Pith. sign in

REVIEW 2 major objections 2 minor 28 references

Non-Perturbative Topological Gadgets for Many-Body Coupling

T0 review · 2 major / 2 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 This paper claims a non-perturbative gadget built from a domain-wall chain and three- or five-body drivers implements the effective interaction αZ^{⊗nd} exactly in a gapped low-energy subspace.

desk verdict The factor-of-two eigenvalue error is real and breaks the central calibration, but the underlying non-perturbative gadget idea is novel enough to warrant a serious referee. read the letter →

arxiv 2506.05323 v2 pith:FPSCSQMU submitted 2025-06-05 quant-ph

David Headley , Nicholas Chancellor This is my paper
classification quant-ph
keywords non-perturbativegadgetsdomain-wallencodingmany-bodyinteractionstopologicaldefectsquantumannealingminorembeddingsubspacedriversHamiltonian
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

Continuous-time quantum computers generally lack native multi-body terms, so Hamiltonian gadgets that synthesize them from simpler couplings matter for compilation and annealing. This paper shows how to do that without perturbation theory: build a chain of overlapping three-body parity terms between data and ancilla qubits so that unsatisfied terms appear as mobile domain-wall defects, then drive the single-defect subspace. With the driver strength set to β = (γ − α)/(2 cos(π/(nd+1))), the gadget's low-energy manifold acts as the exact effective interaction $αZ^{{⊗nd}}$, with a confinement gap of order γ and no need for paths through penalized high-energy states. Because the construction is non-perturbative, it avoids the exponentially growing energy scales of standard kth-order perturbative gadgets, and it supports encoded logical bit-flips and adjacent XX terms useful for minor-embedded chains.

What carries the argument

The central object is the domain-wall defect sector of a chain Hamiltonian built from overlapping three-body terms, ½(1 − Z_i Z^a_{i−1}Z^a_i), where ancilla qubits carry cumulative parity of the data qubits. When the data parity is unsatisfiable, exactly one term is broken and a topological defect sits at that position j; the single-defect subspace is a one-dimensional hopping model with a tridiagonal Toeplitz Hamiltonian whose eigenstates are sine vectors S_{kj}. The subspace driver (five-body in its clean form, three-body after a gauge transform) moves the defect along the chain, and the gauge unitary U_enc diagonalizes the defect sectors. What carries the argument is the combination: the chain sets the defect energy scale γ, the driver provides a spectrum that can be exactly cancelled by choosing β via Equation (21), and the ground-state overlap S0j between defect states of differing data configurations supplies the matrix elements for encoded logical operators.

What would settle it

Diagonalize the nd×nd tridiagonal matrix with off-diagonal entries 2 appearing in Equation (12) for small nd (say nd = 2, 3, 4). If its lowest eigenvalue is 4 cos(π/(nd+1)) rather than the stated −2 cos(π/(nd+1)), then the β in Equation (21) does not put the one-defect ground state at energy α, and the claimed identity H_eff = $αZ^{{⊗nd}}$ fails.

Watch

Extended reading notes

Core claim

The central claim is that the Hamiltonian H_Gadget = γH_Chain + βH_Subspace, with β = (γ − α)/(2 cos(π/(nd+1))), has a low-energy subspace on which it acts as H_eff = $αZ^{{⊗nd}}$: an nd-body interaction formed from only three-body chain terms plus five-body subspace drivers (or gauge-reduced three-body drivers). The chain term penalizes the number of defects, the subspace driver delocalizes a single defect like a particle in a box, and the chosen β cancels the defect kinetic energy so the one-defect ground state sits at energy α. The paper further shows that a physical X on a data qubit implements a logical X with a known amplitude correction 1/S0j, and that XiXi+1X^a_i implements adjacent logical XX. These operations let one flip a minor-embedded logical chain or drive it as a whole, with leakage suppressed by the O(γ) gap.

Load-bearing premise

Everything rests on one exact eigenvalue calculation: the energy of a single mobile domain wall must be exactly −2 cos(π/(nd+1)) for the chosen driver strength to cancel it; if that number is off, the gadget stops doing what is claimed.

Editorial extensions

If this is right

  • If the construction is correct, an nd-body Z^{⊗nd} term for any nd can be implemented on a linear-connectivity device using three-body chain interactions and at most five-body drivers, with no perturbative expansion needed.
  • The encoded single-qubit X operation, with correction 1/S0j, gives a way to flip a single data qubit inside a minor-embedding chain without breaking the chain constraints, directly addressing the tension between constraint enforcement and spin-flip dynamics on quasi-two-dimensional hardware.
  • The three-body physical term XiXi+1X^a_i implements a logical XX interaction, enabling whole-chain driving and effective many-body X-catalysts in the Hadamard-rotated basis.
  • The performance metrics (leakage, conditional fidelity, absolute fidelity) define an operating regime where high confinement γ makes the gadget act like the ideal logical Hamiltonian while suppressing single-qubit noise perturbations.
  • The same defect-parity mechanism is proposed as a template for ice-like two-body systems, where boundary conditions alone determine whether an odd or even number of defects exist, potentially yielding approximate many-body couplings from purely two-body terms.

Reading between the lines

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

  • Beyond the paper's explicit claims, this suggests that any product-of-Z diagonal Hamiltonian could be compiled from linear-connectivity three-body terms with an energy overhead set by γ rather than by perturbation order, a qualitative improvement over kth-order perturbative gadgets if the spectrum identity holds.
  • The position-dependent overlap S0j implies logical driving is stronger at the chain ends than in the middle; the paper mentions heterogeneous subspace driving as a way to bias amplitude toward the ends, but does not numerically explore this tuning knob.
  • A natural testable extension is to apply the same gadget in the Hadamard basis to build effective XX products (not just ZZ products) for catalyst terms in adiabatic computation; the authors list catalysts as an application but do not simulate them.
  • If the ice-like two-body analogue works, it would beat the three-body chain gadget in locality, and the quasi-one-dimensional geometry suggests matrix-product-state simulations could verify it; this last point is an editorial extrapolation from the paper's outlook.
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

2 major / 2 minor

Summary. The manuscript proposes a class of non-perturbative quantum gadgets based on domain-wall defects in a linear chain of three-body parity terms. A chain Hamiltonian penalizes unsatisfied clauses, while subspace drivers move a single defect along the chain, leading to an effective tight-binding model on the defect position. The authors claim that with a suitable driver strength beta (Eq. 21) the gadget realizes the exact effective interaction H_eff = alpha Z^{⊗nd} in a gapped low-energy subspace (Eq. 22), and they use this to construct encoded logical X and XX operations for minor-embedded chains. The paper includes numerical simulations of leakage, conditional fidelity, and a minor-embedding task.

Significance. If correct, the construction would be a useful alternative to perturbative gadgets, avoiding exponentially small energy scales and providing encoded logical operations with modest-weight physical terms. The paper is clearly written and the idea of exploiting the odd-even, boundary-dependent defect structure is appealing. The analytical diagonalization of the single-defect subspace via a sine transform is elegant, and the performance metrics in Sec. 6 are well defined. However, the central parameter-setting equation rests on an incorrect eigenvalue formula, so the headline result is not currently established. The authors also appropriately acknowledge limitations on composability and the approximate nature of the ice-like extension, which are not themselves problems.

major comments (2)
  1. [Sec. 3.1.2, Eqs. (12), (14), (21), (22)] The matrix in Eq. (12) is a tridiagonal Toeplitz matrix with off-diagonal matrix elements equal to 2. Its eigenvalues are 4 cos(pi (k+1)/(nd+1)) for k = 0,...,nd-1, not -2 cos(...) as stated in Eq. (14). Moreover, with the explicit +2 off-diagonal entries, the lowest eigenvalue occurs at k = nd-1 rather than k = 0, so Eq. (15) assigns the ground state to the wrong eigenvector. Because Eq. (21) is derived from this eigenvalue, the calibration does not place the one-defect subspace at the intended energy: with the authors' beta the one-defect ground-state energy is gamma - 4 beta cos(pi/(nd+1)) = 2 alpha - gamma in the large-nd limit, not alpha. Consequently Eq. (22) does not follow, and the numerical demonstrations in Figures 5-8 do not validate the claimed effective Hamiltonian. The eigenvalue formula and the calibration must be corrected before the central claim can be assessed.
  2. [Sec. 4, Eqs. (20)-(22)] Even if the eigenvalue formula is repaired, the stated calibration gives the zero-defect (satisfiable) sector energy 0 and the one-defect ground-state energy alpha, so the effective Hamiltonian on the two-dimensional parity subspace has eigenvalues {0, alpha}. For the unkinked chain this is (alpha/2)(I - Z^{⊗nd}), not alpha Z^{⊗nd}; the coefficient of the many-body term is alpha/2, and to realize exactly alpha Z^{⊗nd} one would need to place the odd-parity level at -alpha (up to an irrelevant constant), which corresponds to a different sign and magnitude of beta. The authors should either correct Eq. (22) or explicitly define alpha as the energy splitting rather than the coefficient of Z^{⊗nd}.
minor comments (2)
  1. [Sec. 5.1, Eq. (26)] The normalization constant S0j is given as sqrt(2/(nd-1)), but Eq. (13) and the simulation amplitudes in the Figure 6 caption (0.577, 0.5, 0.289 for nd=5) require sqrt(2/(nd+1)). This typo affects the correction factor 1/S0j for encoded single-qubit X operations and should be fixed.
  2. [Figure 5 caption] The statement that evolution from a logical |+> state under Z^{⊗n} for time t = pi/2 produces a GHZ state is not consistent with the standard identity e^{-i Z^{⊗n} t}|+>^n = cos(t)|+>^n - i sin(t)|->^n, which at t = pi/2 gives the product state -i|->^n. Please clarify what quantity labeled 'GHZ fidelity' is actually plotted, or correct the caption.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the gadget is an explicit construction whose parameter beta is a design equation, not a fitted value, and the cited self-work on domain-wall encoding is not load-bearing.

full rationale

The paper's central claim is that H_Gadget = gamma H_Chain + beta H_Subspace realizes H_eff = alpha Z^(⊗nd) in a gapped low-energy subspace. The parameter beta is set by Eq. 21 from the analytically derived single-defect spectrum in Eq. 14. This is a design or calibration equation, not a prediction extracted from data, and Eq. 22 follows by construction from diagonalizing the driver. The domain-wall encoding language is attributed to ref. [21] by one of the authors, but the chain Hamiltonian and defect subspace are derived within the paper itself (Eqs. 3-8); the self-citation only names an established concept and is not load-bearing. The apparent issue with Eq. 14's eigenvalue formula, if real, would be a mathematical correctness concern rather than circularity, because the derivation does not assume the desired conclusion; it attempts to prove it from a stated spectral lemma. Since no 'prediction' reduces to a fitted input and no load-bearing argument reduces to a self-citation, the derivation is self-contained for the purpose of this circularity analysis.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim rests mainly on the spectrum of the one-defect Toeplitz Hamiltonian, the invariance of the one-defect subspace, and the availability of three- or five-body physical terms. No new particles, forces, or conserved quantities are introduced. The free parameters γ and α are design choices for the gadget, not hidden fits.

free parameters (2)
  • Confinement strength γ = Values such as 1.05, 3, 8, 128, 1000 used in figures
    Energy scale penalizing defect states; chosen large relative to α to confine dynamics to the logical subspace. It is a design parameter, not fitted to external data, but is central to the gadget's operation.
  • Target coupling strength α = Set to 1 in numerical demonstrations
    Desired strength of the effective many-body term αZ^{⊗nd}. The user of the gadget chooses α, and β is derived from γ and α via Eq. 21.
assumptions (4)
  • standard math The one-defect effective Hamiltonian (Eq. 12) has eigenvalues λ_k = −2 cos(π(k+1)/(nd+1)) as stated in Eq. 14.
    The paper's parameter choice (Eq. 21) and the claimed effective Hamiltonian (Eq. 22) depend directly on this spectrum. However, for a Toeplitz matrix with off-diagonal elements 2, the standard eigenvalues are 4 cos(π(k+1)/(nd+1)), so this premise is actually violated. Flagged in red_flags.
  • domain assumption The one-defect subspace is invariant under the gadget Hamiltonian, and transitions between different defect-number sectors are suppressed by the confinement gap.
    The subspace invariance holds exactly for the chain and subspace driver as constructed, but the suppression of leakage relies on the gap being large. The scaling of this gap with chain length is not uniform O(γ) as claimed, see red_flags.
  • domain assumption Physical hardware can realize the required three-body chain terms (Eq. 3) and five-body subspace drivers (Eq. 11), or the gauge-reduced three-body variant (Eq. 18).
    The entire gadget construction assumes the native availability of these higher-order physical terms, which is the stated motivation for building gadgets in the first place.
  • domain assumption The virtual ancilla qubits at the chain ends can be fixed to constant +1 or −1 values as boundary conditions.
    The kinked and unkinked chain variants rely on pinning the boundary ancillae, which is a physical assumption about the hardware implementation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Non-Perturbative Topological Gadgets for Many-Body Coupling." pith.science (2026). https://pith.science/paper/FPSCSQMU

@misc{pith2026250605323,
  author       = {Pith},
  title        = {Pith review of: Non-Perturbative Topological Gadgets for Many-Body Coupling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FPSCSQMU}},
  note         = {Machine review of arXiv:2506.05323}
}
read the original abstract

Continuous-time quantum hardware implementations generally lack the native capability to implement high-order terms that would facilitate efficient compilation of quantum algorithms. This limitation has, in part, motivated the development of perturbative gadgets -- multi-qubit constructions used to effect a desired Hamiltonian using engineered low-energy subspaces of a larger system constructed using simpler, usually two-body, primitives. In this work, we demonstrate how a class of non-perturbative gadgets can produce high-order multi-body interactions by taking advantage of the odd-even properties of topological defect subspaces. The simplest example is based on domain-wall defects forming an effective Ising spin-chain based on three-body coupling with linear connectivity, alongside three-, or five-body driving terms depending on the intended use. Although this will be the main focus of the paper due to conceptual simplicity, there exist systems constructed with only two-body couplings where the boundaries determine whether there are an odd or even number of defects, namely ice-like systems which may yield more complex gadget-like constructions.

Figures

Figures reproduced from arXiv: 2506.05323 by the authors.

Figure 1
Figure 1. Left: spin-ice systems follow local ice rules. For example, on a tetrahedral pyrochlore spin-ice lattice, a 2-in 2-out rule is adhered to, represented here in 2d. Black vertices adhere to the rule and the red vertex does not, as three di￾rected edges flow into the vertex and only one out. Right: the same dynamics can be expressed with anti-ferromagnetic two-body interactions and four spins per vertex of the left-han… view at source ↗
Figure 2
Figure 2. Left: Perturbative gadgets work by penalizing a subspace and applying a perturbation which [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. A factor-graph representation of Equation 4 in which blue vertices represent terms in the [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: A factor-graph representation of Equation 11 in which blue vertices represent terms in the [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: a) The gadget Hamiltonian evolves a state initialized in the logical [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Illustration of bit-flipping dynamics in a [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: the sources of error that accumulate from leakage and the impact of the return of leaked am￾plitude to the low-energy sector. This shows the ex￾pected relationship with the instantaneously leaked fraction remaining constant over time for a given level of containment af…
Figure 8
Figure 8. Figure 8: A gadget of five data qubits and four ancilla qubits is initialized in the logical |0⟩ state and allowed to evolve under the action of the gad￾get Hamiltonian in the rotated basis in which it acts as a many-body X operator upon the logical subspace. Single-qubit X term…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

28 extracted references · 23 canonical work pages

  1. [1]

    Minor-embedding in adiabatic quantum computation: I

    Vicky Choi. Minor-embedding in adiabatic quantum computation: I. The parameter set- ting problem. Quantum Information Process- ing, 7(5):193–209, 2008

    work page 2008

  2. [2]

    Error-corrected quantum an- nealing with hundreds of qubits.Nature com- munications, 5(1):3243, 2014

    Kristen L Pudenz, Tameem Albash, and Daniel A Lidar. Error-corrected quantum an- nealing with hundreds of qubits.Nature com- munications, 5(1):3243, 2014

    work page 2014

  3. [3]

    Exponential speed-up of quantum anneal- ing via n-local catalysts

    Roopayan Ghosh, Luca A Nutricati, Natasha Feinstein, PA Warburton, and Sougato Bose. Exponential speed-up of quantum anneal- ing via n-local catalysts. arXiv preprint arXiv:2409.13029, 2024

    arXiv 2024

  4. [4]

    Kitaev, Alexander Shen, and Mikhail N

    Alexei Yu. Kitaev, Alexander Shen, and Mikhail N. Vyalyi. Classical and Quantum Computation, volume 47 ofGraduate Studies in Mathematics. American Mathematical So- ciety, Providence, RI, 2002

    work page 2002

  5. [5]

    3-local hamil- tonian is qma-complete, 2003

    Julia Kempe and Oded Regev. 3-local hamil- tonian is qma-complete, 2003

    work page 2003

  6. [6]

    The complexity of the local hamiltonian prob- lem

    Julia Kempe, Alexei Kitaev, and Oded Regev. The complexity of the local hamiltonian prob- lem. Siam journal on computing, 35(5):1070– 1097, 2006

    work page 2006

  7. [7]

    A quantum annealing architecture with all-to-all connectivity from local interactions

    Wolfgang Lechner, Philipp Hauke, and Peter Zoller. A quantum annealing architecture with all-to-all connectivity from local interactions. Science advances, 1(9):e1500838, 2015

    work page 2015

  8. [8]

    A transmon quantum annealer: Decom- posing many-body ising constraints into pair interactions

    Martin Leib, Peter Zoller, and Wolfgang Lech- ner. A transmon quantum annealer: Decom- posing many-body ising constraints into pair interactions. Quantum Science and Technol- ogy, 1(1):015008, 2016

    work page 2016

Show all 28 references
  1. [9]

    Estarellas

    Ana Palacios, Artur Garcia-Saez, Bruno Juliá- Díaz, and Marta P. Estarellas. Scalable 2-local architecture for quantum annealing of ising models with arbitrary dimensions.Phys. Rev. Appl., 23:054070, May 2025

    2025

  2. [10]

    King, and Jack Raymond

    Kelly Boothby, Andrew D. King, and Jack Raymond. Zephyr topology of D-Wave quan- tum processors, 2021. D-Wave technical re- port: 14-1056A-A

    2021

  3. [11]

    Lukin, Sheng-Tao Wang, and Hannes Pichler

    Minh-Thi Nguyen, Jin-Guo Liu, Jonathan Wurtz, Mikhail D. Lukin, Sheng-Tao Wang, and Hannes Pichler. Quantum optimiza- tion with arbitrary connectivity using rydberg atom arrays. PRX Quantum, 4:010316, Feb 2023

    2023

  4. [12]

    Daley, and Jonathan D

    Johannes Kombe, Gerard Pelegrí, Andrew J. Daley, and Jonathan D. Pritchard. A quan- tum wire approach to weighted combina- torial graph optimisation problems, 2025. arXiv:2503.17115

    2025 arXiv

  5. [13]

    Quantum simulation of one-dimensional fermionic systems with ising hamiltonians

    Matthias Werner, Artur García-Sáez, and Marta P Estarellas. Quantum simulation of one-dimensional fermionic systems with ising hamiltonians. Physical Review B , 111(15):155441, 2025

    2025

  6. [14]

    King, Cristiano Nisoli, Edward D

    Andrew D. King, Cristiano Nisoli, Edward D. Dahl, Gabriel Poulin-Lamarre, and Alejandro Lopez-Bezanilla. Qubit spin ice. Science, 373(6554):576–580, July 2021

    2021

  7. [15]

    Artificial ’spin ice’ in a geometrically frustrated lattice of nanoscale ferromagnetic islands

    1 RF Wang, Cristiano Nisoli, Rafael Sá de Freitas, J Li, W McConville, BJ Cooley, MS Lund, N Samarth, C Leighton, VH Crespi, et al. Artificial ’spin ice’ in a geometrically frustrated lattice of nanoscale ferromagnetic islands. Nature, 439(7074):303–306, 2006

    2006

  8. [16]

    A practical introduction to tensor networks: Matrix product states and projected entangled pair states

    Román Orús. A practical introduction to tensor networks: Matrix product states and projected entangled pair states. Annals of Physics, 349:117–158, 2014

    2014

  9. [17]

    Tensor net- works for quantum computing, 2025

    Aleksandr Berezutskii et al. Tensor net- works for quantum computing, 2025. arXiv:2503.08626

    2025

  10. [18]

    Relation betweentheandersonandkondohamiltonians

    John R Schrieffer and Peter A Wolff. Relation betweentheandersonandkondohamiltonians. Physical Review, 149(2):491, 1966

    1966

  11. [19]

    Per- turbative gadgets at arbitrary orders

    Stephen P Jordan and Edward Farhi. Per- turbative gadgets at arbitrary orders. Phys- ical Review A–Atomic, Molecular, and Optical Physics, 77(6):062329, 2008

    2008

  12. [20]

    Faehrmann, Sumeet Khatri, and Jens Eisert

    Simon Cichy, Paul K. Faehrmann, Sumeet Khatri, and Jens Eisert. Perturbative gadgets for gate-based quantum computing: Nonre- cursive constructionswithout subspacerestric- tions. Physical Review A, 109(5), May 2024. 14

    2024

  13. [21]

    Domain wall encoding of discrete variables for quantum annealing and QAOA

    Nicholas Chancellor. Domain wall encoding of discrete variables for quantum annealing and QAOA. Quantum Science and Technol- ogy, 4(4):045004, aug 2019

    2019

  14. [22]

    Quan- tum annealing correction with minor embed- ding

    Walter Vinci, Tameem Albash, Gerardo Paz Silva, Itay Hen, and Daniel Lidar. Quan- tum annealing correction with minor embed- ding. Physical Review A, 92, 07 2015

    2015

  15. [23]

    Effects of XX catalysts on quantum annealing spectra with perturbative crossings

    Natasha Feinstein, Louis Fry-Bouriaux, Sougato Bose, and PA Warburton. Effects of XX catalysts on quantum annealing spectra with perturbative crossings. Physical Review A, 110(4):042609, 2024

    2024

  16. [24]

    Gadgets for simulating a non-nativeXX inter- action in quantum annealing

    Robert J Banks, Natasha Feinstein, Roopayan Ghosh, Sougato Bose, and PA Warburton. Gadgets for simulating a non-nativeXX inter- action in quantum annealing. arXiv preprint arXiv:2503.16663, 2025

    2025 arXiv

  17. [25]

    Mohan Sarovar and G J. Milburn. Continuous quantum error correction by cooling.Physical Review A, 72, 07 2005

    2005

  18. [26]

    Jordan, Edward Farhi, and Pe- ter W

    Stephen P. Jordan, Edward Farhi, and Pe- ter W. Shor. Error-correcting codes for adi- abatic quantum computation. Phys. Rev. A, 74:052322, Nov 2006

    2006

  19. [27]

    Young, Mohan Sarovar, and Robin Blume-Kohout

    Kevin C. Young, Mohan Sarovar, and Robin Blume-Kohout. Error suppression and error correction in adiabatic quantum computation: Techniques and challenges.Physical Review X, 3, 07 2013

    2013

  20. [28]

    Er- ror suppression and error correction in adiabatic quantum computation: non- equilibrium dynamics

    Mohan Sarovar and Kevin C Young. Er- ror suppression and error correction in adiabatic quantum computation: non- equilibrium dynamics. New Journal of Physics, 15(12):125032, dec 2013. 15

    2013

Pith tools

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