analysis of the discretization structure of TTM, clarifying the origin of the initial-time correction and establishing a consistent relationship between the TTM discrete-time memory kernel \( K_N \), and the continuous-time NZ-QME kernel \( \mathcal{K}(N\Delta t) \). This relationship is validated numerically using the spin-boson model, demonstrating convergence of reconstructed memory kernels and accurate dynamical evolution as \( \Delta t \to 0 \). While TTM provides a consistent discretization, we note that alternative schemes are also viable, such as the midpoint derivative/midpoint integral scheme proposed in Makri's work. The relative performance of various schemes for either computing accurate \( \mathcal{K}(N\Delta t) \) from exact dynamics, or obtaining accurate dynamics from exact \( \mathcal{K}(N\Delta t) \), warrants further investigation.

Figures

Figures reproduced from arXiv: 2507.19323 by the authors.

Figure 1
Figure 1. FIG. 1: Elements of continuous time memory kernel [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2: Error of continuous time memory kernel [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3: RDM elements Re [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: FIG. 4: log [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Process Tensor Approaches to Non-Markovian Quantum Dynamics

    quant-ph 2025-09 conditional novelty 2.0 of 10

    A review of process tensor methods for non-Markovian open quantum systems, showing how tensor network representations unify several existing approaches.

Reference graph

Works this paper leans on

44 extracted references · 37 canonical work pages · cited by 1 Pith paper

  1. [28]

    Makri, Discrete generalized quantum master equa- tions, Journal of Chemical Theory and Computation 0, null (0), pMID: 40326041

    N. Makri, Discrete generalized quantum master equa- tions, Journal of Chemical Theory and Computation 0, null (0), pMID: 40326041

    work page

  2. [1]

    Nakajima, On quantum theory of transport phenom- ena: Steady diffusion, Progress of Theoretical Physics 20, 948 (1958)

    S. Nakajima, On quantum theory of transport phenom- ena: Steady diffusion, Progress of Theoretical Physics 20, 948 (1958)

    1958

  3. [2]

    Zwanzig, Ensemble method in the theory of irre- versibility, The Journal of Chemical Physics 33, 1338 (1960)

    R. Zwanzig, Ensemble method in the theory of irre- versibility, The Journal of Chemical Physics 33, 1338 (1960)

    1960

  4. [3]

    Ivander, L

    F. Ivander, L. P. Lindoy, and J. Lee, Unified framework for open quantum dynamics with memory, Nature Com- munications 15, 8087 (2024)

    work page 2024

  5. [4]

    Feynman and F

    R. Feynman and F. Vernon, The theory of a general quan- tum system interacting with a linear dissipative system, Annals of Physics 24, 118 (1963)

    1963

  6. [5]

    Makri and D

    N. Makri and D. E. Makarov, Tensor propagator for it- erative quantum time evolution of reduced density ma- trices. i. theory, The Journal of Chemical Physics 102, 4600 (1995)

    work page 1995

  7. [6]

    D. E. Makarov and N. Makri, Path integrals for dissipa- tive systems by tensor multiplication. condensed phase quantum dynamics for arbitrarily long time, Chemical Physics Letters 221, 482 (1994)

    work page 1994

  8. [7]

    Makri, Small matrix path integral with extended memory, Journal of Chemical Theory and Computation 17, 1 (2021), pMID: 33430598

    N. Makri, Small matrix path integral with extended memory, Journal of Chemical Theory and Computation 17, 1 (2021), pMID: 33430598

    work page 2021

Show all 44 references
  1. [8]

    Makri, Small matrix disentanglement of the path in- tegral: Overcoming the exponential tensor scaling with memory length, The Journal of Chemical Physics 152, 041104 (2020)

    N. Makri, Small matrix disentanglement of the path in- tegral: Overcoming the exponential tensor scaling with memory length, The Journal of Chemical Physics 152, 041104 (2020)

    2020

  2. [9]

    Kilgour, B

    M. Kilgour, B. K. Agarwalla, and D. Segal, Path-integral methodology and simulations of quantum thermal trans- port: Full counting statistics approach, The Journal of Chemical Physics 150, 084111 (2019)

    2019

  3. [10]

    Simine and D

    L. Simine and D. Segal, Path-integral simulations with fermionic and bosonic reservoirs: Transport and dissi- pation in molecular electronic junctions, The Journal of Chemical Physics 138, 214111 (2013)

    2013

  4. [11]

    Makri, Modular path integral methodology for real- time quantum dynamics, The Journal of Chemical Physics 149, 214108 (2018)

    N. Makri, Modular path integral methodology for real- time quantum dynamics, The Journal of Chemical Physics 149, 214108 (2018)

    2018

  5. [12]

    Kundu and N

    S. Kundu and N. Makri, Pathsum: A c++ and fortran suite of fully quantum mechanical real-time path inte- gral methods for (multi-)system + bath dynamics, The Journal of Chemical Physics 158, 224801 (2023)

    2023

  6. [13]

    Bloch, Nuclear induction, Phys

    F. Bloch, Nuclear induction, Phys. Rev. 70, 460 (1946)

    1946

  7. [14]

    A. G. Redfield, On the theory of relaxation processes, IBM Journal of Research and Development 1, 19 (1957)

    1957

  8. [15]

    Mulvihill and E

    E. Mulvihill and E. Geva, A road map to various path- ways for calculating the memory kernel of the general- ized quantum master equation, The Journal of Physical Chemistry B 125, 9834 (2021), pMID: 34424700

    2021

  9. [16]

    Zhang, B

    M.-L. Zhang, B. J. Ka, and E. Geva, Nonequilibrium quantum dynamics in the condensed phase via the gen- eralized quantum master equation, The Journal of Chem- ical Physics 125, 044106 (2006)

    2006

  10. [17]

    Shi and E

    Q. Shi and E. Geva, A new approach to calculating the memory kernel of the generalized quantum master equa- tion for an arbitrary system–bath coupling, The Journal of Chemical Physics 119, 12063 (2003)

    2003

  11. [18]

    Mulvihill, A

    E. Mulvihill, A. Schubert, X. Sun, B. D. Dunietz, and E. Geva, A modified approach for simulating electroni- cally nonadiabatic dynamics via the generalized quantum master equation, The Journal of Chemical Physics 150, 034101 (2019)

    2019

  12. [19]

    Kidon, H

    L. Kidon, H. Wang, M. Thoss, and E. Rabani, On the memory kernel and the reduced system propagator, The Journal of Chemical Physics 149, 104105 (2018)

    2018

  13. [20]

    Chatterjee and N

    S. Chatterjee and N. Makri, Real-time path integral methods, quantum master equations, and classical vs quantum memory, The Journal of Physical Chemistry B 123, 10470 (2019), pMID: 31721584

    2019

  14. [21]

    Cerrillo and J

    J. Cerrillo and J. Cao, Non-markovian dynamical maps: Numerical processing of open quantum trajectories, Phys. Rev. Lett. 112, 110401 (2014)

    2014

  15. [22]

    F. A. Pollock, E. Gull, K. Modi, and G. Cohen, Reduced Dynamics of Full Counting Statistics, SciPost Phys. 13, 027 (2022)

    2022

  16. [23]

    A. A. Kananenka, C.-Y. Hsieh, J. Cao, and E. Geva, Accurate long-time mixed quantum-classical liouville dy- namics via the transfer tensor method, The Journal of Physical Chemistry Letters 7, 4809 (2016), pMID: 27934045

    2016

  17. [24]

    Buser, J

    M. Buser, J. Cerrillo, G. Schaller, and J. Cao, Initial system-environment correlations via the transfer-tensor method, Phys. Rev. A 96, 062122 (2017)

    2017

  18. [25]

    Gelzinis, E

    A. Gelzinis, E. Rybakovas, and L. Valkunas, Applicabil- ity of transfer tensor method for open quantum system dynamics, The Journal of Chemical Physics 147, 234108 (2017)

    2017

  19. [26]

    F. A. Pollock and K. Modi, Tomographically recon- structed master equations for any open quantum dynam- ics, Quantum 2, 76 (2018)

    2018

  20. [27]

    M. R. Jørgensen and F. A. Pollock, Discrete memory ker- nel for multitime correlations in non-markovian quantum processes, Phys. Rev. A 102, 052206 (2020)

    2020

  21. [29]

    W. H. Young, On the multiplication of successions of fourier constants, Proceedings of the Royal Society of London. Series A, Containing Papers of a Mathematical and Physical Character 87, 331 (1912)

    1912

  22. [30]

    V. I. Bogachev, Measure Theory , 2007th ed., Vol. 1 (Springer, Berlin, Germany, 2006)

    2006

  23. [31]

    Kelly, A

    A. Kelly, A. Montoya-Castillo, L. Wang, and T. E. Mark- land, Generalized quantum master equations in and out of equilibrium: When can one win?, The Journal of Chemical Physics 144, 184105 (2016)

    2016

  24. [32]

    Mulvihill and E

    E. Mulvihill and E. Geva, Simulating the dynamics of electronic observables via reduced-dimensionality gener- alized quantum master equations, The Journal of Chem- ical Physics 156, 044119 (2022)

    2022

  25. [33]

    Montoya-Castillo and D

    A. Montoya-Castillo and D. R. Reichman, Approximate but accurate quantum dynamics from the mori formal- ism: I. nonequilibrium dynamics, The Journal of Chem- ical Physics 144, 184104 (2016). 8

    2016

  26. [34]

    Montoya-Castillo and D

    A. Montoya-Castillo and D. R. Reichman, Approximate but accurate quantum dynamics from the mori formal- ism. ii. equilibrium time correlation functions, The Jour- nal of Chemical Physics 146, 084110 (2017)

    2017

  27. [35]

    Brian and X

    D. Brian and X. Sun, Generalized quantum master equa- tion: A tutorial review and recent advances, Chinese Journal of Chemical Physics 34, 497 (2021)

    2021

  28. [36]

    28 there is an additional factor of 1 /2 from T to K, which we believe is a typo from comparing their Eq

    Note that in Table 1 of Ref. 28 there is an additional factor of 1 /2 from T to K, which we believe is a typo from comparing their Eq. (33) and (36), and this was also numerically confirmed

  29. [37]

    Tanimura and R

    Y. Tanimura and R. Kubo, Time evolution of a quantum system in contact with a nearly gaussian-markoffian noise bath, Journal of the Physical Society of Japan 58, 101 (1989)

    1989

  30. [38]

    Tanimura, Numerically “exact” approach to open quantum dynamics: The hierarchical equations of motion (heom), The Journal of Chemical Physics 153, 020901 (2020)

    Y. Tanimura, Numerically “exact” approach to open quantum dynamics: The hierarchical equations of motion (heom), The Journal of Chemical Physics 153, 020901 (2020)

    2020

  31. [39]

    M. Xu, Y. Yan, Q. Shi, J. Ankerhold, and J. T. Stock- burger, Taming quantum noise for efficient low temper- ature simulations of open quantum systems, Phys. Rev. Lett. 129, 230601 (2022)

    2022

  32. [40]

    Q. Shi, Y. Xu, Y. Yan, and M. Xu, Efficient propagation of the hierarchical equations of motion using the matrix product state method, The Journal of Chemical Physics 148, 174102 (2018)

    2018

  33. [42]

    Y. Yan, M. Xu, T. Li, and Q. Shi, Efficient propagation of the hierarchical equations of motion using the tucker and hierarchical tucker tensors, The Journal of Chemical Physics 154, 194104 (2021)

    2021

  34. [43]

    Mangaud, A

    E. Mangaud, A. Jaouadi, A. Chin, and M. Desouter- Lecomte, Survey of the hierarchical equations of motion in tensor-train format for non-markovian quantum dy- namics, The European Physical Journal Special Topics 232, 1847 (2023)

    2023

  35. [44]

    L. P. Lindoy, D. Rodrigo-Albert, Y. Rath, and I. Rung- ger, pyttn: An open source toolbox for open and closed system quantum dynamics simulations using tree tensor networks (2025), arXiv:2503.15460 [quant-ph]

    2025 arXiv

  36. [45]

    Nakatsukasa, O

    Y. Nakatsukasa, O. S` ete, and L. N. Trefethen, The aaa algorithm for rational approximation, SIAM Journal on Scientific Computing 40, A1494 (2018)

    2018

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.