Pith. sign in

REVIEW 5 major objections 6 minor 2 cited by

Differential Privacy Preserving Distributed Quantum Computing

T0 review · 5 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Quantum Rényi differential privacy extends Rényi DP to distributed quantum computing via measurement-outcome distributions.

desk verdict QRDP is classical RDP on measurement outcomes; the composability claim for every POVM is unproven and the post-processing proof reverses the data-processing inequality. read the letter →

arxiv 2412.12387 v2 pith:HC2N6L4V submitted 2024-12-16 quant-ph cs.DC

Hui Zhong , Keyi Ju , Jiachen Shen , Xinyue Zhang , Xiaoqi Qin , Tomoaki Ohtsuki , Miao Pan , Zhu Han This is my paper · ORCID
classification quant-phcs.DC MSC 68P2781P6894A17 PACS 03.67.-a03.67.Lx
keywords QuantumdistributedcomputingnoisesdifferentialprivacyRényidivergencebudgetcompositionFidelity
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 proposes quantum Rényi differential privacy (QRDP), a privacy definition for distributed quantum computing modeled on classical Rényi differential privacy. The central idea is to measure privacy loss by the Rényi divergence between the probability distributions obtained by measuring two neighboring quantum states after the same quantum operation, rather than by comparing the states directly. QRDP inherits the properties that make Rényi privacy attractive in distributed settings—post-processing immunity and composition—so the total privacy budget of a multi-round distributed computation is the sum of the per-round budgets. For three noise mechanisms (generalized amplitude damping, phase plus amplitude damping, and depolarizing noise), the paper derives privacy budgets and a simplified formula $\hat{\epsilon}(\alpha)$ that depends only on noise parameters and the trace-distance bound, not on the input state. Numerical simulations illustrate the expected tradeoff: stronger noise lowers the privacy budget but also lowers fidelity, meaning weaker data utility.

What carries the argument

The load-bearing object is the quantum Rényi divergence defined in Eq. (5): for a POVM (a set of positive matrices summing to the identity) $M=\{M_m\}$ and two neighboring states $\rho,\sigma$, it is $D_\alpha(E_m(\rho)\|E_m(\sigma))=\frac{1}{\alpha-1}\log\sum_m \frac{(\operatorname{Tr}(M_m\rho))^\alpha}{(\operatorname{Tr}(M_m\sigma))^{\alpha-1}}$. This converts quantum states into classical probability distributions at the measurement stage, which lets the paper import the classical Rényi DP toolbox—post-processing, basic and adaptive composition, and the conversion to $(\epsilon,\delta)$-DP. Theorem 4 is built by substituting the QDP budgets of the three noise channels into the tight classical RDP-conversion bound of [35], so the noise parameters alone control the QRDP budget.

What would settle it

Choose neighboring single-qubit states with trace distance $d=0.1$, depolarizing noise with $p=0.5$, and numerically maximize $D_2(E_m(\rho)\|E_m(\sigma))$ over all projective measurements; if the maximum exceeds $\hat{\epsilon}(2)$ from Theorem 4 (with $\epsilon=\ln(1+\frac{1-p}{p}dD)$, $D=2$), the claimed universal QRDP bound fails. This is directly checkable because the channels and formula are explicit.

Watch

Extended reading notes

Core claim

The central claim is that classical Rényi differential privacy can be lifted to quantum distributed computing by defining privacy through a new quantum Rényi divergence on measurement-outcome distributions. The key quantitative result is Theorem 4: if a noise channel is known to satisfy QDP with budget $\epsilon$ for neighboring states with trace distance at most $d$, then the same channel satisfies $(\alpha,\hat{\epsilon}(\alpha))$-QRDP with $\hat{\epsilon}(\alpha)=\epsilon-\frac{1}{\alpha-1}\log\frac{1+e^{-\epsilon}}{1+e^{-(2\alpha-1)\epsilon}}$, where $\epsilon$ is the GAD, PAD, or depolarizing QDP budget. The bound is tight in the classical case, so QRDP gives a principled $\alpha$-tunable budget that interpolates between $\epsilon$-QDP and $(\epsilon,\delta)$-QDP. Composition theorems make the total budget additive across the quantum operations that appear in a distributed computation, which is the paper's main argument for why QRDP suits QDC.

Load-bearing premise

The whole scheme rests on treating the output of a quantum computation as a classical probability distribution over measurement results; if an attacker can extract information from the quantum state itself, for example by choosing a different measurement or exploiting quantum correlations invisible in the measured outcomes, the QRDP guarantee may not cover that attack.

Editorial extensions

If this is right

  • For a distributed quantum computation split across QPUs, each round's QRDP budget can be summed under the composition theorems, giving a total budget that grows linearly in the number of rounds rather than exploding combinatorially.
  • The parameter $\alpha$ lets a user interpolate between stricter and looser interpretations: as $\alpha\to\infty$, QRDP reduces to QDP, while finite $\alpha$ gives smaller budgets that are easier to satisfy.
  • Because Theorem 4 is independent of the input state and measurement, privacy budgeting can be done at design time from the noise parameters and the trace-distance bound $d$ alone.
  • The same noise that degrades a quantum computation—amplitude damping, phase damping, depolarization—can be counted as a privacy mechanism, so QRDP can be implemented with inherent or tunable physical noise.
  • The conversion to $(\epsilon,\delta)$-QDP gives QRDP an operational interpretation compatible with existing QDP guarantees.

Reading between the lines

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

  • Beyond the paper: the outcome-distribution definition suggests a stronger adversarial model in which the attacker chooses the POVM adaptively after seeing the state; whether Theorem 4 holds under that supremum-over-measurements definition is not settled by the paper's state-specific calculations.
  • Beyond the paper: because Theorem 4 needs only a QDP budget $\epsilon$, the same conversion would apply to any future quantum noise mechanism for which a QDP budget is derived, such as projection-operator or amplitude-damping variants.
  • Beyond the paper: the fidelity analysis is carried out for single-qubit states; extending it to entangled multi-qubit inputs could reveal whether the privacy-utility tradeoff sharpens or weakens in distributed settings.
  • Beyond the paper: framing QRDP as Rényi divergence over measurement outcomes connects naturally to quantum local differential privacy and to distinguishing measurements studied in shadow tomography, though the paper does not explore that link.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 6 minor

Summary. The paper proposes 'quantum Rényi differential privacy' (QRDP), defined by applying classical Rényi divergence to the classical measurement-outcome distributions of a quantum mechanism, with the privacy condition required for every POVM. It claims post-processing and composition properties for QRDP, gives privacy-budget expressions for generalized amplitude damping, phase-and-amplitude damping, and depolarizing noise, converts QRDP to (ε,δ)-QDP, and presents simulations of the privacy-fidelity tradeoff.

Significance. If the central claims were correct, the paper would provide a Rényi-version of quantum differential privacy with composition guarantees suitable for distributed quantum computing, and explicit privacy-budget formulas for common noise channels. The manuscript is not without useful pieces: Theorem 4's formula is a plausible and potentially applicable upper bound, and the simulation study addresses a relevant tradeoff. However, the main conceptual contribution is essentially classical RDP applied to quantum-induced outcome distributions, and the composition and post-processing proofs contain concrete errors. The claimed novelty and the QDC-specific composition guarantee are therefore not established in the present form.

major comments (5)
  1. [Definition 5 / Definition 6, Eq. (5)] Definition 5 defines D_α(E_m(ρ)||E_m(σ)) as the classical Rényi divergence between the classical distributions induced by a fixed POVM M. Consequently, QRDP in Definition 6 is exactly classical RDP applied to the classical channel M∘E, quantified over all POVMs; it is not a new quantum Rényi divergence. The discussion in Section 7.1 contrasts Eq. (5) with the established quantum Rényi divergence in Eq. (30), but Eq. (5) is the classical divergence on measurement outcomes. This is not merely a naming issue: the paper's novelty claim and the interpretation of the composition theorems depend on treating Eq. (5) as a quantum generalization, which it is not.
  2. [Proposition 1, Eq. (11)] The proof of post-processing invokes the wrong direction of the data-processing inequality. For Rényi divergence, data processing gives D_α(g(P)||g(Q)) ≤ D_α(P||Q) for any map g, whereas Eq. (11) states the reverse inequality. As written, the proof does not establish that QRDP is preserved under post-processing. The statement may be true, but it must be proved with the correct inequality and with care about the POVM on F(E(ρ)).
  3. [Propositions 2 and 3] The central composition claim is not established for the definition's quantification over all POVMs. In Proposition 2, the proof in Eq. (14) factorizes the joint outcome distribution as a product of the two mechanisms' outcome probabilities; this is valid only for product POVMs on R1⊗R2. Definition 6 requires the Rényi bound for every POVM on the joint output, including entangled measurements whose distribution does not factorize. Proposition 3 is the sequential composition relevant to QDC, but its proof is a citation to the classical theorem plus the assertion that E_m(ρ) and E_m(σ) are 'two special probability distributions'; no argument addresses adaptive composition when the second mechanism receives quantum side information or is followed by an arbitrary joint POVM. Thus the paper's key claim that 'QRDP composition is well suited for QDC' is unsupported.
  4. [Theorems 1-4, Eqs. (18), (22), (25), (26)] Theorems 1-3 restate Definition 6 rather than derive a privacy bound: the left-hand side of Eq. (18), (22), and (25) is exactly the divergence expression, and the inequality merely names that expression as ε. No closed-form or input-independent bound is obtained, so the abstract's claim of 'the lowest privacy budget provided by these mechanisms' is not supported. Theorem 4 does give a closed-form formula, but it is an upper bound inherited from the classical tight pure-DP-to-RDP conversion applied per POVM; the proof's justification 'can be seen as two special probability distributions' is not a derivation. The 'lowest' terminology is unjustified unless tightness over all POVMs is explicitly shown.
  5. [Appendix A.1, proof of Theorem 6] The derivative argument in the proof of Theorem 6 contains a sign error. With t = √(1−γ), dF/dγ = (dF/dt)(dt/dγ); the paper states dF/dt < 0 and dt/dγ < 0, which would imply dF/dγ > 0, contradicting the claimed conclusion dF/dγ ≤ 0. The assertion that the cubic term makes the negative part dominate is also not a proof. Since Theorem 5 is used in Section 6 to interpret the privacy-utility tradeoff, this theorem needs a valid proof.
minor comments (6)
  1. [Eq. (6)] The expression for D_1 is incorrect: the Kullback-Leibler divergence is E_{E_m(σ)}[log(E_m(ρ)/E_m(σ))], not log E_{E_m(σ)}[(E_m(ρ)/E_m(σ))^α].
  2. [Definition 4] The text says '(α,ε)-QRDP for short' while defining classical RDP; this should be '(α,ε)-RDP'.
  3. [Section 6, first paragraph] The text says 'we choose Theorem 5 to calculate the privacy budget', but Theorem 5 concerns fidelity; the formula used is Theorem 4.
  4. [Throughout] There are recurring typos and inconsistencies: 'simplex' should be 'simplify', and DQC/QDC are used inconsistently.
  5. [Section 4.4] The phrase 'since we do not scale the formulas, thereby allowing for a more precise estimation of the privacy budget' is unclear, and the word 'lowest' is not justified; an upper bound would be the accurate description.
  6. [Reference [35]] Reference [35] is a non-archival blog post; for the tight pure-DP-to-RDP conversion, please cite the peer-reviewed source or provide a self-contained proof.

Circularity Check

2 steps flagged · score 7.0 of 10

QRDP's derivation chain reduces to classical RDP over measurement-outcome distributions: Theorems 1–3 restate Definition 6 as 'privacy budgets', and the composition theorem is imported from Mironov via the 'special probability distributions' phrase; the universal-POVM requirement is never proven.

  1. self definitional [Section 4, Theorems 1–3 and proof of Theorem 1; Eqs. (18), (22), (25); compare Definition 6, Eq. (8)]
    "Theorem 1 ... provides (α,ε)-quantum Rényi differential private, 1/(α−1) log E_{E_m(σ)}(E_m(ρ)/E_m(σ))^α = 1/(α−1) log Σ_m [Tr(M(ρ''))]^α/[Tr(M(σ''))]^{α−1} ≤ ε ... Proof. ... Eq.(5) can be expanded as ... = 1/(α−1) log Σ_m P(E_m(ρ)=m)^α/P(E_m(σ)=m)^{α−1} ≤ ε."

    Definition 6 defines QRDP exactly as D_α(E_m(ρ)||E_m(σ)) ≤ ε, and Definition 5 defines that D_α as the classical Rényi divergence of the outcome distributions. The LHS of (18) is precisely this divergence expanded via Tr(Mρ'') probabilities. The proof only rewrites Definition 5 and appends '≤ ε'; no upper bound is computed. Thus the 'derivation of the lowest privacy budget' for GAD (and similarly PAD, Dep) is the QRDP definition restated with ε set equal to the divergence expression. Any mechanism would satisfy the theorem by taking ε to be its own divergence, so the claimed result is equivalent to its input by construction.

  2. renaming known result [Section 3.1, Definition 5 and Section 3.2, Proposition 3 proof; Eqs. (5) and (14)]
    "Definition 5 ... For two probability distribution E_m(ρ) and E_m(σ), the Rényi divergence of order α > 1 is D_α(E_m(ρ)||E_m(σ)) = 1/(α−1) log E_{E_m(σ)}(E_m(ρ)/E_m(σ))^α. ... Proof. Proposition 1 in [30] shows that the associated composition theorem holds in classical computing. Since our definitions of E_m(ρ) and E_m(σ) can be seen as two special probability distributions, the associated composition theorem also holds in quantum computing."

    The 'quantum Rényi divergence' is not a quantum divergence: it is the classical Rényi divergence of the discrete measurement-outcome distributions. Once QRDP is so defined, the composition result is the classical Mironov RDP composition theorem applied to those distributions. Proposition 3's proof says exactly this: it cites Mironov and uses the phrase 'can be seen as two special probability distributions' to transfer the classical theorem. The paper presents this import as a new QRDP composition theorem for QDC, but it is a relabeling of classical RDP by construction; no quantum argument addresses the universal-POVM quantifier in Definition 6.

full rationale

The two steps above capture the main circularity. Definition 5 merely instantiates classical Rényi divergence on POVM-outcome distributions, so 'quantum Rényi divergence' is a renaming rather than a new object; Proposition 3 (and the recurring 'special probability distributions' sentence) imports Mironov's classical composition rather than proving the quantum statement. Theorems 1–3 are more directly circular: their LHS is the Definition 5 divergence and the proof appends '≤ ε' without deriving any bound, so the claimed privacy budgets are restatements of the QRDP definition. Theorem 4 is not itself circular: it transparently substitutes the Zhou–Ying QDP epsilons into Steinke's classical PDP-to-RDP bound (Eq. (27)), which is a legitimate corollary of prior work, though it means the headline 'intuitive QRDP' formula is derivative. There is no load-bearing self-citation; the authors' prior work is cited only for tuning noise or as related literature. Two non-circular correctness gaps should also be noted: Proposition 2's proof factorizes the joint output distribution (Eq. (14)), which is valid only for tensor-product POVMs and does not cover the arbitrary POVMs required by Definition 6, and Proposition 1's proof cites a data-processing inequality in the reversed direction (Eq. (11) has D_α(P||Q) ≤ D_α(g(P)||g(Q)), whereas classical Rényi post-processing gives ≥). These gaps do not by themselves make the paper circular, but they reinforce that the derivation chain relies on the classical RDP analogy rather than on a proven quantum argument. Overall, the central privacy-budget and composition claims reduce by construction to classical RDP or to the QRDP definition, so the paper is substantially circular despite containing independent analyses of fidelity and noise-parameter trends.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The framework pulls its mathematical core from classical RDP (Mironov, Steinke) and QDP (Zhou-Ying); its one novel piece, QRDP, is a re-labeling of RDP for measurement outcome distributions. The main 'derivations' either restate the definition or substitute known formulas, so the ledger is dominated by imported results and a domain assumption about what privacy means.

free parameters (3)
  • α (Rényi order) = user-specified
    The privacy definition and all budget formulas are parameterized by α; it is not fitted but is the knob that makes QRDP flexible.
  • p (GAD/PAD thermal mixing probability) = 0.5 (assumed)
    The paper fixes p=0.5 'to simplex the mechanism as [43]', so Theorems 1 and 2 only cover the symmetric thermal case.
  • d (trace distance between neighboring states) = d=0.1 in simulations; arbitrary in definition
    All privacy budgets depend on the chosen adjacency distance d; it is an input to the model, not derived.
assumptions (5)
  • standard math Standard quantum mechanics: density matrices, unitary gates, POVM measurements, trace distance (Section 2.1).
    Background formalism for states and measurements.
  • domain assumption Privacy is defined on the classical probability distribution of measurement outcomes, not on the quantum states themselves (Definition 5).
    This is the central modeling choice that lets the paper import classical RDP results.
  • ad hoc to paper Classical RDP results from Mironov [30] and Steinke [35] apply unchanged to quantum measurement outcome distributions.
    Propositions 1-4 and Theorem 4 rely on the repeated assertion that E_m(ρ) and E_m(σ) 'can be seen as two special probability distributions'; no quantum-specific verification is provided.
  • domain assumption QDP privacy budgets for GAD, PAD, and depolarizing mechanisms from Zhou-Ying [43] are correct and directly substitutable.
    Theorem 4 derives its epsilon values from [43] Theorems 1-3 rather than proving them.
  • ad hoc to paper p=0.5 in GAD and PAD mechanisms (Section 4.1).
    The paper restricts to the symmetric thermal case without justification beyond matching [43].

how reviews work

0 comments
Cite this review

Pith. "Pith review of Differential Privacy Preserving Distributed Quantum Computing." pith.science (2026). https://pith.science/paper/HC2N6L4V

@misc{pith2026241212387,
  author       = {Pith},
  title        = {Pith review of: Differential Privacy Preserving Distributed Quantum Computing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HC2N6L4V}},
  note         = {Machine review of arXiv:2412.12387}
}
abstract

Existing quantum computers can only operate with hundreds of qubits in the Noisy Intermediate-Scale Quantum (NISQ) state, while quantum distributed computing (QDC) is regarded as a reliable way to address this limitation, allowing quantum computers to achieve their full computational potential. However, similar to classical distributed computing, QDC also faces the problem of privacy leakage. Existing research has introduced quantum differential privacy (QDP) for privacy protection in central quantum computing, but there is no dedicated privacy protection mechanisms for QDC. To fill this research gap, our paper introduces a novel concept called quantum R\'enyi differential privacy (QRDP), which incorporates the advantages of classical R\'enyi DP and is applicable in the QDC domain. Based on the new quantum R\'enyi divergence, QRDP provides delicate and flexible privacy protection by introducing parameter $\alpha$. In particular, the QRDP composition is well suited for QDC, since it allows for more precise control of the total privacy budget in scenarios requiring multiple quantum operations. We analyze a variety of noise mechanisms that can implement QRDP, and derive the lowest privacy budget provided by these mechanisms. Finally, we investigate the impact of different quantum parameters on QRDP. Through our simulations, we also find that adding noise will make the data less usable, but increase the level of privacy protection.

Figures

Figures reproduced from arXiv: 2412.12387 by the authors.

Figure 1
Figure 1. Execution of multiple quantum circuit instances [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Relationship between 𝜖 and 𝛼 under generalized amplitude damping 𝛾 [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 5
Figure 5. Relationship between 𝜖 and 𝛼 under depolarizing mechanism 𝑝. Depolarizing mechanism. According to Eq.(24) and (26), under the Dep mechanism, when the quantum dataset and circuit are determined, 𝐷 is fixed (𝐷 = 2 for the single qubit input), so (𝛼, 𝜖)- QRDP is only related to 𝑝 [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figures from the paper (2 more)
Figure 6
Figure 6. Figure 6: Relationship between 𝜖 and 𝐹 under generalized amplitude damping 𝛾 [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 9
Figure 9. Figure 9: Relationship between 𝜖 and 𝐹 under depolarizing mechanism 𝑝. 7 Discussion 7.1 Rationale for the proposed quantum Rényi divergence In preparation for QRDP, we found that quantum Rényi divergence has already been proposed in [38] from the perspective of quantum informati…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Differentially private quantum sensor networks

    quant-ph 2026-07 conditional novelty 7.0 of 10

    Differentially private quantum sensor network protocols are introduced that inject noise into the sensing Hamiltonian, achieving (O(1), δ)-differential privacy while retaining Heisenberg-limited MSE scaling under hone...

  2. A conditional no-go for resource-free magic-axis measurement on a static surface code

    eess.SY 2026-07 conditional novelty 6.0 of 10

    Under three stated assumptions, a resource-free static surface-code patch cannot sharply measure the magic axis at polynomial acceptance; it must pay with a resource, leave the dilute regime, or accept exponentially rarely.

Reference graph

Works this paper leans on

48 extracted references · 42 canonical work pages · cited by 2 Pith papers

  1. [1]

    Scott Aaronson and Guy N Rothblum. 2019. Gentle measurement of quantum states and differential privacy. In Proceedings of the 51st Annual ACM SIGACT Symposium on Theory of Computing . 322–333

    work page 2019

  2. [2]

    Shahab Asoodeh, Jiachun Liao, Flavio P Calmon, Oliver Kosut, and Lalitha Sankar

    work page

  3. [3]

    Jean-Philippe Aumasson. 2017. The impact of quantum computing on cryptog- raphy. Computer Fraud & Security 2017, 6 (June 2017), 8–11

    work page 2017

  4. [4]

    Borja Balle, Gilles Barthe, Marco Gaboardi, Justin Hsu, and Tetsuya Sato. 2020. Hypothesis testing interpretations and renyi differential privacy. In International Conference on Artificial Intelligence and Statistics . 2496–2506

    work page 2020

  5. [5]

    David Barral, F Javier Cardama, Guillermo Díaz, Daniel Faílde, Iago F Llovo, Mariamo Mussa Juane, Jorge Vázquez-Pérez, Juan Villasuso, César Piñeiro, Na- talia Costas, et al. 2024. Review of distributed quantum computing. from single qpu to high performance quantum computing. arXiv preprint arXiv:2404.01265 (April 2024)

    arXiv 2024

  6. [6]

    Carlo Cafaro and Peter van Loock. 2014. Approximate quantum error correction for generalized amplitude-damping errors. Physical Review A 89, 2 (February 2014), 022316

    work page 2014

  7. [7]

    Marcello Caleffi, Michele Amoretti, Davide Ferrari, Jessica Illiano, Antonio Man- zalini, and Angela Sara Cacciapuoti. 2024. Distributed quantum computing: a survey. Computer Networks (August 2024), 110672

    work page 2024

  8. [8]

    Marcello Caleffi, Angela Sara Cacciapuoti, and Giuseppe Bianchi. 2018. Quantum internet: From communication to distributed computing!. In Proceedings of the 5th ACM international conference on nanoscale computing and communication . 1–4

    work page 2018

Show all 48 references
  1. [9]

    Jing-Ling Chen and Abraham A Ungar. 2002. The Bloch gyrovector. Foundations of Physics 32 (2002), 531–565

    2002

  2. [10]

    Fida Kamal Dankar and Khaled El Emam. 2013. Practicing differential privacy in health care: A review. Trans. Data Priv. 6, 1 (2013), 35–67

    2013

  3. [11]

    Vasil S Denchev and Gopal Pandurangan. 2008. Distributed quantum computing: A new frontier in distributed systems or science fiction? ACM SIGACT News 39, 3 (September 2008), 77–95

    2008

  4. [12]

    Yuxuan Du, Min-Hsiu Hsieh, Tongliang Liu, Shan You, and Dacheng Tao. 2022. Quantum differentially private sparse regression learning. IEEE Transactions on Information Theory 68, 8 (April 2022), 5217–5233

    2022

  5. [13]

    Cynthia Dwork. 2006. Differential privacy. In International colloquium on au- tomata, languages, and programming . Springer, 1–12

    2006

  6. [14]

    Cynthia Dwork. 2008. Differential privacy: A survey of results. In International conference on theory and applications of models of computation . Springer, 1–19

    2008

  7. [15]

    Michael Elkin, Hartmut Klauck, Danupon Nanongkai, and Gopal Pandurangan

  8. [16]

    Davide Ferrari, Stefano Carretta, and Michele Amoretti. 2023. A modular quan- tum compilation framework for distributed quantum computing. IEEE Transac- tions on Quantum Engineering (August 2023)

    2023

  9. [17]

    Gambetta

    J. Gambetta. 2023. The hardware and software for the era of quantum utility is here. https://www.ibm.com/quantum/blog/quantum-roadmap-2033

    2023

  10. [18]

    Ziv Goldfeld, Dhrumil Patel, Sreejith Sreekumar, and Mark M Wilde. 2024. Quan- tum neural estimation of entropies. Physical Review A 109, 3 (March 2024), 032431

    2024

  11. [19]

    Harper R Grimsley, Sophia E Economou, Edwin Barnes, and Nicholas J Mayhall

  12. [20]

    Ji Guan. 2024. Optimal Mechanisms for Quantum Local Differential Privacy. arXiv preprint arXiv:2407.13516 (July 2024)

    2024

  13. [21]

    Ji Guan, Wang Fang, Mingyu Huang, and Mingsheng Ying. 2023. Detecting violations of differential privacy for quantum algorithms. In Proceedings of the 2023 ACM SIGSAC Conference on Computer and Communications Security . 2277– 2291

    2023

  14. [22]

    Kuk-Hyun Han and Jong-Hwan Kim. 2000. Genetic quantum algorithm and its application to combinatorial optimization problem. In Proceedings of the 2000 congress on evolutionary computation. CEC00 (Cat. No. 00TH8512) , Vol. 2. 1354– 1360

    2000

  15. [23]

    Christoph Hirche, Cambyse Rouzé, and Daniel Stilck França. 2023. Quantum differential privacy: An information theory perspective. IEEE Transactions on Information Theory 69, 9 (September 2023), 5771–5787

    2023

  16. [24]

    Richard Jozsa. 1994. Fidelity for mixed quantum states. Journal of modern optics 41, 12 (December 1994), 2315–2323

    1994

  17. [25]

    Keyi Ju, Xiaoqi Qin, Hui Zhong, Xinyue Zhang, Miao Pan, and Baoling Liu

  18. [26]

    Keyi Ju, Xiaoqi Qin, Hui Zhong, Xinyue Zhang, Miao Pan, and Baoling Liu. 2024. Privacy Preserving Quantum Search Mechanism Using Grover’s Algorithm. In 2024 International Conference on Quantum Communications, Networking, and Computing (QCNC). 204–210

    2024

  19. [27]

    Keyi Ju, Hui Zhong, Xinyue Zhang, Xiaoqi Qin, and Miao Pan. 2024. Quan- tum Computing Catalyzes Future Quantum Networks: Efficiency and Inherent Privacy. IEEE Network (May 2024)

    2024

  20. [28]

    Yuqing Li, Yusheng Zhao, Xinyue Zhang, Hui Zhong, Miao Pan, and Chi Zhang

  21. [29]

    Yeong-Cherng Liang, Yu-Hao Yeh, Paulo EMF Mendonça, Run Yan Teh, Mar- garet D Reid, and Peter D Drummond. 2019. Quantum fidelity measures for mixed states. Reports on Progress in Physics 82, 7 (June 2019), 076001

    2019

  22. [30]

    Ilya Mironov. 2017. Rényi differential privacy. In2017 IEEE 30th computer security foundations symposium (CSF). 263–275

    2017

  23. [31]

    Jack Murtagh and Salil Vadhan. 2015. The complexity of computing the optimal composition of differential privacy. In Theory of Cryptography Conference . 157– 175

    2015

  24. [32]

    Stefano Pirandola, Ulrik L Andersen, Leonardo Banchi, Mario Berta, Darius Bunandar, Roger Colbeck, Dirk Englund, Tobias Gehring, Cosmo Lupo, Carlo Ottaviani, et al. 2020. Advances in quantum cryptography. Advances in optics and photonics 12, 4 (2020), 1012–1236

    2020

  25. [33]

    In 2024 International Conference on Quantum Communications, Networking, and Computing (QCNC)

    Differential privacy preserving quantum computing via projection operator measurements. In 2024 International Conference on Quantum Communications, Networking, and Computing (QCNC) . 236–243

    2024

  26. [34]

    Abhishek Sharma and Amit Kumar. 2019. A survey on quantum key distribution. In 2019 International Conference on Issues and Challenges in Intelligent Computing Techniques (ICICT), Vol. 1. 1–4

    2019

  27. [35]

    Thomas Steinke. 2024. Tight RDP & zCDP Bounds from Pure DP. Differential- Privacy.org. https://differentialprivacy.org/pdp-to-zcdp/

    2024

  28. [36]

    Swamit S Tannu and Moinuddin K Qureshi. 2019. Not all qubits are created equal: A case for variability-aware policies for NISQ-era quantum computers. In Proceedings of the twenty-fourth international conference on architectural support for programming languages and operating s...

    2019

  29. [37]

    Kristan Temme, Sergey Bravyi, and Jay M Gambetta. 2017. Error mitigation for short-depth quantum circuits. Physical review letters 119, 18 (November 2017), 180509

    2017

  30. [38]

    Yi Qiao, Zhaobin Liu, Haoze Lv, Minghui Li, Zhiyi Huang, Zhiyang Li, and Weijiang Liu. 2019. An effective data privacy protection algorithm based on differential privacy in edge computing. IEEE Access 7 (September 2019), 136203– 136213

    2019

  31. [39]

    transition probability

    Armin Uhlmann. 1976. The “transition probability” in the state space of a𝐶∗- algebra. Reports on Mathematical Physics 9, 2 (April 1976), 273–279

    1976

  32. [40]

    William M Watkins, Samuel Yen-Chi Chen, and Shinjae Yoo. 2023. Quantum machine learning with differential privacy. Scientific Reports 13, 1 (February 2023), 2453

    2023

  33. [41]

    Yusheng Zhao, Hui Zhong, Xinyue Zhang, Chi Zhang, and Miao Pan. 2024. Bridging Quantum Computing and Differential Privacy: A Survey on Quantum Computing Privacy. arXiv preprint arXiv:2403.09173 (March 2024)

    2024 arXiv

  34. [42]

    Hui Zhong, Keyi Ju, Manojna Sistla, Xinyue Zhang, Xiaoqi Qin, Xin Fu, and Miao Pan. 2023. Tuning quantum computing privacy through quantum error correction. arXiv preprint arXiv:2312.14521 (December 2023)

    2023 arXiv

  35. [43]

    Marco Tomamichel. 2015. Quantum information processing with finite resources: mathematical foundations. Vol. 5. Springer

    2015

  36. [48]

    Li Zhou and Mingsheng Ying. 2017. Differential privacy in quantum computation. In 2017 IEEE 30th Computer Security Foundations Symposium (CSF) . 249–262. A Appendix: Theorem 5 We provide a more detailed explanation of Theorem 5, outlining the fidelity trends under three differ...

    2017

  37. [2014]

    In Proceedings of the 2014 ACM symposium on Principles of distributed computing

    Can quantum communication speed up distributed computation?. In Proceedings of the 2014 ACM symposium on Principles of distributed computing . 166–175

    2014

  38. [2019]

    Nature communications 10, 1 (July 2019), 3007

    An adaptive variational algorithm for exact molecular simulations on a quantum computer. Nature communications 10, 1 (July 2019), 3007

    2019

  39. [2021]

    IEEE Journal on Selected Areas in Information Theory 2, 1 (January 2021), 208–222

    Three variants of differential privacy: Lossless conversion and applications. IEEE Journal on Selected Areas in Information Theory 2, 1 (January 2021), 208–222

    2021

  40. [2024]

    In ICC 2024-IEEE International Conference on Communications

    Harnessing inherent noises for privacy preservation in quantum machine Hui Zhong, Keyi Ju, Jiachen Shen, Xinyue Zhang, Xiaoqi Qin, Tomoaki Ohtsuki, Miao Pan, and Zhu Han learning. In ICC 2024-IEEE International Conference on Communications . 1121– 1126

    2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.