pith. sign in

arxiv: 2605.30142 · v1 · pith:CMHR2THXnew · submitted 2026-05-28 · 🪐 quant-ph

Koopman--von Neumann Molecular Dynamics for Green--Kubo Transport Coefficients

Masari Watanabe , Hirofumi Nishi , Taichi Kosugi , Shigekazu Hidaka , Ryo Sakurai , Yu-ichiro Matsushita This is my paper

Pith reviewed 2026-06-29 07:18 UTC · model grok-4.3

classification 🪐 quant-ph
keywords Koopman-von Neumann representationGreen-Kubo transport coefficientsquantum phase estimationmolecular dynamicsunitary evolutionNVE NVT ensemblesamplitude estimation
0
0 comments X

The pith

The Koopman-von Neumann representation turns classical molecular dynamics into unitary quantum evolutions so Green-Kubo transport coefficients become quantum phase estimation readouts.

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

The paper shows how to recast the integrals that yield transport coefficients such as viscosity from classical particle trajectories as a quantum readout task. It lifts the classical phase-space evolution for both constant-energy and constant-temperature dynamics into unitary operators acting on a Hilbert space whose dimension is set by a position-momentum grid. Because the grid error in the resulting correlation functions falls as a power law in the number of points, the number of qubits needed for a given accuracy grows only logarithmically. Quantum phase estimation applied to a flux-excited state then extracts a windowed version of the integral, and maximum-likelihood amplitude estimation improves the sampling cost from the usual square-root scaling to nearly linear. A reader would care because molecular-dynamics calculations of material properties remain expensive on classical machines, and the construction supplies an explicit route for quantum hardware to contribute to those calculations.

Core claim

Both NVE and Nosé-Hoover-type NVT dynamics are derived as unitary evolutions on Hilbert spaces associated with the corresponding classical phase spaces. Numerical benchmarks on finite grids show that the discretization error in the correlation function decreases as a power law in the number of grid points Nz. Equivalently, with Nz=2^nz the error decreases exponentially in the register size nz. To read out a transport coefficient, a flux-excited state is input to quantum phase estimation; the probability P0 of measuring the ancilla register in the all-zero state corresponds to a Bartlett-windowed Green-Kubo integral. With maximum-likelihood amplitude estimation the statistical estimation of P

What carries the argument

The Koopman-von Neumann representation, which associates classical phase-space functions with vectors in a Hilbert space so that the classical time-evolution operator becomes a unitary operator on that space.

If this is right

  • One step of the NVE propagator can be implemented with O(n^2) CX gates where n equals the total number of position and momentum qubits.
  • The centered-difference implementation of the Nosé-Hoover friction term for NVT scales as O(n_ξ n_p 2^{n_p}).
  • A target accuracy ε for the transport coefficient requires only O(log(1/ε)) qubits for the grid register.
  • Maximum-likelihood amplitude estimation applied to the QPE oracle yields statistical scaling close to N_queries^{-1} rather than the direct-sampling N_queries^{-1/2}.

Where Pith is reading between the lines

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

  • The same unitary-lifting construction could be applied to other classical statistical-mechanical ensembles whose propagators admit a phase-space representation.
  • Hybrid algorithms that interleave the KvN propagator with existing quantum Hamiltonian-simulation routines might allow larger phase-space grids without increasing qubit count proportionally.
  • The logarithmic qubit scaling assumes the power-law convergence continues to arbitrarily fine grids; practical limits set by floating-point precision or memory would cap the achievable accuracy on current hardware.

Load-bearing premise

The finite-grid discretization of the classical phase space produces a unitary operator whose long-time correlation functions converge to the continuous Green-Kubo integrals at a power-law rate in the number of grid points.

What would settle it

A numerical test on successively refined grids in which the measured correlation-function error fails to decrease as a power law with Nz or deviates systematically from the classical integral outside statistical fluctuations.

Figures

Figures reproduced from arXiv: 2605.30142 by Hirofumi Nishi, Masari Watanabe, Ryo Sakurai, Shigekazu Hidaka, Taichi Kosugi, Yu-ichiro Matsushita.

Figure 1
Figure 1. Figure 1: conceptually illustrates the readout workflow used in this paper. The real-time dynamics of classical MD are mapped to unitary evolution in KvN-MD, and the time evolution of the flux-excited state is input to QPE. As shown below, the bin-zero probability in QPE corresponds to a Bartlett-windowed Green–Kubo inte￾gral. II. FORMULATION OF KVN-MD WITH A NOSÉ–HOOVER THERMOSTAT In this section, we map classical … view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: shows a numerical validation of the QPE bin￾zero readout for the two-particle coupled cosine system. We apply the same readout formula to both NVE dy￾namics and Nosé–Hoover-type NVT dynamics. The blue curve is the finite-time Green–Kubo integral DGK(t), ob- [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: compares three estimates for implement￾ing e −iδtHˆ3 . The first is the centered-difference Pauli￾decomposition method. The blue solid line gives the an￾alytic prediction for centered-difference Pauli evolution, and the blue circles show the CX counts after Qiskit tran￾spilation. Over the measured range, both exhibit the same np2 np -type growth. Equation (119) gives the closed form of the weighted Pauli s… view at source ↗
Figure 8
Figure 8. Figure 8: shows the cost of the NVT one-step circuit. The fits for nξ = 2, 3, 4 are N NVT CX [PITH_FULL_IMAGE:figures/full_fig_p017_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: FIG. 9 [PITH_FULL_IMAGE:figures/full_fig_p020_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: compares the reduced phase-space distribu￾tions for the two-particle coupled cosine system. For the system register composed of the q and p registers of each particle, we repeatedly apply the same one-step propa￾gator and evaluate P (1)(q1, p1;t) at t = 2.56. The dis￾tributions obtained from the discrete KvN state-vector calculation and from the Qiskit state-vector simulation agree after propagation, with… view at source ↗
Figure 11
Figure 11. Figure 11: (b) shows the state-vector simulation of the QPE circuit. The ancilla-bin distribution obtained from the discrete KvN calculation and that obtained from the Qiskit state-vector simulation agree over all bins to within machine precision. This confirms that the full QPE circuit, including the controlled- U 2 j gates, the in￾verse QFT, and the bit ordering convention, generates the same ancilla probability d… view at source ↗
read the original abstract

We formulate the Green--Kubo transport coefficients of classical molecular dynamics as a readout problem for quantum algorithms using the Koopman--von Neumann (KvN) representation. Both NVE and Nos\'e--Hoover-type NVT dynamics are derived as unitary evolutions on Hilbert spaces associated with the corresponding classical phase spaces. Numerical benchmarks on finite grids show that the discretization error in the correlation function decreases as a power law in the number of grid points $N_z$. Equivalently, with $N_z=2^{n_z}$, the error decreases exponentially in the register size $n_z$, so a target accuracy $\epsilon$ requires $n_z=\mathcal{O}(\log(1/\epsilon))$ qubits. To read out a transport coefficient, we input a flux-excited state to quantum phase estimation (QPE). The probability $P_0$ of measuring the QPE ancilla register in the all-zero state corresponds to a Bartlett-windowed Green--Kubo integral. With maximum-likelihood amplitude estimation, the statistical estimation of $P_0$ defined by this QPE oracle improves from the $N_{\rm queries}^{-1/2}$ scaling of direct shot sampling to scaling close to $N_{\rm queries}^{-1}$. Our circuit-resource analysis shows that one step of the NVE propagator can be built with $\mathcal{O}(n^2)$ CX gates, where $n=n_x+n_p$ is the total number of position and momentum qubits. For the NVT propagator, the centered-difference Pauli-decomposition implementation of the Nos\'e--Hoover friction term scales as $\mathcal{O}(n_\xi n_p\,2^{n_p})$, where $n_p$ and $n_\xi$ are the numbers of momentum and thermostat qubits, respectively. The proposed framework is a concrete step toward translating the principles of quantum algorithms into the transport-coefficient calculations required in practical molecular simulation.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 0 minor

Summary. The paper formulates Green-Kubo transport coefficients of classical molecular dynamics as a quantum readout problem via the Koopman-von Neumann (KvN) representation. It derives unitary evolutions on associated Hilbert spaces for both NVE and Nosé-Hoover NVT dynamics, reports numerical benchmarks showing power-law decay of discretization error in correlation functions with grid size N_z (hence exponential decay in qubit number n_z = log2(N_z)), and proposes using quantum phase estimation (QPE) on a flux-excited state with maximum-likelihood amplitude estimation to extract a Bartlett-windowed integral of the correlation function. Circuit-resource estimates are given for the propagators.

Significance. If the discretization convergence holds, the work supplies a concrete mapping from classical transport-coefficient calculations to quantum algorithms, with explicit O(n^2) CX scaling for the NVE step and an amplitude-estimation improvement from N_queries^{-1/2} to near N_queries^{-1}. The numerical benchmarks on small grids and the parameter-free character of the KvN-to-quantum mapping are positive features.

major comments (1)
  1. [Abstract] Abstract (discretization benchmarks paragraph): the headline claim that n_z = O(log(1/ε)) suffices for target accuracy ε rests on the finite-grid unitary U_{N_z} producing correlation functions whose time integrals converge to the continuous Green-Kubo value at a power-law rate in N_z. Only numerical benchmarks on small grids are reported; no operator-norm bound, weak-convergence argument, or analysis establishing that the grid KvN propagator approximates the continuous Liouville evolution in the topology required for ∫C(t)dt as N_z→∞ and t→∞ simultaneously is supplied. This assumption is load-bearing for both the qubit scaling and the subsequent QPE readout argument.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading and for identifying this key assumption in our work. We respond to the major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract (discretization benchmarks paragraph): the headline claim that n_z = O(log(1/ε)) suffices for target accuracy ε rests on the finite-grid unitary U_{N_z} producing correlation functions whose time integrals converge to the continuous Green-Kubo value at a power-law rate in N_z. Only numerical benchmarks on small grids are reported; no operator-norm bound, weak-convergence argument, or analysis establishing that the grid KvN propagator approximates the continuous Liouville evolution in the topology required for ∫C(t)dt as N_z→∞ and t→∞ simultaneously is supplied. This assumption is load-bearing for both the qubit scaling and the subsequent QPE readout argument.

    Authors: We agree that the headline qubit scaling n_z = O(log(1/ε)) is supported only by the reported numerical benchmarks on small grids showing power-law decay of discretization error in the correlation functions, rather than by a rigorous operator-norm bound or weak-convergence analysis of the discretized KvN propagator in the topology needed for the integrated Green-Kubo quantity under the joint limits N_z o ∞ and t o ∞. The manuscript presents these benchmarks as empirical evidence but does not claim or supply such a proof. We will revise the abstract to state explicitly that the power-law convergence (and consequent logarithmic qubit scaling) is observed in numerical benchmarks on finite grids, and we will add a brief discussion noting that a rigorous continuum-limit analysis remains an open question for future work. This revision will be made to ensure the load-bearing assumption is presented accurately. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation is a direct mapping with numerical support

full rationale

The paper constructs a formulation mapping classical Green-Kubo integrals to quantum readout via KvN unitary operators on discretized phase space, derives the NVE/NVT propagators explicitly, and reports numerical power-law decay of discretization error in correlation functions to support the logarithmic qubit claim. No equation reduces the target transport coefficient to a fitted input by construction, no self-citation is load-bearing for the central claim, and the discretization convergence is presented as an observed benchmark rather than an unverified self-referential assumption. The chain remains independent of its own outputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the standard mathematical properties of the Koopman-von Neumann lift and on the assumption that a finite-grid truncation yields a unitary operator whose correlation functions converge to the continuous case; no free parameters or invented entities are introduced in the abstract.

axioms (1)
  • domain assumption The Koopman-von Neumann representation maps classical phase-space functions to operators on a Hilbert space such that the Liouville operator becomes a Hermitian operator generating unitary evolution.
    Invoked to obtain unitary NVE and NVT propagators from classical equations of motion.

pith-pipeline@v0.9.1-grok · 5907 in / 1365 out tokens · 26797 ms · 2026-06-29T07:18:18.274797+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

43 extracted references · 2 canonical work pages · 2 internal anchors

  1. [1]

    (A1) The coordinate q is a physically periodic variable

    T est Model and Error Metric For the test system, we use the one-dimensional cosine potential V (q) = V0 cos q, q 2 [0, 2π). (A1) The coordinate q is a physically periodic variable. By contrast, the momentum p and the Nosé–Hoover variable ξ are nonperiodic variables and are numerically truncated to the finite intervals p 2 [pmax, pmax), ξ 2 [ξmax, ξmax). ...

  2. [2]

    By contrast, the p and ξ axes are intrinsically nonperiodic

    Boundaries of Nonperiodic Axes and W rap-Around Error The q axis is physically periodic and is therefore com- patible with Fourier differentiation. By contrast, the p and ξ axes are intrinsically nonperiodic. When Fourier differentiation is used on a finite grid, however, these fi- nite intervals are numerically treated as periodically con- nected. Conseq...

  3. [3]

    Figures 9(c) and 9(d) show the results obtained by fixing either ξmax or pmax to 8σ and varying Nξ and Np

    Grid Resolution and the Time-Discretization Floor After the boundary width is made sufficiently large, the remaining errors are controlled by the number of grid points and the time step. Figures 9(c) and 9(d) show the results obtained by fixing either ξmax or pmax to 8σ and varying Nξ and Np. In the low-resolution regime, even when the Gaussian tail is con...

  4. [4]

    K + 2 K−1X s=1 (K s)cD(s∆t) # . (B13) Finally, multiplying Eq. ( B13) by 1 2 C (D) JJ (0)τQPE and using τQPE = K∆t, we obtain 1 2 C (D) JJ (0)τQPEP0 = C (D) JJ (0)∆t

    Recommended Discretization and the N = 2 n Grid Representation Table II summarizes the practical discretization con- ditions inferred from Fig. 9 for the one-particle one- dimensional cosine potential. These values are represen- tative for the present model. Systems with stiffer poten- tials, lower-temperature distributions, or longer correla- tion times ...

  5. [5]

    For a diffusion coefficient, setting CJJ = Cvv gives Eq

    The window vanishes at s = K, so no endpoint contri- bution appears there. For a diffusion coefficient, setting CJJ = Cvv gives Eq. ( 92) in the main text. 22 Appendix C: Details of the Centered-Difference Pauli-Decomposition Method This appendix summarizes the centered-difference Pauli-decomposition method used in Sec. VII A. The tar- get operator is the ...

    1918

  6. [6]

    First, to validate SD 2 (∆t) without QPE ancil- las, we compare the discrete KvN state-vector calcula- tion with the Qiskit state-vector simulation

    V alidation Design The quantum-circuit simulations are performed in two stages. First, to validate SD 2 (∆t) without QPE ancil- las, we compare the discrete KvN state-vector calcula- tion with the Qiskit state-vector simulation. As the ob- servable, we use the reduced phase-space distribution of particle 1, P (1)(q1, p1; t) = X q2,p2 jhq1, p1, q2, p2jψ(t)...

  7. [7]

    For the system register composed of the q and p registers of each particle, we repeatedly apply the same one-step propa- gator and evaluate P (1)(q1, p1; t) at t = 2 .56

    Agreement of Reduced Phase-Space Distributions Figure 10 compares the reduced phase-space distribu- tions for the two-particle coupled cosine system. For the system register composed of the q and p registers of each particle, we repeatedly apply the same one-step propa- gator and evaluate P (1)(q1, p1; t) at t = 2 .56. The dis- tributions obtained from th...

  8. [8]

    Here we use manc = 7 , corresponding to K = 2 7

    QPE Ancilla-Bin Distribution and Finite-Shot Statistics Figure 11(a) shows the convergence of ˆD when the same QPE circuit is executed with a finite number of shots. Here we use manc = 7 , corresponding to K = 2 7. Since ∆t = 0.02, τQPE = K∆t = 2 manc ∆t = 2.56. (E3) 25 102 103 104 105 Nshots 0.08 0.09 0.10 0.11 0.12 ^D (a) DBart(m = 7) = 0:0921 Aer shot ...

  9. [9]

    M. E. Tuckerman, Statistical Mechanics: Theory and Molecular Simulation (Oxford University Press, 2023)

    2023

  10. [10]

    Fallahzadeh and N

    R. Fallahzadeh and N. Farhadian, Solid State Ionics 280, 10 (2015)

    2015

  11. [11]

    Baktash, J

    A. Baktash, J. C. Reid, T. Roman, and D. J. Searles, npj Computational Materials 6, 162 (2020)

    2020

  12. [12]

    Muralidharan, M

    A. Muralidharan, M. I. Chaudhari, L. R. Pratt, and S. B. Rempe, Scientific Reports 8, 10736 (2018)

    2018

  13. [13]

    Kassal, S

    I. Kassal, S. P. Jordan, P. J. Love, M. Mohseni, and A. Aspuru-Guzik, Proc. Natl. Acad. Sci. U.S.A. 105, 18681 (2008)

    2008

  14. [14]

    J. D. Whitfield, J. Biamonte, and A. Aspuru-Guzik, Mol. Phys. 109, 735 (2011)

    2011

  15. [15]

    Y. Su, D. W. Berry, N. Wiebe, N. Rubin, and R. Bab- bush, PRX Quantum 2, 040332 (2021)

    2021

  16. [16]

    P. J. Ollitrault, A. Miessen, and I. Tavernelli, Acc. Chem. Res. 54, 4229 (2021)

    2021

  17. [17]

    Joseph, Phys

    I. Joseph, Phys. Rev. Research 2, 043102 (2020)

    2020

  18. [18]

    Simon, R

    S. Simon, R. Santagati, M. Degroote, N. Moll, M. Streif, and N. Wiebe, PRX Quantum 5, 010343 (2024)

    2024

  19. [19]

    Fullqubit alchemist: Quantum algorithm for alchemical free energy calculations

    P.-W. Huang, G. Boyd, G.-L. R. Anselmetti, M. De- groote, N. Moll, R. Santagati, M. Streif, B. Ries, D. Marti-Dafcik, H. Jnane, S. Simon, N. Wiebe, T. R. Bromley, and B. Koczor, arXiv arXiv:2508.16719, 10.48550/arXiv.2508.16719 (2025), arXiv:2508.16719

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2508.16719 2025

  20. [20]

    S. Nosé, J. Chem. Phys. 81, 511 (1984)

    1984

  21. [21]

    W. G. Hoover, Phys. Rev. A 31, 1695 (1985)

    1985

  22. [22]

    B. O. Koopman, Proc. Natl. Acad. Sci. U.S.A. 17, 315 (1931)

    1931

  23. [23]

    J. v. Neumann, Annals of Mathematics 33, 587 (1932)

    1932

  24. [24]

    Brassard, P

    G. Brassard, P. Høyer, M. Mosca, and A. Tapp, in Quantum Computation and Information , Contemporary Mathematics, Vol. 305 (American Mathematical Society,

  25. [25]

    Suzuki, S

    Y. Suzuki, S. Uno, R. Raymond, T. Tanaka, T. Onodera, and N. Yamamoto, Quantum Inf. Process. 19, 75 (2020)

    2020

  26. [26]

    Grinko, J

    D. Grinko, J. Gacon, C. Zoufal, and S. Woerner, npj Quantum Inf. 7, 52 (2021)

    2021

  27. [27]

    D. I. Bondar, F. Gay-Balmaz, and C. Tronci, Proc. R. Soc. A 475, 20180879 (2019)

    2019

  28. [28]

    G. J. Martyna, M. L. Klein, and M. Tuckerman, J. Chem. Phys. 97, 2635 (1992)

    1992

  29. [29]

    H. F. Trotter, Proceedings of the American Mathematical Society 10, 545 (1959) . 26

    1959

  30. [30]

    Suzuki, Communications in Mathematical Physics 51, 183 (1976)

    M. Suzuki, Communications in Mathematical Physics 51, 183 (1976)

    1976

  31. [31]

    Suzuki, Phys

    M. Suzuki, Phys. Lett. A 146, 319 (1990)

    1990

  32. [32]

    Yoshida, Phys

    H. Yoshida, Phys. Lett. A 150, 262 (1990)

    1990

  33. [33]

    I. D. Kivlichan, C. Gidney, D. W. Berry, N. Wiebe, J. McClean, W. Sun, Z. Jiang, N. Rubin, A. Fowler, A. Aspuru-Guzik, H. Neven, and R. Babbush, Quantum 4, 296 (2020)

    2020

  34. [34]

    Welch, D

    J. Welch, D. Greenbaum, S. Mostame, and A. Aspuru- Guzik, New J. Phys. 16, 033040 (2014)

    2014

  35. [35]

    Leimkuhler and C

    B. Leimkuhler and C. Matthews, J. Chem. Phys. 138, 174102 (2013)

    2013

  36. [36]

    Quantum computing with Qiskit

    A. Javadi-Abhari, M. Treinish, K. Krsulich, C. J. Wood, J. Lishman, J. Gacon, S. Martiel, P. D. Nation, L. S. Bishop, A. W. Cross, B. R. Johnson, and J. M. Gambetta, Quantum computing with Qiskit (2024), arXiv:2405.08810 [quant-ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  37. [37]

    C. Z. Cheng and G. Knorr, J. Comput. Phys. 22, 330 (1976)

    1976

  38. [38]

    A. J. Klimas, J. Comput. Phys. 68, 202 (1987)

    1987

  39. [39]

    A. J. Klimas and W. M. Farrell, J. Comput. Phys. 110, 150 (1994)

    1994

  40. [40]

    G. V. Vogman, P. Colella, and U. Shumlak, J. Comput. Phys. 277, 101 (2014)

    2014

  41. [41]

    V. V. Shende, S. S. Bullock, and I. L. Markov, IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst. 25, 1000 (2006)

    2006

  42. [42]

    A. M. Childs and N. Wiebe, Quantum Inf. Comput. 12, 901 (2012)

    2012

  43. [43]

    G. H. Low and I. L. Chuang, Quantum 3, 163 (2019)

    2019