Pith. sign in

REVIEW 3 major objections 4 minor 50 references

This paper claims that a post-selection circuit using one or two ancillas and two Clifford gates filters out the amplitude-damping jump branch, purifying both quantum states and quantum channels.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

Postselecting on the no-jump measurement in a one- or two-ancilla circuit improves state and channel fidelity under amplitude-damping noise, leaving a residual amplitude attenuation.

T0 review reviewed 2026-08-05 challenge →

load-bearing objection A correct but standard no-jump filter for AD states, with a flawed extension to channel purification that needs major revision. the 3 major comments →

arxiv 2509.05709 v1 pith:CMBON5HG submitted 2025-09-06 quant-ph

Quantum Purification for Amplitude Damping Noise

Kai Wang , Zhen-Yang Peng This is my paper
classification quant-ph
keywords amplitude damping noisequantum state purificationchannel purificationpost-selectionChoi–Jamiolkowski isomorphismClifford gatesfidelitynoise probing
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The reading

This paper proposes a post-selection protocol that purifies a single copy of a quantum state—or, via the Choi state, a quantum channel—corrupted by amplitude-damping (AD) noise. The circuit needs one or two ancilla qubits and only Hadamard and controlled-Z gates. Because AD's Kraus operators either commute or anticommute with Z, measuring the ancilla and keeping the |0> outcome discards the E1 "jump" branch and leaves the attenuated no-jump branch E0|ψ>, giving the closed-form fidelity of Eq. (14). The same construction applied to a Choi state suppresses the dominant channel error at second order in the damping probability. The authors support the protocol with numerical fidelities and success probabilities, arguing it is practical for near-term devices because it avoids multi-copy purification and keeps the success probability above 0.5.

Core claim

The central claim is that amplitude-damping noise can be filtered at the level of a single noisy copy. Prepare an ancilla in |+>, apply a controlled-Z between it and the data qubit, and measure the ancilla; the |0> outcome corresponds to the no-jump branch E0|ψ>, while the jump branch E1|ψ> is routed to the discarded |1> outcome. For an input α|0>+β|1>, the post-selected state is α|0>+√(1−γ)β|1>, with fidelity (α²+√(1−γ)β²)²/(α²+(1−γ)β²) and success probability α²+(1−γ)β². For channels, the same filter on a Choi state keeps the E0⊗E0 branch and suppresses the leading E1⊗E1 error with probability γ². Numerical sampling shows single-qubit fidelity above 0.99 for γ up to 0.4184 with success pro

What carries the argument

The key object is a Z-parity filter built from AD's commutation relations E0Z=ZE0 and E1Z=−ZE1. The circuit entangles the ancilla with the data qubit so that the E0 branch picks up no relative phase and the E1 branch picks up a π phase; after a final Hadamard, that phase rotates the E1 branch into the |1> ancilla outcome. Post-selecting |0> then realizes the non-unitary map E0, which attenuates the |1> amplitude by √(1−γ). For channels, the Choi state is damped on both sides, so the kept branch is E0⊗E0 and the leading discarded error is E1⊗E1 with probability γ².

Load-bearing premise

The scheme rests on the assumption that the only error is amplitude damping on the data qubit(s), that the ancilla and controlled-Z gate are perfect, and that the input is a pure state—any additional noise or impurity leaks into the post-selected branch and breaks Eq. (14).

What would settle it

Prepare a known state α|0>+β|1>, run the circuit with a noisy ancilla or CZ gate, and measure the excited-state population of the post-selected output. Eq. (13) predicts it equals (1−γ)|β|²/(α²+(1−γ)β²); observing a different population, or any admixture of the E1 branch in the |0> outcome, would falsify the claim that the circuit filters AD noise.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • For a single damped qubit, post-selecting the |0> ancilla outcome yields the state α|0>+√(1−γ)β|1> with fidelity given by Eq. (14); the protocol keeps fidelity above 0.99 up to γ=0.4184.
  • For an AD-affected channel, applying the same filter to its Choi state suppresses the dominant error branch E1⊗E1, whose probability is O(γ²), so the protected channel stays close to ideal for small γ.
  • With two ancillas, two-qubit states and channels are purified with fidelity above 0.95 up to γ=0.5, at the cost of a lower success probability—about 70.9% at the γ=0.2667 threshold where fidelity stays above 0.99.
  • The circuit doubles as a noise probe: preparing the data qubit in |1> makes the ancilla "1" outcome occur with probability exactly γ, giving a direct estimate of the damping strength.
  • Because only Hadamard and controlled-Z gates are used, the protocol is compatible with current near-term hardware and can be applied to a single noisy copy, sidestepping the no-cloning restrictions that limit multi-copy purification.

Where Pith is reading between the lines

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

  • The parity-filter idea is not tied to AD specifically: any noise channel whose Kraus operators split into Z-commuting and Z-anticommuting parts would be filtered the same way, so the protocol may transfer to mixtures of dephasing and damping; the authors do not claim this.
  • With γ estimated from the probe mode, the attenuation in Eq. (13) can be inverted via the operator in Eq. (17) to restore the original amplitudes; this turns the purification into a near-deterministic AD-error correction whenever the damping rate is known.
  • For channel purification, the Choi-state version effectively renormalizes the noisy channel; a natural follow-up is to concatenate it with gate teleportation to produce higher-fidelity logical gates, which the paper leaves implicit.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The manuscript proposes a one- or two-ancilla circuit (Hadamard and CZ gates) that postselects the no-jump branch of amplitude-damping noise. For a single-qubit pure state, the circuit conditions the data qubit into the E0 branch, giving the analytical success probability and fidelity in Eqs. (12)-(14). The authors also propose a parameter-estimation scheme for γ and an optional compensation operation in Eq. (17). They then extend the same idea to channel purification via Choi states, claiming that the circuit purifies an amplitude-damped channel, and present numerical fidelities and success probabilities for state, two-qubit, and channel cases.

Significance. If the claims were fully correct, the paper would offer a strikingly simple, low-overhead purification protocol for AD noise: single-copy operation, only two Clifford gate types, explicit analytical fidelity formulas, and a built-in noise-strength probe. The state-purification part is largely sound and transparent; Eqs. (12)-(14) follow directly from the circuit and the AD Kraus operators, with no fitted parameters. However, the channel-purification extension, which is a central claim in the title, abstract, introduction, and numerical section, is not valid as presented. The postselected operation is trace-decreasing and after renormalization is not a CPTP map, so the output is not a quantum channel and the reported channel fidelity is undefined. The optional compensation operator in Eq. (17) is also non-unitary and unphysical as a deterministic gate. The paper therefore needs substantial revision before the channel claims can be accepted.

major comments (3)
  1. [Section IV, Eqs. (20)-(21)] The derivation of channel purification is founded on an incorrect use of the Choi isomorphism. Eq. (20) is a density matrix with an incoherent sum over the Kraus index α, but Eq. (21) replaces it by a pure-state superposition |ξ> over α and m. For a non-unitary channel ε the Choi state is mixed; only unitary (isometric) channels have pure Choi states. The statement in Sec. I that 'any quantum channel can be mathematically represented as a pure state' is therefore false. All subsequent simplifications, including Eqs. (28)-(29), inherit this error, and the conclusion that postselecting |0> realizes channel purification is not justified.
  2. [Section IV, Fig. 5] Even if the algebra of Eq. (28) were correct, the postselected operation is trace-decreasing and its normalized form depends nonlinearly on the input state, so it does not define a CPTP map and hence does not define a quantum channel. Fig. 5 reports a 'channel fidelity' but no definition of channel fidelity is given. A concrete counterexample is ε=I: the dominant postselected term is proportional to |00>+(1−γ)|11> after normalization, whose reduced state is not I/2, so it cannot be the Choi state of any channel. Thus the channel-purification claim and Fig. 5 are unsupported.
  3. [Section III, Eq. (17)] The proposed compensation operator diag(1, 1/√(1−γ)) is non-unitary and has operator norm >1, so it cannot be a deterministic quantum gate. It also violates the trace-nonincreasing condition K†K ≤ I for a physical Kraus operator. This step is not realizable under the paper's stated resource assumptions, and if interpreted as a probabilistic filter it would require a success probability greater than 1 for the |1> input. The 'further purification' claim should be removed or replaced by a proper physical implementation.
minor comments (4)
  1. [Section I] The sentence 'any quantum channel can be mathematically represented as a pure state' should be corrected; the Choi state of a general channel is mixed.
  2. [Section IV, Eqs. (21)-(29)] The notation is very hard to follow: the indices p_i^m, q_α^m, p_j^{αm} are not defined clearly, and the transition from the mixture in Eq. (20) to a pure-state notation in Eq. (21) should be explained or avoided.
  3. [General] The figures are not visible in the manuscript text provided; please ensure they are included in the submission.
  4. [References] Reference [2] has a typo: 'Accesible via' should be 'Accessible via'. The reference list would also benefit from a discussion of no-jump/quantum-Zeno error suppression methods, which are closely related to the state-purification idea.

Circularity Check

0 steps flagged

No significant circularity: the state-purification fidelities are direct circuit computations; the channel-extension caveat is a correctness issue, not a circular reduction.

full rationale

The derivation chain is self-contained. In Sec. III, Eqs. (5)-(14) explicitly compute the action of the H-CZ-N-CZ-H circuit on an arbitrary pure state. The postselected |0> branch reduces to E0|ψ> because of the parity relations [E0,Z]=0 and {E1,Z}=0 (Eq. 9), giving the measurement probability p0 (Eq. 12) and fidelity (Eq. 14) directly from the circuit and the AD Kraus operators. No parameter is fitted, no external benchmark is invoked, and the 'purification' claim is not definitional: the target is the original unknown state |ψ>, and the fidelity is computed against that state, not assumed. The channel-purification extension in Sec. IV repeats the same circuit calculation on a Choi state (Eqs. 20-29). There is a genuine correctness concern that the Choi state of a non-unitary channel is generally mixed and that the postselected E0-filter is trace-decreasing, so the normalized output is not generally a CPTP map; however, that is a validity flaw in the channel-extension claim, not a circularity. The derivation of the postselected expressions follows from the circuit and the Kraus operators rather than being equivalent to the conclusion by construction. There is no load-bearing self-citation, no fitted-input-called-prediction pattern, and no imported uniqueness theorem. The only mild conceptual point is that the protocol achieves purification by selecting the no-jump branch, which is the intended mechanism of postselection-based purification and is not circular.

Axiom & Free-Parameter Ledger

0 free parameters · 5 axioms · 0 invented entities

No free parameters are fitted: the protocol is derived analytically from the AD Kraus operators. The optional compensation step in Eq. (17) requires estimating gamma, but that is an input from the noise model, not a fitted parameter of the protocol. No new physical entities are introduced.

axioms (5)
  • domain assumption The amplitude-damping channel is CPTP with Kraus operators E0 and E1 as in Eq. (2), satisfying [E0,Z]=0 and {E1,Z}=0.
    The parity filter in Eqs. (7)-(8) depends on these commutation relations; if the noise model deviates, the |0> branch is not exactly E0|psi>.
  • domain assumption The ancilla and the controlled-Z gate are noiseless and the measurement is perfect.
    The derivation of Eqs. (5)-(11) assumes ideal operations; noise on the ancilla or CZ gate would spoil the cancellation.
  • domain assumption Input states are pure and unknown; purification targets a pure state.
    The paper states purification applies exclusively to pure states (Sec. I); the fidelity formulas assume |psi> is pure.
  • standard math Choi-Jamiolkowski isomorphism for channel purification.
    Used in Sec. IV to convert channel purification into state purification.
  • standard math Born rule and standard projective measurement.
    Used to compute probabilities p0 and p1 in Eqs. (12), (15)-(16).

reviewed 2026-08-05 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Quantum Purification for Amplitude Damping Noise." pith.science (2026). https://pith.science/paper/CMBON5HG

@misc{pith2026250905709,
  author       = {Pith},
  title        = {Pith review of: Quantum Purification for Amplitude Damping Noise},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CMBON5HG}},
  note         = {Machine review of arXiv:2509.05709}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Noise poses a fundamental challenge to quantum information processing, with amplitude-damping (AD) noise being particularly detrimental. Preserving high-fidelity quantum systems therefore relies critically on effective error correction and purification methods. In this work, we introduce a novel approach for mitigating AD noise that can be applied to both state and channel purification. Our method achieves a substantial enhancement in the fidelity of affected states or channels while maintaining a low resource overhead, requiring only one or two ancilla qubits in combination with two Clifford gates, and exhibits a relatively high success probability. This approach provides a practical and scalable framework for addressing AD noise in realistic quantum systems.

Figures

Figures reproduced from arXiv: 2509.05709 by Kai Wang, Zhen-Yang Peng.

Figure 1
Figure 1. Figure 1: FIG. 1. (Color online) Circuit for state purification under AD [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. (Color online) Circuit for channel purification under [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. (Color online) Circuit for two-qubit state purification [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. (Color online) Fidelity and probability as a function [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. (Color online) Fidelity and probability as a func [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗

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

50 extracted references · 36 canonical work pages · 1 internal anchor

  1. [1]

    M. A. Nielsen and I. L. Chuang,Quantum Computa- tion and Quantum Information: 10th Anniversary Edi- tion(Cambridge University Press, 2010)

    2010

  2. [2]

    Preskill, Accesible via http://www

    J. Preskill, Accesible via http://www. theory. caltech. edu/people/preskill/ph2291997(2015)

    work page 2015

  3. [3]

    Kashif, E

    M. Kashif, E. Sychiuco, and M. Shafique, 2024 Interna- tional Joint Conference on Neural Networks (IJCNN) , 1 (2024)

    work page 2024

  4. [4]

    D. A. Lidar and T. A. Brun,Quantum error correction (Cambridge university press, 2013)

    work page 2013

  5. [5]

    P. W. Shor, Phys. Rev. A52, R2493 (1995)

    1995

  6. [6]

    Steane, Proceedings of the Royal Society of London

    A. Steane, Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sci- ences452, 2551 (1996)

    1996

  7. [7]

    Laflamme, C

    R. Laflamme, C. Miquel, J. P. Paz, and W. H. Zurek, Phys. Rev. Lett.77, 198 (1996)

    1996

  8. [8]

    Gottesman, arXiv: Quantum Physics (1997)

    D. Gottesman, arXiv: Quantum Physics (1997)

    work page 1997

  9. [9]

    Georgescu, Nature Reviews Physics2, 519 (2020)

    I. Georgescu, Nature Reviews Physics2, 519 (2020)

    work page 2020

  10. [10]

    Gidney, M

    C. Gidney, M. Newman, P. Brooks, and C. Jones, Nature Communications16, 4498 (2025)

    2025

  11. [11]

    Duclos-Cianci and D

    G. Duclos-Cianci and D. Poulin, Physical review letters 104, 050504 (2010)

    work page 2010

  12. [12]

    Bomb ´ ın, Physical Review A—Atomic, Molecular, and Optical Physics81, 032301 (2010)

    H. Bomb ´ ın, Physical Review A—Atomic, Molecular, and Optical Physics81, 032301 (2010)

    work page 2010

  13. [13]

    Kubica, B

    A. Kubica, B. Yoshida, and F. Pastawski, New Journal of Physics17, 083026 (2015)

    2015

  14. [14]

    A. G. Fowler, Physical Review A—Atomic, Molecular, and Optical Physics83, 042310 (2011)

    work page 2011

  15. [15]

    Postler, F

    L. Postler, F. Butt, I. Pogorelov, C. D. Marciniak, S. Heußen, R. Blatt, P. Schindler, M. Rispler, M. M¨ uller, and T. Monz, PRX Quantum5, 030326 (2024)

    2024

  16. [16]

    Zheng, W

    G. Zheng, W. He, G. Lee, and L. Jiang, Phys. Rev. Lett. 132, 250602 (2024)

    2024

  17. [17]

    I. L. Chuang, D. W. Leung, and Y. Yamamoto, Phys. Rev. A56, 1114 (1997)

    work page 1997

  18. [18]

    A. S. Fletcher, P. W. Shor, and M. Z. Win, IEEE Trans- actions on Information Theory54, 5705 (2008)

    work page 2008

  19. [19]

    R. Duan, M. Grassl, Z. Ji, and B. Zeng, in2010 IEEE International Symposium on Information Theory(2010) pp. 2672–2676

    work page 2010

  20. [20]

    P. W. Shor, G. Smith, J. A. Smolin, and B. Zeng, IEEE Transactions on Information Theory57, 7180 (2011)

    work page 2011

  21. [21]

    M. F. Ezerman and M. Grassl, 2013 IEEE International Symposium on Information Theory , 922 (2013)

    work page 2013

  22. [22]

    Cafaro and P

    C. Cafaro and P. van Loock, Phys. Rev. A89, 022316 (2014)

    work page 2014

  23. [23]

    Ouyang and R

    Y. Ouyang and R. Chao, IEEE Transactions on Informa- tion Theory66, 2921 (2020)

    work page 2020

  24. [24]

    J. I. Cirac, A. K. Ekert, and C. Macchiavello, Phys. Rev. Lett.82, 4344 (1999)

    work page 1999

  25. [25]

    Banaszek and I

    K. Banaszek and I. Devetak, Phys. Rev. A64, 052307 (2001)

    work page 2001

  26. [26]

    Zhao and G

    X. Zhao and G. Chiribella, Phys. Rev. A95, 042303 (2017)

    work page 2017

  27. [27]

    M. J. Gullans and D. A. Huse, Phys. Rev. X10, 041020 (2020)

    2020

  28. [28]

    Z.-L. Cao, M. Yang, and G.-C. Guo, Physics Letters A 308, 349 (2003)

    work page 2003

  29. [29]

    Fang and Z.-W

    K. Fang and Z.-W. Liu, PRX Quantum3, 010337 (2022)

    2022

  30. [30]

    S. Das, J. Sun, M. Hanks, B. Koczor, and M. Kim, arXiv preprint arXiv:2407.20173 (2024)

    work page arXiv 2024

  31. [31]

    Tsubouchi, Y

    K. Tsubouchi, Y. Mitsuhashi, R. Takagi, and N. Yosh- ioka, arXiv preprint arXiv:2503.13114 (2025)

    work page arXiv 2025

  32. [32]

    Z. Liu, X. Zhang, Y.-Y. Fei, and Z. Cai, PRX Quantum 6, 020325 (2025)

    work page 2025

  33. [33]

    Yamamoto, S

    K. Yamamoto, S. Endo, H. Hakoshima, Y. Matsuzaki, and Y. Tokunaga, Physical Review Letters129, 250503 (2022)

    work page 2022

  34. [34]

    H. Kwon, C. Oh, Y. Lim, H. Jeong, and L. Jiang, Phys- ical Review A109, 022410 (2024)

    work page 2024

  35. [35]

    X. Yuan, B. Regula, R. Takagi, and M. Gu, Physical Review Letters132, 050203 (2024)

    work page 2024

  36. [36]

    Wang and D.-S

    K. Wang and D.-S. Wang, New Journal of Physics25, 043013 (2023)

    work page 2023

  37. [37]

    Verstraete, J

    F. Verstraete, J. Dehaene, and B. DeMoor, Physical Re- view A64, 010101 (2001)

    work page 2001

  38. [38]

    Ecker, F

    S. Ecker, F. Bouchard, L. Bulla, F. Brandt, O. Kohout, F. Steinlechner, R. Fickler, M. Malik, Y. Guryanova, R. Ursin,et al., Physical Review X9, 041042 (2019)

    work page 2019

  39. [39]

    L. Chen, K. Xue, J. Li, Z. Li, R. Li, N. Yu, Q. Sun, and J. Lu, IEEE Journal on Selected Areas in Communica- tions42, 1723 (2024)

    work page 2024

  40. [40]

    Cozzolino, B

    D. Cozzolino, B. Da Lio, D. Bacco, and L. K. Oxenløwe, Advanced Quantum Technologies2, 1900038 (2019)

    work page 2019

  41. [41]

    Montanaro, npj Quantum Information2, 1 (2016)

    A. Montanaro, npj Quantum Information2, 1 (2016)

    work page 2016

  42. [42]

    Cerezo, A

    M. Cerezo, A. Arrasmith, R. Babbush, S. C. Benjamin, S. Endo, K. Fujii, J. R. McClean, K. Mitarai, X. Yuan, L. Cincio,et al., Nature Reviews Physics3, 625 (2021)

    2021

  43. [43]

    Jamio lkowski, Reports on Mathematical Physics3, 275 (1972)

    A. Jamio lkowski, Reports on Mathematical Physics3, 275 (1972)

    work page 1972

  44. [44]

    Choi, Linear Algebra and its Applications10, 285 (1975)

    M.-D. Choi, Linear Algebra and its Applications10, 285 (1975)

    1975

  45. [45]

    W. K. Wootters and W. H. Zurek, Physics Today62, 76 (2009)

    work page 2009

  46. [46]

    Kraus, A

    K. Kraus, A. B¨ ohm, J. D. Dollard, and W. Wootters, States, effects, and operations fundamental notions of quantum theory: Lectures in mathematical physics at the university of Texas at Austin(Springer, 1983)

    work page 1983

  47. [47]

    A. S. Fletcher, arXiv preprint arXiv:0706.3400 (2007)

    work page internal anchor Pith review Pith/arXiv arXiv 2007

  48. [48]

    Buscemi, G

    F. Buscemi, G. Chiribella, and G. Mauro D’Ariano, Phys. Rev. Lett.95, 090501 (2005)

    work page 2005

  49. [49]

    J. S. Bell, Physics Physique Fizika1, 195 (1964)

    1964

  50. [50]

    Zx-calculus for the working quan- tum computer scientist,

    J. van de Wetering, “Zx-calculus for the working quan- tum computer scientist,” (2020), arXiv:2012.13966 [quant-ph]

    Pith/arXiv arXiv 2020

This paper was first reviewed by deepseek-v4-flash on August 5, 2026.