Pith. sign in

REVIEW 2 major objections 5 minor 1 cited by

Heisenberg-limited Hamiltonian learning continuous variable systems via engineered dissipation

T0 review · 2 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read The paper establishes that engineered dissipation confines the dynamics of a bosonic Hamiltonian to a two-dimensional subspace per mode, allowing all coefficients to be recovered in total evolution time \(O(1/\epsilon)\) — the Heisenberg…

desk verdict Genuinely new adiabatic theorem and learning protocol, but the stated dissipation strength is wrong and Theorem 2.2 needs a constant fix. read the letter →

arxiv 2506.00606 v1 pith:5ZDPNN2G submitted 2025-05-31 quant-ph math-phmath.MP

classification quant-phmath-phmath.MP
keywords HamiltonianlearningbosonicsystemsHeisenberglimitengineereddissipationadiabaticapproximationcatcodescontinuous-variablequantumunboundedLindbladgenerators
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 tries to establish that Hamiltonian learning is possible for continuous-variable bosonic systems at the Heisenberg limit, the best scaling allowed by quantum metrology. It proposes adding a strong engineered dissipation, of the same family used to stabilize cat qubits, so that the unknown Hamiltonian's time evolution is effectively projected onto a small, controllable subspace. From there, phase-estimation experiments read out expectation values \(\langle\$\alpha$|H|\$\alpha$\rangle\), and polynomial interpolation plus Fourier inversion recovers every coefficient of a low-intersection Hamiltonian. The total time spent evolving under the unknown Hamiltonian is \(O((1/\epsilon)\log(m/\delta))\) for \(m\) modes, precision \(\epsilon\), and failure probability \(\delta\), matching the fundamental bound up to logarithmic factors.

What carries the argument

The load-bearing object is the modified photon-dissipation Lindbladian \(\mathcal{L} = \mathcal{L}[b(b-\$\alpha$)] + \mathcal{L}[b^r(b-\$\alpha$)]\), where \(\mathcal{L}[X] = X\cdot X^\dagger - \tfrac12\{X^\dagger X,\cdot\}\) and \(r=\lceil d/2\rceil-1\). Its kernel is exactly \(\operatorname{span}\{|0\rangle,|\$\alpha$\rangle\}\), and the proof establishes a spectral gap \(\eta \ge 1\) on the complement of this subspace. The adiabatic theorem (Theorem A.1) then controls the full dissipative-plus-Hamiltonian evolution in trace norm by the projected effective evolution, with constants \(C,C'\) that depend polynomially on \(d\) and \(|\$\alpha$|\). That bound is what makes strong dissipation act as a rigorous projector: it decouples clusters in multi-mode systems, diagonalizes the effective Hamiltonian within each cluster, and lets phase-estimation read out \(\langle\$\alpha$|H|\$\alpha$\rangle\).

What would settle it

The central claim can be settled by computing the smallest eigenvalue of \(L' = \sum_i (I-P)L_i^\dagger L_i(I-P)\) for \(\mathcal{L} = \mathcal{L}[b(b-\$\alpha$)] + \mathcal{L}[b^r(b-\$\alpha$)]\) on the complement of \(\operatorname{span}\{|0\rangle,|\$\alpha$\rangle\}\): if it is ever below 1, assumption (i) of the adiabatic theorem fails. A direct experimental check would simulate the single-mode protocol with a test Hamiltonian such as \(H = h_{2,0}(b^\dagger)^2 + h_{0,2}$b^{2}$ + h_{1,1}b^\dagger b\) at \(\$\alpha$ = A $e^{{i\theta}}$\), \(A = \Theta(\sqrt{\log(1/\epsilon)})\), \(\gamma = \Theta(\$epsilon^{{-1}}$\$log^{{2d+1/2}}$(1/\epsilon))\), and verify that the reconstructed coefficients satisfy the claimed \(\epsilon\)-accuracy.

Watch

Extended reading notes

Core claim

The paper's central claim is that low-intersection bosonic Hamiltonians—polynomials of bounded degree in creation and annihilation operators, with each term touching \(O(1)\) modes and each mode appearing in \(O(1)\) terms—can be learned with Heisenberg-limited scaling. The algorithm adds dissipation with jump operators \(L_{1,\$\alpha$}=b(b-\$\alpha$)\) and \(L_{r,\$\alpha$}=b^r(b-\$\alpha$)\) per mode, with strength \(\gamma=O($m^{2}$\$epsilon^{{-1}}$\$log^{{2d+1/2}}$(1/\epsilon))\), so the effective dynamics is confined to \(\operatorname{span}\{|0\rangle,|\$\alpha$\rangle\}\). In that subspace the effective Hamiltonian is approximately \(|\$\alpha$\rangle\langle\$\alpha$|H|\$\alpha$\rangle\langle\$\alpha$|\), and measurements give access to \($e^{{-i\langle\alpha|H|\alpha\rangle t}}$\); robust frequency estimation plus polynomial and Fourier interpolation reconstructs every coefficient to precision \(\epsilon\) with probability at least \(1-\delta\). The proof rests on a new quantitative adiabatic approximation for general Lindbladian evolutions with unbounded generators, giving the explicit trace-norm bound \(\|$e^{{t(\gamma\mathcal{L}}$+H)}(\rho)-$e^{{tH_{\mathrm{proj}}$}}(\rho)\|_1 \le tC/\gamma + C'/\gamma\).

Load-bearing premise

The whole argument rests on the engineered dissipation having a strictly positive spectral gap on the complement of the protected subspace and on the unknown Hamiltonian being relatively bounded by the dissipation generator with constants that grow only polynomially in \(|\$\alpha$|\), since if the gap collapsed or those constants blew up, the adiabatic projection error \(tC/\gamma + C'/\gamma\) would not stay small.

Editorial extensions

If this is right

  • For any low-intersection bosonic Hamiltonian of bounded degree, all coefficients can be learned to precision \(\epsilon\) with total Hamiltonian evolution time \(O((1/\epsilon)\log(m/\delta))\), so the number of modes enters only logarithmically.
  • The same engineered dissipation can stabilize a cat code, and the proof gives explicit trace-norm convergence to the code space plus an adiabatic limit for bounded-degree Hamiltonians, making rotations about the \(x\)-axis on the code space rigorous with explicit constants.
  • Strong dissipation regularizes the unbounded bosonic dynamics, preventing the exponential particle-number growth caused by squeezing and higher-order terms that would otherwise break Lieb-Robinson-type bounds.
  • The multi-mode protocol decouples overlapping interaction clusters by projecting boundary modes to vacuum, enabling parallel cluster-by-cluster learning with only \(O(\log^2(\log(1/\epsilon)/\epsilon)\log(m/\delta))\) experiments.

Reading between the lines

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

  • A natural next step, left open by the paper, is to remove the \(m^2\) factor in the dissipation strength: if the analysis is refined under energy-bounded dynamics, \(\gamma\) might become independent of \(m\) while retaining the same total evolution time.
  • The robust frequency-estimation subroutine, which converts phase estimation into a binary decision tree with majority voting, could likely be reused in classical continuous-variable learning tasks such as estimating the generator of a diffusion process from noisy observations.
  • The adiabatic approximation for unbounded Lindbladians is likely to transfer to other infinite-dimensional control problems, such as learning fermionic or optomechanical Hamiltonians, once an appropriate dissipative kernel with a known spectral gap is identified.
  • The explicit trace-norm convergence to the cat-code subspace may be useful for quantitative fault-tolerance analyses of cat-qubit architectures, since it upgrades earlier weighted convergence statements to full state-distance convergence.
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 / 5 minor

Summary. The paper studies Hamiltonian learning for continuous-variable bosonic systems governed by low-intersection Hamiltonians of bounded degree. It introduces a strong engineered-dissipation framework based on modified photon-loss jump operators L_{r,alpha}=b^r(b-alpha), whose kernel is span{|0>,|alpha>}, and proves a quantitative adiabatic theorem for unbounded Lindblad generators. This theorem is used to show that strong dissipation stabilizes an effective two-dimensional dynamics, from which the coefficients of the Hamiltonian can be extracted via robust frequency estimation. The claimed main result is a Heisenberg-limited learning algorithm with O((1/epsilon) log(m/delta)) total Hamiltonian evolution time and dissipation strength gamma = O(m^2 epsilon^{-1} log^{2d+1/2}(1/epsilon)). The same adiabatic framework is also applied to analyze the convergence of r-photon-driven dissipation in bosonic cat codes.

Significance. If the technical claims hold, this is a substantial contribution: it provides the first Heisenberg-limited Hamiltonian learning algorithm for a broad class of continuous-variable bosonic Hamiltonians, going beyond previous Bose-Hubbard-type restrictions. The new quantitative adiabatic theorem for Lindblad evolutions with unbounded generators is independently valuable and will likely be useful beyond the learning application, for example in the rigorous analysis of cat-qubit stabilization. The paper is unusually careful about unbounded operators and provides detailed appendices verifying spectral-gap and relative-boundedness assumptions. However, a specific exponent error in the derivation of the dissipation strength affects the formal statements of the main theorems; this is local and fixable but requires revision.

major comments (2)
  1. [Appendix A.3.1, Prop. A.4; Cor. C.5; Sec. 4.4; Theorems 2.2, 4.2, B.3, B.4] The stress-test concern about the displayed simplification in Prop. A.4 does not land: sqrt(c/(4r+4) ((4r+3)/(4r+4) 2c)^{4r+3}) equals 2^{(4r+3)/2} ((4r+3)/(4r+4))^{(4r+3)/2} c^{2r+2}/sqrt(4r+4), which is at most (2c)^{2(r+1)}/sqrt(4r+4). The actual problem is the next step. From Cor. C.5 with delta=1/2, the constant c is (r+2)(r+1)+16|alpha|(1+|alpha|^2)+8sqrt(2)|alpha|^2 = Theta(|alpha|^3), so the second constant in the relative bound is Theta(|alpha|^{6(r+1)}). With r=ceil(d/2)-1, this is Theta(|alpha|^{3d}) for even d and Theta(|alpha|^{3d+3}) for odd d, not O(|alpha|^{3d+2}) as stated. Consequently the constant C in Prop. 3.1 is Theta(|alpha|^{4d}) for even d and Theta(|alpha|^{4d+3}) for odd d, not O(|alpha|^{4d+2}). Section 4.4 and Theorems 2.2, 4.2, B.3, and B.4 choose gamma = O(t|alpha|^{4d+2}); for odd d this gives tC/gamma = Omega(|alpha|) = Omega(sqrt(log(1/epsilon))), so the adiabatic error is not bounded by a small constant and the condition |X(t)-cos(theta t)|<1/sqrt(8) in Theorem 4.1 is not guaranteed. The stated dissipation-strength claims are therefore not proven for odd d. The fix is local: choose gamma = O(t|alpha|^{4d+3}) and adjust the log exponents accordingly; the Heisenberg-limited total evolution time claim is unaffected.
  2. [Cor. C.5 / Lemma C.3 versus Prop. A.4 (r=ceil(d/2)-1)] For d=1 and d=2 the parameter r=ceil(d/2)-1 equals 0. The relative-boundedness result used for iota_2, Lemma C.3, is proved only for r>=1; its proof explicitly says that the case r<1 is dispatched by invoking Lemma C.2, but Lemma C.2 bounds powers of (N+I) by L=b^r-alpha^r and does not directly cover the jump operator L_{0,alpha}=b-alpha appearing in the dissipation. Thus the small-degree cases d=1,2 are not covered by the proof as written. A separate direct verification of the bound ||(N+I)|psi>|| <= a||L'|psi>||+b|||psi>|| for L'=L^dagger_{0,alpha}L_{0,alpha}+L^dagger_{1,alpha}L_{1,alpha} should be supplied. This is a completeness gap in the proof of Prop. 3.1, though it is local and does not affect the asymptotic structure for large d.
minor comments (5)
  1. [Sec. 4.4, Eq. (10)] The condition A_- = O(sqrt(log(1/epsilon))) is written as an upper bound, but the purpose is to make the exponential error O(|alpha|^d d^2 e^{-|alpha|^2/2} t) small; this requires a sufficiently large constant, i.e., A_- = Theta(sqrt(log(1/epsilon))) with an explicit constant threshold. The current wording is misleading.
  2. [Appendix B, first paragraph] There is a typo: 'ommitted' should be 'omitted'.
  3. [Lemma C.6 proof] In the final sentence of the proof, 'shoes' should be 'shows'; the proof also contains a few bracketed expressions that are hard to parse and would benefit from a clean rewrite.
  4. [Algorithm 1, ExpectVal] The line 'Evolve under -i[H,.] + gamma L' should clarify that this is the physical evolution to be implemented in the experiment, not a classical simulation, since H is the unknown Hamiltonian being learned.
  5. [Lemma D.2] The phrase 'd+1 th Chebyshev polynomial' should be '(d+1)-th Chebyshev polynomial'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the learning protocol is self-contained and the adiabatic bounds depend only on algorithm parameters, not on the target coefficients.

full rationale

The learning scheme extracts each coefficient from measured values of ⟨α|H|α⟩, which is a phase read out from the actual unknown-Hamiltonian evolution; no target coefficient is assumed in order to construct the estimator. The adiabatic approximation (Theorem A.1) is proved by a Duhamel expansion, and the constants C and C′ in Proposition A.4 are bounded using the degree d, the dissipation parameter α, and the uniform coefficient bound |h|≤1; they do not involve the specific values of hj,j′. Self-citations to [18], [20], [24], and [69] are used as technical lemmas (e.g., an elementary minimization lemma, robust frequency estimation), and the load-bearing robust-frequency-estimation theorem is re-proved in Appendix B.1. This makes those citations non-circular support rather than imported conclusions. The numerical inequality in Proposition A.4 flagged by a skeptical reader concerns the validity of an exponent reduction; even if it were an error, it is a correctness or scaling issue, not a reduction of the result to its inputs, so it does not constitute circularity.

Assumptions & free parameters 0 free parameters · 3 assumptions · 1 invented entities

The algorithm introduces no fitted parameters; all internal protocol parameters (e.g., A_-, A_+, gamma) are chosen from known bounds and do not depend on the unknown Hamiltonian. The central claim relies on standard semigroup theory and on the low-intersection Hamiltonian model. The modified photon dissipation is a new control object without independent experimental evidence for general r.

assumptions (3)
  • standard math Existence and uniqueness of quantum dynamical semigroups for unbounded Lindblad generators.
    Assumed in Section 3 (eq. 2) and Appendix A to define e^{tL}; the paper cites [64, Sec. 3.3] for the general theory.
  • domain assumption The unknown Hamiltonian is low-intersection with bounded degree d and coefficients bounded in absolute value by 1.
    Definition 2.1; this defines the problem class and is used throughout to bound constants C_d and C'_d independently of the coefficients.
  • domain assumption Experimental access to controlled displacements D(alpha) and to the engineered dissipation L_{r,alpha}=b^r(b-alpha).
    Used in the experimental protocol (Section 4.2 and Algorithm 1). The implementations are discussed in Section 5 and Appendix F but not demonstrated, especially for r>1.
invented entities (1)
  • Modified photon dissipation jump operator L_{r,alpha}=b^r(b-alpha)
    purpose: Projects the bosonic dynamics onto span{|0>,|alpha>} and regulates unbounded Hamiltonian evolution, enabling the learning protocol.
    New family of jump operators introduced in Section 2; for r=2 it resembles cat-code dissipation, and Appendix F sketches circuit-QED realizations, but no experimental demonstration is provided for general r. It is a control object, not a newly discovered physical entity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Heisenberg-limited Hamiltonian learning continuous variable systems via engineered dissipation." pith.science (2026). https://pith.science/paper/5ZDPNN2G

@misc{pith2026250600606,
  author       = {Pith},
  title        = {Pith review of: Heisenberg-limited Hamiltonian learning continuous variable systems via engineered dissipation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5ZDPNN2G}},
  note         = {Machine review of arXiv:2506.00606}
}
read the original abstract

Discrete and continuous variables oftentimes require different treatments in many learning tasks. Identifying the Hamiltonian governing the evolution of a quantum system is a fundamental task in quantum learning theory. While previous works mostly focused on quantum spin systems, where quantum states can be seen as superpositions of discrete bit-strings, relatively little is known about Hamiltonian learning for continuous-variable quantum systems. In this work we focus on learning the Hamiltonian of a bosonic quantum system, a common type of continuous-variable quantum system. This learning task involves an infinite-dimensional Hilbert space and unbounded operators, making mathematically rigorous treatments challenging. We introduce an analytic framework to study the effects of strong dissipation in such systems, enabling a rigorous analysis of cat qubit stabilization via engineered dissipation. This framework also supports the development of Heisenberg-limited algorithms for learning general bosonic Hamiltonians with higher-order terms of the creation and annihilation operators. Notably, our scheme requires a total Hamiltonian evolution time that scales only logarithmically with the number of modes and inversely with the precision of the reconstructed coefficients. On a theoretical level, we derive a new quantitative adiabatic approximation estimate for general Lindbladian evolutions with unbounded generators. Finally, we discuss possible experimental implementations.

Figures

Figures reproduced from arXiv: 2506.00606 by the authors.

Figure 1
Figure 1. (a) The algorithm for estimating coefficients [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗

Discussion (0). Sign in 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. Continuous Variable Hamiltonian Learning at Heisenberg Limit via Displacement-Random Unitary Transformation

    quant-ph 2025-10 conditional novelty 5.0 of 10

    D-RUT: displace-and-twirl each mode, measure one phase, interpolate and Fourier-invert to recover all finite-order bosonic Hamiltonian coefficients at the Heisenberg limit, with hierarchical multi-mode recovery and SP...

Reference graph

Works this paper leans on

76 extracted references · 62 canonical work pages · cited by 1 Pith paper

  1. [24]

    Heisenberg-limited Hamilto- nian learning for interacting bosons.npj Quantum Information, 10(1):83, 2024

    Haoya Li, Yu Tong, Tuvia Gefen, Hongkang Ni, and Lexing Ying. Heisenberg-limited Hamilto- nian learning for interacting bosons.npj Quantum Information, 10(1):83, 2024

  2. [18]

    Caro, Albert H Werner, and Cambyse Rouzé

    Tim Möbus, Andreas Bluhm, Matthias C. Caro, Albert H Werner, and Cambyse Rouzé. Dissipation-enabled bosonic Hamiltonian learning via new information-propagation bounds. arXiv preprint arXiv:2307.15026, 2023

  3. [1]

    Bypassing the noisy parity barrier: Learning higher-order Markov random fields from dynamics.arXiv preprint arXiv:2409.05284, 2024

    Jason Gaitonde, Ankur Moitra, and Elchanan Mossel. Bypassing the noisy parity barrier: Learning higher-order Markov random fields from dynamics.arXiv preprint arXiv:2409.05284, 2024

  4. [2]

    Compressed Sensing Measurement of Long-Range Correlated Noise

    Alireza Seif, Mohammad Hafezi, and Yi-Kai Liu. Compressed sensing measurement of long- range correlated noise.arXiv preprint arXiv:2105.12589, 2021

  5. [3]

    Evans, Robin Harper, and Steven T

    Tim J. Evans, Robin Harper, and Steven T. Flammia. Scalable Bayesian Hamiltonian learning. arXiv preprint arXiv:1912.07636, 2019

  6. [4]

    Zhi Li, Liujun Zou, and Timothy H. Hsieh. Hamiltonian tomography via quantum quench. Physical Review Letters, 124(16):160502, 2020

  7. [5]

    Learning quantum Hamiltonians from single-qubit measurements.Physical Review Research, 3(2):023246, 2021

    Liangyu Che, Chao Wei, Yulei Huang, Dafa Zhao, Shunzhong Xue, Xinfang Nie, Jun Li, Dawei Lu, and Tao Xin. Learning quantum Hamiltonians from single-qubit measurements.Physical Review Research, 3(2):023246, 2021

  8. [6]

    Optimal learning of quantum Hamiltonians from high-temperature Gibbs states

    Jeongwan Haah, Robin Kothari, and Ewin Tang. Optimal learning of quantum Hamiltonians from high-temperature Gibbs states. In2022 IEEE 63rd Annual Symposium on Foundations of Computer Science (FOCS), pages 135–146. IEEE, 2022

Show all 76 references
  1. [7]

    Robust and efficient Hamiltonian learning

    Wenjun Yu, Jinzhao Sun, Zeyao Han, and Xiao Yuan. Robust and efficient Hamiltonian learning. Quantum, 7:1045, 2023

  2. [8]

    Robustly learning the Hamiltonian dynamics of a superconducting quantum processor.Nature Communi- cations, 15(1):9595, 2024

    Dominik Hangleiter, Ingo Roth, Jonáš Fuksa, Jens Eisert, and Pedram Roushan. Robustly learning the Hamiltonian dynamics of a superconducting quantum processor.Nature Communi- cations, 15(1):9595, 2024

  3. [9]

    Markovich, V

    Daniel Stilck França, Liubov A. Markovich, V . V . Dobrovitski, Albert H. Werner, and Johannes Borregaard. Efficient and robust estimation of many-qubit Hamiltonians.Nature Communica- tions, 15(1):311, Jan 2024

  4. [10]

    Lindner, and Eyal Bairey

    Assaf Zubida, Elad Yitzhaki, Netanel H. Lindner, and Eyal Bairey. Optimal short-time measure- ments for Hamiltonian learning.arXiv preprint arXiv:2108.08824, 2021

  5. [11]

    Eyal Bairey, Itai Arad, and Netanel H. Lindner. Learning a local Hamiltonian from local measurements.Physical Review Letters, 122(2):020504, 2019

  6. [12]

    Lindner, and Itai Arad

    Eyal Bairey, Chu Guo, Dario Poletti, Netanel H. Lindner, and Itai Arad. Learning the dynamics of open quantum systems from their steady states.New Journal of Physics, 22(3):032001, 2020

  7. [13]

    Granade, Christopher Ferrie, Nathan Wiebe, and David G

    Christopher E. Granade, Christopher Ferrie, Nathan Wiebe, and David G. Cory. Robust online Hamiltonian learning.New Journal of Physics, 14(10):103013, 2012

  8. [14]

    Practical Hamiltonian learning with unitary dynamics and Gibbs states.Nature Communications, 15(1):312, 2024

    Andi Gu, Lukasz Cincio, and Patrick J Coles. Practical Hamiltonian learning with unitary dynamics and Gibbs states.Nature Communications, 15(1):312, 2024

  9. [15]

    Scalably learning quantum many-body Hamiltonians from dynamical data.arXiv preprint arXiv:2209.14328, 2022

    Frederik Wilde, Augustine Kshetrimayum, Ingo Roth, Dominik Hangleiter, Ryan Sweke, and Jens Eisert. Scalably learning quantum many-body Hamiltonians from dynamical data.arXiv preprint arXiv:2209.14328, 2022

  10. [16]

    Flammia, and Liang Jiang

    Stefan Krastanov, Sisi Zhou, Steven T. Flammia, and Liang Jiang. Stochastic estimation of dynamical variables.Quantum Science and Technology, 4(3):035003, 2019

  11. [17]

    Matthias C. Caro. Learning quantum processes and Hamiltonians via the Pauli transfer matrix. ACM Transactions on Quantum Computing, 5(2):1–53, June 2024

  12. [19]

    Milan Holzäpfel, Tillmann Baumgratz, Marcus Cramer, and Martin B. Plenio. Scalable recon- struction of unitary processes and Hamiltonians.Physical Review A, 91(4):042129, 2015. 11

  13. [20]

    Learning many-body Hamiltonians with Heisenberg-limited scaling.Physical Review Letters, 130(20):200403, 2023

    Hsin-Yuan Huang, Yu Tong, Di Fang, and Yuan Su. Learning many-body Hamiltonians with Heisenberg-limited scaling.Physical Review Letters, 130(20):200403, 2023

  14. [21]

    O’Brien, and Thomas Schuster

    Alicja Dutkiewicz, Thomas E. O’Brien, and Thomas Schuster. The advantage of quantum control in many-body Hamiltonian learning.Quantum, 8:1537, 2024

  15. [22]

    Learning interacting fermionic Hamiltonians at the Heisen- berg limit.Physical Review A, 110(6):062421, 2024

    Arjun Mirani and Patrick Hayden. Learning interacting fermionic Hamiltonians at the Heisen- berg limit.Physical Review A, 110(6):062421, 2024

  16. [23]

    Quantum Hamiltonian learning for the Fermi- Hubbard model.Acta Applicandae Mathematicae, 191(1):1–16, 2024

    Hongkang Ni, Haoya Li, and Lexing Ying. Quantum Hamiltonian learning for the Fermi- Hubbard model.Acta Applicandae Mathematicae, 191(1):1–16, 2024

  17. [25]

    Sergio Boixo and Rolando D. Somma. Parameter estimation with mixed-state quantum compu- tation.Physical Review A—Atomic, Molecular, and Optical Physics, 77(5):052320, 2008

  18. [26]

    Structure learning of Hamiltonians from real-time evolution

    Ainesh Bakshi, Allen Liu, Ankur Moitra, and Ewin Tang. Structure learning of Hamiltonians from real-time evolution. In2024 IEEE 65th Annual Symposium on Foundations of Computer Science (FOCS), pages 1037–1050. IEEE, 2024

  19. [27]

    Simulation-assisted learning of open quantum systems.Quantum, 8:1407, 2024

    Ke Wang and Xiantao Li. Simulation-assisted learning of open quantum systems.Quantum, 8:1407, 2024

  20. [28]

    Universal algorithm for transforming Hamiltonian eigenvalues.Physical Review Research, 7(1):013331, 2025

    Tatsuki Odake, Hlér Kristjánsson, Philip Taranto, and Mio Murao. Universal algorithm for transforming Hamiltonian eigenvalues.Physical Review Research, 7(1):013331, 2025

  21. [29]

    Hamiltonian learning at Heisenberg limit for hybrid quantum systems.arXiv preprint arXiv:2502.20373, 2025

    Lixing Zhang, Ze-Xun Lin, Prineha Narang, and Di Luo. Hamiltonian learning at Heisenberg limit for hybrid quantum systems.arXiv preprint arXiv:2502.20373, 2025

  22. [30]

    Flammia, John Preskill, and Yu Tong

    Muzhou Ma, Steven T. Flammia, John Preskill, and Yu Tong. Learningk-body Hamiltonians via compressed sensing.arXiv preprint arXiv:2410.18928, 2024

  23. [31]

    Flammia, and Susanne F

    Hong-Ye Hu, Muzhou Ma, Weiyuan Gong, Qi Ye, Yu Tong, Steven T. Flammia, and Susanne F. Yelin. Ansatz-free Hamiltonian learning with Heisenberg-limited scaling.arXiv preprint arXiv:2502.11900, 2025

  24. [32]

    Orlando, Simon Gustavsson, and William D

    Philip Krantz, Morten Kjaergaard, Fei Yan, Terry P. Orlando, Simon Gustavsson, and William D. Oliver. A quantum engineer’s guide to superconducting qubits.Applied Physics Reviews, 6(2), 2019

  25. [33]

    A. A. Clerk, K. W. Lehnert, P. Bertet, J. R. Petta, and Y . Nakamura. Hybrid quantum systems with circuit quantum electrodynamics.Nature Physics, 16(3):257–267, 2020

  26. [34]

    Grimsmo, Steven M

    Alexandre Blais, Arne L. Grimsmo, Steven M. Girvin, and Andreas Wallraff. Circuit quantum electrodynamics.Reviews of Modern Physics, 93(2):025005, 2021

  27. [35]

    Thompson

    Jianwei Wang, Fabio Sciarrino, Anthony Laing, and Mark G. Thompson. Integrated photonic quantum technologies.Nature Photonics, 14(5):273–284, 2020

  28. [36]

    Kippenberg, and Florian Marquardt

    Markus Aspelmeyer, Tobias J. Kippenberg, and Florian Marquardt. Cavity optomechanics. Reviews of Modern Physics, 86(4):1391, 2014

  29. [37]

    Applications of cavity optomechanics.Applied Physics Reviews, 1(3), 2014

    Michael Metcalfe. Applications of cavity optomechanics.Applied Physics Reviews, 1(3), 2014

  30. [38]

    Supersonic quantum communication.Physical Review Letters, 102(24):240501, 2009

    Jens Eisert and David Gross. Supersonic quantum communication.Physical Review Letters, 102(24):240501, 2009

  31. [39]

    Albert, Steven Touzard, Robert J

    Mazyar Mirrahimi, Zaki Leghtas, Victor V . Albert, Steven Touzard, Robert J. Schoelkopf, Liang Jiang, and Michel H. Devoret. Dynamically protected cat-qubits: a new paradigm for universal quantum computation.New Journal of Physics, 16(4), 2014

  32. [40]

    Rémi Azouit, Alain Sarlette, and Pierre Rouchon. Well-posedness and convergence of the Lindblad master equation for a quantum harmonic oscillator with multi-photon drive and damping.ESAIM: Control, Optimisation and Calculus of Variations, 22(4), 2016. 12

  33. [41]

    Repetition cat qubits for fault-tolerant quantum computation.Physical Review X, 9, 2019

    Jérémie Guillaud and Mazyar Mirrahimi. Repetition cat qubits for fault-tolerant quantum computation.Physical Review X, 9, 2019

  34. [42]

    Quantum computation with cat qubits.SciPost Physics Lecture Notes, 2023

    Jérémie Guillaud, Joachim Cohen, and Mazyar Mirrahimi. Quantum computation with cat qubits.SciPost Physics Lecture Notes, 2023

  35. [43]

    Coherent quantum dynamics in steady-state manifolds of strongly dissipative systems.Physical Review Letters, 113(24), 2014

    Paolo Zanardi and Lorenzo Campos Venuti. Coherent quantum dynamics in steady-state manifolds of strongly dissipative systems.Physical Review Letters, 113(24), 2014

  36. [44]

    Geometry, robustness, and emerging unitarity in dissipation-projected dynamics.Physical Review A, 91(5), 2015

    Paolo Zanardi and Lorenzo Campos Venuti. Geometry, robustness, and emerging unitarity in dissipation-projected dynamics.Physical Review A, 91(5), 2015

  37. [45]

    Quantum Zeno dynamics from general quantum operations.Quantum, 4:289, 2020

    Daniel Burgarth, Paolo Facchi, Hiromichi Nakazato, Saverio Pascazio, and Kazuya Yuasa. Quantum Zeno dynamics from general quantum operations.Quantum, 4:289, 2020

  38. [46]

    A note on approximation of operator semigroups.Archiv der Mathematik, 106(3):265–273, 2016

    Jochen Glück. A note on approximation of operator semigroups.Archiv der Mathematik, 106(3):265–273, 2016

  39. [47]

    Generalized quantum Zeno dynamics and ergodic means.arXiv preprint arXiv:1811.02509, 2018

    Norbert Barankai and Zoltán Zimborás. Generalized quantum Zeno dynamics and ergodic means.arXiv preprint arXiv:1811.02509, 2018

  40. [48]

    Quantitative quantum Zeno and strong damping limits in strong topology

    Robert Salzmann. Quantitative quantum Zeno and strong damping limits in strong topology. arXiv preprint arXiv:2409.06469, 2024

  41. [49]

    Touzard, A

    S. Touzard, A. Grimm, Z. Leghtas, S. O. Mundhada, P. Reinhold, C. Axline, M. Reagor, K. Chou, J. Blumoff, K. M. Sliwa, S. Shankar, L. Frunzio, R. J. Schoelkopf, M. Mirrahimi, and M. H. Devoret. Coherent oscillations inside a quantum manifold stabilized by dissipation. Physical...

  42. [50]

    On learning Markov chains.Advances in Neural Information Processing Systems, 31, 2018

    Yi Hao, Alon Orlitsky, and Venkatadheeraj Pichapati. On learning Markov chains.Advances in Neural Information Processing Systems, 31, 2018

  43. [51]

    Optimal prediction of Markov chains with and without spectral gap.Advances in Neural Information Processing Systems, 34:11233–11246, 2021

    Yanjun Han, Soham Jana, and Yihong Wu. Optimal prediction of Markov chains with and without spectral gap.Advances in Neural Information Processing Systems, 34:11233–11246, 2021

  44. [52]

    Minimax learning of ergodic Markov chains

    Geoffrey Wolfer and Aryeh Kontorovich. Minimax learning of ergodic Markov chains. In Algorithmic Learning Theory, pages 904–930. PMLR, 2019

  45. [53]

    Statistical estimation of ergodic Markov chain kernel over discrete state space.Bernoulli, 27(1):532–553, 2021

    Geoffrey Wolfer and Aryeh Kontorovich. Statistical estimation of ergodic Markov chain kernel over discrete state space.Bernoulli, 27(1):532–553, 2021

  46. [54]

    Deep unsuper- vised learning using nonequilibrium thermodynamics

    Jascha Sohl-Dickstein, Eric Weiss, Niru Maheswaranathan, and Surya Ganguli. Deep unsuper- vised learning using nonequilibrium thermodynamics. InInternational conference on machine learning, pages 2256–2265. pmlr, 2015

  47. [55]

    Denoising diffusion probabilistic models.Advances in neural information processing systems, 33:6840–6851, 2020

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models.Advances in neural information processing systems, 33:6840–6851, 2020

  48. [56]

    Generative modeling by estimating gradients of the data distribution.Advances in neural information processing systems, 32, 2019

    Yang Song and Stefano Ermon. Generative modeling by estimating gradients of the data distribution.Advances in neural information processing systems, 32, 2019

  49. [57]

    Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole

    Yang Song, Jascha Sohl-Dickstein, Diederik P. Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations.arXiv preprint arXiv:2011.13456, 2020

  50. [58]

    Discrete diffusion modeling by estimating the ratios of the data distribution.arXiv preprint arXiv:2310.16834, 2023

    Aaron Lou, Chenlin Meng, and Stefano Ermon. Discrete diffusion modeling by estimating the ratios of the data distribution.arXiv preprint arXiv:2310.16834, 2023

  51. [59]

    Rotskoff, and Lexing Ying

    Yinuo Ren, Haoxuan Chen, Grant M. Rotskoff, and Lexing Ying. How discrete and continuous diffusion meet: Comprehensive analysis of discrete diffusion models via a stochastic integral framework.arXiv preprint arXiv:2410.03601, 2024. 13

  52. [60]

    Rotskoff, Molei Tao, and Lexing Ying

    Yinuo Ren, Haoxuan Chen, Yuchen Zhu, Wei Guo, Yongxin Chen, Grant M. Rotskoff, Molei Tao, and Lexing Ying. Fast solvers for discrete diffusion models: Theory and applications of high-order algorithms.arXiv preprint arXiv:2502.00234, 2025

  53. [61]

    Exponential suppression of bit-flips in a qubit encoded in an oscillator.Nature Physics, 16(5):509–513, 2020

    Raphaël Lescanne, Marius Villiers, Théau Peronnin, Alain Sarlette, Matthieu Delbecq, Benjamin Huard, Takis Kontos, Mazyar Mirrahimi, and Zaki Leghtas. Exponential suppression of bit-flips in a qubit encoded in an oscillator.Nature Physics, 16(5):509–513, 2020

  54. [62]

    Campbell, Con- nor T

    Christopher Chamberland, Kyungjoo Noh, Patricio Arrangoiz-Arriola, Earl T. Campbell, Con- nor T. Hann, Joseph Iverson, Harald Putterman, Thomas C. Bohdanowicz, Steven T. Flammia, Andrew Keller, et al. Building a fault-tolerant quantum computer using concatenated cat codes. PRX...

  55. [63]

    American Mathematical Society, 2015

    Barry Simon.Operator theory, volume 4 ofA Comprehensive Course in Analysis. American Mathematical Society, 2015

  56. [64]

    Holevo.Statistical Structure of Quantum Theory

    Alexander S. Holevo.Statistical Structure of Quantum Theory. Springer, 2001

  57. [65]

    Vittorio Gorini, Andrzej Kossakowski, and E. C. George Sudarshan. Completely positive dynamical semigroups of n-level systems.Journal of Mathematical Physics, 17(5):821–825, 1976

  58. [66]

    On the generators of quantum dynamical semigroups.Communications in Mathematical Physics, 48(2):119–130, 1976

    Göran Lindblad. On the generators of quantum dynamical semigroups.Communications in Mathematical Physics, 48(2):119–130, 1976

  59. [67]

    Springer, 2012

    Konrad Schmüdgen.Unbounded self-adjoint operators on Hilbert space, volume 265 of Graduate Texts in Mathematics. Springer, 2012

  60. [68]

    Holevo.Quantum Systems, Channels, Information: A Mathematical Introduction

    Alexander S. Holevo.Quantum Systems, Channels, Information: A Mathematical Introduction. De Gruyter, 2012

  61. [69]

    Energy preserving evolutions over bosonic systems.Quantum, 8:1551, 2024

    Paul Gondolf, Tim Möbus, and Cambyse Rouzé. Energy preserving evolutions over bosonic systems.Quantum, 8:1551, 2024

  62. [70]

    Contributions to the theory of generalized inverses

    Adi Ben-Israel and Abraham Charnes. Contributions to the theory of generalized inverses. Journal of the Society for Industrial and Applied Mathematics, 11(3):667–699, 1963

  63. [71]

    Y . Y . Tseng. Generalized inverses of unbounded operators between two unitary spaces. In Doklady Akademii Nauk SSSR, volume 67, pages 431–434, 1949

  64. [72]

    Lawrence J. Lardy. A series representation for the generalized inverse of a closed linear operator. Atti della Accademia Nazionale dei Lincei. Classe di Scienze Fisiche, Matematiche e Naturali. Rendiconti, 59:152–157, 1975

  65. [73]

    Academic press, 1981

    Michael Reed and Barry Simon.Functional analysis, volume 1 ofMethods of Modern Mathe- matical Physics. Academic press, 1981

  66. [74]

    Springer, 1966

    Tosio Kato.Perturbation theory for linear operators, volume 132 ofDie Grundlagen der mathematischen Wissenschaften in Einzeldarstellungen. Springer, 1966

  67. [75]

    Pedersen.Analysis now, volume 118 ofGraduate Texts in Mathematics

    Gert K. Pedersen.Analysis now, volume 118 ofGraduate Texts in Mathematics. Springer, 2nd edition, 1995. 14 A Adiabatic evolutions A.1 Adiabatic approximation for general Lindbladian evolutions with unbounded generators In this appendix, we will prove our main technical tool, t...

  68. [76]

    beam-splitter

    · · ·e−iθ(jk−j′ k). From the above we can see that ⟨α|H|α⟩ is a polynomial of A1, A2,· · ·, Aq, whose coefficients hl1···lk (θ) are Fourier sums of the anglesθ1,· · ·, θk. Therefore we first use polynomial interpolation to estimate the polynomial coefficientshl1···lk (θ) for a...

Pith tools

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