Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Factorization by extremal privacy mechanisms: new insights into efficiency

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read This paper proves that under $\alpha$-local differential privacy, the maximal Fisher information of a regular one-dimensional parametric model behaves like $\frac{\alpha^2}{4}(\int |s_{\theta_0}(x)|p_{\theta_0}(x)dx)^2$ as $\alpha\to 0$…

desk verdict Solid high-privacy Fisher information characterization with a real gap in the existence proof and an overclaimed uniform application; worth refereeing but needs fixing. read the letter →

arxiv 2507.21769 v1 pith:J4EMW64A submitted 2025-07-29 math.ST math.PRstat.TH

Chiara Amorino , Arnaud Gloter This is my paper · ORCID
classification math.STmath.PRstat.TH MSC 62F1268P2762B1546A55
keywords localdifferentialprivacyFisherinformationextremalmechanismsstaircaseChoquettheoremasymptoticefficiencyuniformdistributionestimationfactorizationlemma
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

This paper identifies the statistical efficiency ceiling for locally differentially private estimation. It proves that any $\alpha$-local differential privacy mechanism can be factorized as an extremal 'staircase' mechanism followed by arbitrary extra randomization, and because the extra step can only destroy Fisher information, only the extremal mechanisms matter. For a regular one-dimensional parametric model the maximal Fisher information under $\alpha$-LDP is asymptotically $\frac{\alpha^2}{4}(\int |s_{\theta_0}(x)|p_{\theta_0}(x)dx)^2$ as $\alpha\to 0$, with matching upper and lower bounds, and for every $\alpha>0$ an extremal mechanism attaining the maximum exists. These results give the exact high-privacy benchmark for efficient estimation and a construction principle that the authors then apply to estimating the range of a uniform distribution.

What carries the argument

The central object is the factorization lemma: any $\alpha$-LDP channel $q$ factorizes as $q=q^{(2)}\circ q^{(\mu)}$, where $q^{(\mu)}$ is extremal, meaning its log-likelihood ratios take only the values $0$ and $\alpha$, and $q^{(2)}$ is arbitrary extra randomization. In finite spaces this follows by writing each ratio vector in the hyperrectangle $[1,e^\alpha]^d$ as a convex combination of its vertices via Carathéodory's theorem; in continuous spaces the same decomposition is obtained by Choquet's theorem on the compact convex set $C=\{1\le v\le e^\alpha\}$ in $L^\infty$ with the weak-$*$ topology, whose extreme points are the measurable functions with values in $\{1,e^\alpha\}$. A measurable point-evaluation operator and Bochner integrals make the pointwise evaluation of these integrals rigorous. Because the extra randomization step can only decrease Fisher information, the optimization problem reduces to choosing the sub-probability measure $\mu$ on the extreme set $E$.

What would settle it

For the central theorem, one could numerically search for an $\alpha$-LDP mechanism on a standard normal location model whose Fisher information exceeds $(e^\alpha-1)^2/(2\pi)$ at a fixed $\alpha$; any such mechanism would refute the upper bound in Theorem 2. For the uniform application, the paper's own Figure 1 provides the disconfirming observation: with $\hat{\theta}_p=1.3\theta_0$ and $\alpha=0.3$, the estimator's empirical mean stays near $1.3\theta_0$ rather than $\theta_0$, exactly the predicted failure mode.

Watch

Extended reading notes

Core claim

The central claim is that the optimization problem $\sup_{q\in\mathcal{Q}_\alpha} I_{\theta_0}(q\circ P)$ has a complete answer in one dimension. For any regular model, $J^{\max,\alpha}_{\theta_0}$ is squeezed between $\frac{(e^\alpha-1)^2}{2e^\alpha(1+e^\alpha)}(\int |s_{\theta_0}|p_{\theta_0})^2$ and $\frac{(e^\alpha-1)^2}{4}(\int |s_{\theta_0}|p_{\theta_0})^2$, so as $\alpha\to 0$ the maximal Fisher information is equivalent to $\frac{\alpha^2}{4}(\int |s_{\theta_0}|p_{\theta_0})^2$. Theorems 2 and 3 further show that the supremum is always attained, by an extremal mechanism $q^{(\mu)}_x(dr)=e_x(r)\mu(dr)$ for some Radon sub-probability measure $\mu$ on the extreme points $E=\{r\text{ measurable}: r\in\{1,e^\alpha\}\text{ a.e.}\}$. In the finite case, Theorem 1 gives the exact value $\frac{(e^\alpha-1)^2}{4}\cdot \frac{E[|s_{\theta_0}(X)|]^2}{[(1-n_{\max})+e^\alpha n_{\max}][n_{\max}+(1-n_{\max})e^\alpha]}$, achieved by a mechanism that randomizes between the two score-sign regions. The authors apply the same two-point mechanism to the uniform model $U[0,\theta]$, proving a consistent, asymptotically Gaussian estimator whose variance matches the bound as $\alpha\to 0$ whenever the preliminary estimate satisfies $\hat{\theta}_p\le \theta_0$.

Load-bearing premise

The uniform-law efficiency result assumes a preliminary estimate $\hat{\theta}_p$ that is no larger than the true $\theta_0$, yet the paper gives no private construction of such an estimate; if $\hat{\theta}_p$ exceeds $\theta_0$, Proposition 4 shows the estimator converges to $\theta_0\vee\hat{\theta}_p$ and stays biased.

Editorial extensions

If this is right

  • In the high-privacy regime, no $\alpha$-LDP estimator for a regular one-dimensional model can beat asymptotic variance $4/(\alpha^2(\int |s_{\theta_0}|p_{\theta_0}dx)^2)$ per observation, and the score-sign staircase mechanism achieves it.
  • Any optimal $\alpha$-LDP mechanism can be taken to be extremal, so the search over all privacy channels collapses to the choice of a sub-probability measure on the extreme set $E$.
  • For every fixed $\alpha>0$, the Fisher-information maximization over $\alpha$-LDP mechanisms has a solution, not merely a supremum.
  • For the uniform range model, the proposed two-point mechanism yields a $\sqrt{n}$-consistent estimator with variance equivalent to $\theta_0^2/\alpha^2$ when the preliminary estimate equals $\theta_0$, matching the Fisher-information upper bound.

Reading between the lines

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

  • The same factorization argument should apply to other utility criteria, since Proposition 2 maximizes any bounded continuous convex functional $j$ over measures on $E$; one could define efficient mechanisms for risks other than Fisher information.
  • The multidimensional extension is left open in the paper, but the factorization itself is dimension-free; a natural next step is to maximize a trace or directional version of the Fisher information matrix, where the scalar score-sign split becomes a hyperplane choice.
  • The uniform example exposes a general two-stage recipe: use a cheap preliminary estimate to locate the score-sign boundary, then apply the extremal Bernoulli mechanism; making the preliminary step itself private without losing the below-truth guarantee is the practical bottleneck.
  • If the conjecture in Remark 13 holds, the continuous staircase mechanism would transfer to metric spaces with a non-atomic measure, opening the way to locally private inference for stochastic processes.
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

3 major / 5 minor

Summary. This paper studies the maximal Fisher information achievable by an alpha-locally differentially private mechanism for a one-dimensional regular parametric model. The authors prove a factorization lemma showing that any alpha-LDP channel can be written as an extremal 'staircase' channel followed by post-processing; in the continuous case the factorization is obtained through Choquet's theorem on the set C={1<=v<=e^alpha} in L^infty. For discrete models they obtain an exact small-alpha formula for the maximal Fisher information (Theorem 1); for continuous models they obtain matching upper and lower bounds with asymptotic constant alpha^2/4 (integral |s_{theta0}| p_{theta0} dx)^2 as alpha->0 (Theorem 2) and prove existence of an extremal mechanism attaining the supremum for every alpha (Theorem 3). The final section applies the framework to estimating the endpoint theta of a uniform distribution on [0,theta], proposing a two-point extremal mechanism and a two-stage estimator whose asymptotic variance matches the derived Fisher information bound when a preliminary estimate theta_hat_p <= theta0 is available. Numerical experiments illustrate the behavior for theta_hat_p above and below theta0.

Significance. The conceptual contribution is valuable: the factorization viewpoint unifies the discrete staircase mechanisms of Kairouz et al. with a continuous Choquet-based representation, and Theorem 2 gives a clean, explicit asymptotic benchmark that matches the Gaussian case in [30]. The proof of Theorem 3 is an interesting application of convex analysis (maximal measures and Zorn's lemma) to a problem where the extremal set E is not closed. The discrete theorem is carefully worked out, with an explicit two-point optimal mechanism. The uniform application is attractive and the numerical experiments support the conditional claims. However, two load-bearing points need repair before the advertised results are fully established: the uniform-continuity step in the proof of Theorem 3 and the missing private preliminary estimator in Section 5.2.

major comments (3)
  1. [7.2.3, proof of Theorem 3] The proof asserts that 't_theta0 and p_tilde_theta0 are uniformly continuous on the compact set C', but the preceding results only establish continuity on E: Lemma 1(3) states continuity on E, and E is not compact in the d_star topology (Remark 12). The functions t_theta0 and p_tilde_theta0 are not defined on all of C in the manuscript. Since the claim that the approximation error in Eq. (69) can be made arbitrarily small depends on this uniform continuity, the existence proof is incomplete as written. The fix is straightforward if formulas (32)-(33) are used to extend p_tilde_theta0 and t_theta0 to all of C and their weak-* continuity is proved there; please do so or replace the argument.
  2. [5.2, Remark 11, and abstract] The abstract's claim that the proposed mechanism 'yields a consistent and asymptotically efficient estimator in high privacy regime' is not established by the submitted proof. Proposition 4(1) shows that if the preliminary value theta_hat_p exceeds theta0, then theta_hat_n converges almost surely to theta0 vee theta_hat_p, i.e. it is inconsistent. The analysis in Proposition 4(2) is conditional on theta_hat_p <= theta0, and Remark 11 only asserts, without proof or construction, that theta_hat_p can be replaced by 'some preliminary consistent estimator based on a small subset of the data'. No alpha-LDP procedure producing a downward-consistent estimator of the support endpoint theta0 is given, and no argument shows that such an estimator can satisfy theta_hat_p/theta0 -> 1 fast enough for the stated variance equivalence. The efficiency statement for the uniform model should either be removed from the abstract or supplied with the missing preliminary estimator and its one-sided consistency proof.
  3. [5.1, Proposition 3] The proof derives a pointwise derivative of theta -> p_tilde_theta(r) for mu-almost every r and then computes the integral of (partial_theta p_tilde)^2 / p_tilde, but it does not verify the differentiability-in-quadratic-mean condition of Definition 2, which is what the statement 'the model is DQM' asserts. Since the subsequent efficiency comparison in Remark 11 uses this Fisher information as the asymptotic variance lower bound, Proposition 3 needs a proof of DQM, or else the statement should be weakened to the property actually established.
minor comments (5)
  1. [3.1 and 4.2] There are two results numbered Lemma 1: the discrete factorization lemma in Section 3.1 and the regularity lemma for continuous extremal mechanisms in Section 4.2; renumbering would prevent confusion.
  2. [4.1, Proposition 1] The definition q(z):=ess inf_x q_x(z) is asserted to satisfy q(z)>0 without argument; as in the discrete Lemma 1, outputs with q(z)=0 have q_x(z)=0 almost everywhere and should be discarded before dividing by q(z).
  3. [7.2.3, proof of Lemma 8] The homogeneity identity i(lambda r)=|lambda| i(r) is written for lambda in R, but i is only defined on C where the denominator is positive; the convexity argument should be restricted to the convex-combination parameters actually needed.
  4. [Remark 7] The comparison with [30] states the Gaussian optimal Fisher information as (2/pi)(e^alpha-1)^2/(e^alpha+1)^2; since the derivation is only sketched, a precise reference to the corresponding equation in [30] would help the reader verify the match.
  5. [Figures 1-2] The captions of Figures 1 and 2 are very terse; in particular, the dashed line in Figure 2 should be identified as the standard-deviation lower bound derived from Proposition 3.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the Fisher-information bounds and extremal mechanisms are proved in-paper from the factorization lemma and a solved linear program, with external DQM and Gaussian benchmarks; the §5.2 preliminary-estimator gap is a proof gap, not circularity.

full rationale

The central chain is self-contained. Lemma 1 and Proposition 1 prove the factorization of arbitrary α-LDP channels through extremal mechanisms using Carathéodory's and Choquet's theorems rather than assuming the reduction. The upper bounds in Theorem 1 and Theorem 2 are derived by bounding the score integrals (e.g., (77)) and solving the LP in Lemma 7, while the lower bounds are attained by explicit two-point Bernoulli/staircase mechanisms (Proposition 7 and the proof of Theorem 2); no quantity is fitted to data and then renamed a prediction. Existence of an optimal mechanism (Theorem 3) follows from Proposition 2, whose proof uses only compactness, convexity, and Choquet machinery. The only load-bearing external inputs are [45] for DQM preservation and the Fisher-information data-processing inequality (16)–(17), and [30] as an independent Gaussian benchmark; author self-citations [2] and [3] appear only as related work and are not used in the proofs. Flagged for completeness, but not circular: Section 5.2 and Remark 11 assume "a preliminary consistent estimator based on a small subset of the data" with θ̂p≤θ0 without constructing one under α-LDP; Proposition 4(1) shows that if θ̂p>θ0 the estimator converges to θ0∨θ̂p, so the abstract's unqualified uniform-efficiency claim is not fully established. This is a correctness/completeness gap in the application, not a reduction of the results to their inputs.

Assumptions & free parameters 0 free parameters · 5 assumptions · 1 invented entities

The theoretical results rely on standard regularity (DQM), the alpha-LDP density framework, and Choquet's theorem. The only ad hoc assumption specific to the paper's application is the availability of a preliminary estimator below the true parameter. No free parameters are fitted to data.

assumptions (5)
  • standard math Differentiability in quadratic mean (DQM) of the parametric model at theta0 with finite second moment of the score.
    Invoked in Section 3.2 (Definition 2) and Lemma 1 in Section 4.2; standard regularity condition in asymptotic statistics.
  • domain assumption Alpha-LDP mechanisms admit densities with respect to a common dominating measure, and the support of the output space is fixed.
    Used in Section 2 to rewrite the privacy constraint as a density ratio bound; needed for the factorization construction.
  • standard math Choquet's theorem, Banach-Alaoglu theorem, and the weak-* compactness of C in L-infinity.
    Central to the continuous factorization in Proposition 1 and existence in Proposition 2; stated in Section 6.
  • ad hoc to paper The evaluation operator e_x defined by a limsup of averages yields r(x) for almost every x and is measurable and Bochner-compatible.
    Developed in Section 6.3 and Proposition 5; essential for pointwise interpretation of Choquet representations. The proof is technical and somewhat delicate.
  • ad hoc to paper Existence of a preliminary estimator theta_hat_p that is independent of the main sample and satisfies theta_hat_p <= theta0.
    Assumed in Section 5.2 for the efficient uniform estimator; no private construction is provided, and the claim fails when theta_hat_p > theta0.
invented entities (1)
  • Continuous extremal (staircase) mechanism q(mu) with output space E = {r:X -> {1,e^alpha}}
    purpose: Represents any alpha-LDP mechanism as an extremal mechanism followed by post-processing, and provides the candidate optimal mechanisms in Theorems 2 and 3.
    This is a mathematical construction, the natural infinite-dimensional analogue of Kairouz's finite extremal mechanisms. It is rigorously defined and its properties are proven, but it is not an empirically observable entity with external falsifiable predictions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Factorization by extremal privacy mechanisms: new insights into efficiency." pith.science (2026). https://pith.science/paper/J4EMW64A

@misc{pith2026250721769,
  author       = {Pith},
  title        = {Pith review of: Factorization by extremal privacy mechanisms: new insights into efficiency},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J4EMW64A}},
  note         = {Machine review of arXiv:2507.21769}
}
abstract

We study the problem of efficiency under $\alpha$ local differential privacy ($\alpha$ LDP) in both discrete and continuous settings. Building on a factorization lemma, which shows that any privacy mechanism can be decomposed into an extremal mechanism followed by additional randomization, we reduce the Fisher information maximization problem to a search over extremal mechanisms. The representation of extremal mechanisms requires working in infinite dimensional spaces and invokes advanced tools from convex and functional analysis, such as Choquet's theorem. Our analysis establishes matching upper and lower bounds on the Fisher information in the high privacy regime ($\alpha \to 0$), and proves that the maximization problem always admits a solution for any $\alpha$. As a concrete application, we consider the problem of estimating the parameter of a uniform distribution on $[0, \theta]$ under $\alpha$ LDP. Guided by our theoretical findings, we design an extremal mechanism that yields a consistent and asymptotically efficient estimator in high privacy regime. Numerical experiments confirm our theoretical results.

Figures

Figures reproduced from arXiv: 2507.21769 by the authors.

Figure 2
Figure 2. Empirical Std [PITH_FULL_IMAGE:figures/full_fig_p024_2.png] view at source ↗

Discussion (0). Continue with ORCID 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. Fundamental Limit of Discrete Distribution Estimation under Utility-Optimized Local Differential Privacy

    cs.CR 2025-09 conditional novelty 8.0 of 10

    The exact asymptotic worst-case mean squared error for discrete distribution estimation under utility-optimized local differential privacy is characterized as a saddle-point optimization, and uBD schemes achieve it.

Reference graph

Works this paper leans on

47 extracted references · 42 canonical work pages · cited by 1 Pith paper

  1. [30]

    Efficient Estimation of a Gaussian Mean with Local Differential Privacy

    Kalinin, N. P., & Steinberger, L. (2024). Efficient estimation of a Gaussian mean with Local Differential Privacy. arXiv preprint arXiv:2402.04840

    work page Pith review arXiv 2024

  2. [1]

    B., Mironov, I., Talwar, K., & Zhang, L

    Abadi, M., Chu, A., Goodfellow, I., McMahan, H. B., Mironov, I., Talwar, K., & Zhang, L. (2016, October). Deep learning with differential privacy. In Proceedings of the 2016 ACM SIGSAC conference on computer and communications security (pp. 308-318)

    work page 2016

  3. [2]

    Amorino, C., & Gloter, A. (2025). Minimax rate for multivariate data under componentwise local differential privacy constraints. The Annals of Statistics, 53(3), 1176-1202

    work page 2025

  4. [3]

    Amorino, C., Gloter, A., & Halconruy, H. (2025). Evolving privacy: drift parameter estima- tion for discretely observed iid diffusion processes under LDP. Stochastic Processes and their Applications, 181, 104557

    work page 2025

  5. [4]

    Apple Differential Privacy Team. (2017). Learning with privacy at scale. Apple Mach. Learn. J, 1(8), 1-25. Available at https://machinelearning.apple.com/2017/12/06/learning- withprivacy-at-scale.html

    work page 2017

  6. [5]

    (2022, June)

    Asi, H., Feldman, V., & Talwar, K. (2022, June). Optimal algorithms for mean estimation under local differential privacy. In International Conference on Machine Learning (pp. 1046- 1056). PMLR

    work page 2022

  7. [6]

    Avella-Medina, M., Bradshaw, C., & Loh, P. L. (2023). Differentially private inference via noisy optimization. The Annals of Statistics, 51(5), 2067-2092

    work page 2023

  8. [7]

    P., Chen, W

    Barnes, L. P., Chen, W. N., & Özgür, A. (2020). Fisher information under local differential privacy. IEEE Journal on Selected Areas in Information Theory, 1(3), 645-659

    work page 2020

Show all 47 references
  1. [8]

    Berrett, T., & Butucea, C. (2020). Locally private non-asymptotic testing of discrete distri- butions is faster using interactive mechanisms. Advances in Neural Information Processing Systems, 33, 3164-3173

    2020

  2. [9]

    Berrett, T., & Yu, Y. (2021). Locally private online change point detection. Advances in Neural Information Processing Systems, 34, 3425-3437

    2021

  3. [10]

    et al (1993)

    Bickel, P. et al (1993). Efficient and adaptive estimation for semiparametric models (Vol. 4). Baltimore: Johns Hopkins University Press

    1993

  4. [11]

    Brezis, H. (1983). Analyse fonctionnelle. Théorie et applications. 49

    1983

  5. [12]

    (2016, October)

    Bun, M., & Steinke, T. (2016, October). Concentrated differential privacy: Simplifications, extensions, and lower bounds. In Theory of cryptography conference (pp. 635-658). Berlin, Heidelberg: Springer Berlin Heidelberg

    2016

  6. [13]

    Butucea, C., Dubois, A., Kroll, M., & Saumard, A. (2020). Local differential privacy: Elbow effect in optimal density estimation and adaptation over Besov ellipsoids

    2020

  7. [14]

    Butucea, C., Klockmann, K., & Krivobokova, T. (2025). Nonparametric spectral den- sity estimation using interactive mechanisms under local differential privacy. arXiv preprint arXiv:2504.00919

    2025

  8. [15]

    Butucea, C., Rohde, A., & Steinberger, L. (2023). Interactive versus noninteractive locally differentially private estimation: Two elbows for the quadratic functional. The Annals of Statistics, 51(2), 464-486

    2023

  9. [16]

    Choquet, G. (1960). Le théorème de représentation intégrale dans les ensembles convexes compacts. In Annales de l’institut Fourier (Vol. 10, pp. 333-344)

    1960

  10. [17]

    and Uhl, J.J

    Diestel, J. and Uhl, J.J. (1977). Vector Measures. American Mathematical Soc. ISBN 978- 0-8218-1515-1

    1977

  11. [18]

    C., Jordan, M

    Duchi, J. C., Jordan, M. I., & Wainwright, M. J. (2018). Minimax optimal procedures for locally private estimation. Journal of the American Statistical Association, 113(521), 182-201

    2018

  12. [19]

    (2019, June)

    Duchi, J., & Rogers, R. (2019, June). Lower bounds for locally private estimation via communication complexity. In Conference on Learning Theory (pp. 1161-1191). PMLR

    2019

  13. [20]

    C., & Ruan, F

    Duchi, J. C., & Ruan, F. (2024). The right complexity measure in locally private estimation: It is not the fisher information. The Annals of Statistics, 52(1), 1-51

    2024

  14. [21]

    Dwork, C., McSherry, F., Nissim, K., & Smith, A. (2006). Calibrating noise to sensitivity in private data analysis. In Theory of Cryptography: Third Theory of Cryptography Con- ference, TCC 2006, New York, NY, USA, March 4-7, 2006. Proceedings 3 (pp. 265-284). Springer Berlin ...

    2006

  15. [22]

    Dwork, C., & Rothblum, G. N. (2016). Concentrated differential privacy. arXiv preprint arXiv:1603.01887

    2016 arXiv

  16. [23]

    (2014, November)

    Erlingsson, Ú., Pihur, V., & Korolova, A. (2014, November). Rappor: Randomized ag- gregatable privacy-preserving ordinal response. In Proceedings of the 2014 ACM SIGSAC conference on computer and communications security (pp. 1054-1067)

    2014

  17. [24]

    Geng, Q., Kairouz, P., Oh, S., & Viswanath, P. (2015). The staircase mechanism in differ- ential privacy. IEEE Journal of Selected Topics in Signal Processing, 9(7), 1176-1184

    2015

  18. [25]

    Geng, Q., & Viswanath, P. (2014). The optimal mechanism in differential privacy. In 2014 IEEE international symposium on information theory (pp. 2371-2375). IEEE

    2014

  19. [26]

    Geng, Q., & Viswanath, P. (2015). Optimal noise adding mechanisms for approximate differential privacy. IEEE Transactions on Information Theory, 62(2), 952-969

    2015

  20. [27]

    (2009, May)

    Ghosh, A., Roughgarden, T., & Sundararajan, M. (2009, May). Universally utility- maximizing privacy mechanisms. In Proceedings of the forty-first annual ACM symposium on Theory of computing (pp. 351-360)

    2009

  21. [28]

    Joseph, M., Kulkarni, J., Mao, J., & Wu, S. Z. (2019). Locally private gaussian estimation. Advances in Neural Information Processing Systems, 32 50

    2019

  22. [29]

    Kairouz, P., Oh, S., & Viswanath, P. (2016). Extremal mechanisms for local differential privacy. Journal of Machine Learning Research, 17(17), 1-51

    2016

  23. [31]

    Kearns, M., & Roth, A. (2019). The ethical algorithm: The science of socially aware algo- rithm design. Oxford University Press

    2019

  24. [32]

    Kifer, D., & Machanavajjhala, A. (2014). Pufferfish: A framework for mathematical privacy definitions. ACM Transactions on Database Systems (TODS), 39(1), 1-36

    2014

  25. [33]

    Kroll, M. (2024). Nonparametric spectral density estimation under local differential privacy. Statistical Inference for Stochastic Processes, 27(3), 725-759

    2024

  26. [34]

    T., & Wang, Y

    Kulesza, A., Suresh, A. T., & Wang, Y. (2023). Mean estimation in the add-remove model of differential privacy. arXiv preprint arXiv:2312.06658

    2023 arXiv

  27. [35]

    Lang, S. (1993). Real and Functional Analysis. Graduate Texts in Mathematics. Springer New York. ISSN: 0072-5285

    1993

  28. [36]

    Lalanne, C. (2023). On the tradeoffs of statistical learning with privacy (Doctoral disserta- tion, Ecole normale supérieure de lyon-ENS LYON)

    2023

  29. [37]

    Lam-Weil, J., Laurent, B., & Loubes, J. M. (2022). Minimax optimal goodness-of-fit testing for densities and multinomials under a local differential privacy constraint. Bernoulli, 28(1), 579-600

    2022

  30. [38]

    B., & Yu, Y

    Li, M., Berrett, T. B., & Yu, Y. (2023). On robustness and local differential privacy. The Annals of Statistics, 51(2), 717-737

    2023

  31. [39]

    H., & Lee, S

    Nam, S. H., & Lee, S. H. (2022). A tighter converse for the locally differentially private discrete distribution estimation under the one-bit communication constraint. IEEE Signal Processing Letters, 29, 1923-1927

    2022

  32. [40]

    Robust de-anonymization of large sparse datasets

    Narayanan, Arvind, and Vitaly Shmatikov. Robust de-anonymization of large sparse datasets. 2008 IEEE Symposium on Security and Privacy (sp 2008). IEEE, 2008

    2008

  33. [41]

    Oxtoby, J. C. (1970). Homeomorphic measures in metric spaces. Proceedings of the Amer- ican Mathematical Society, 24(3), 419-423

    1970

  34. [42]

    Smith, A. (2008). Efficient, differentially private point estimators. arXiv preprint arXiv:0809.4794

    2008 arXiv

  35. [43]

    (2011, June)

    Smith, A. (2011, June). Privacy-preserving statistical estimation with optimal convergence rates. In Proceedings of the forty-third annual ACM symposium on Theory of computing (pp. 813-822)

    2011

  36. [44]

    Tang, J., Korolova, A., Bai, X., Wang, X., & Wang, X. (2017). Privacy loss in apple’s implementation of differential privacy on macos 10.12. arXiv preprint arXiv:1709.02753

    2017 arXiv

  37. [45]

    Steinberger, L.(2024).Efficiencyinlocaldifferentialprivacy.TheAnnalsofStatistics, 52(5), 2139-2166

    2024

  38. [46]

    Van der Vaart, A. W. (2000). Asymptotic statistics (Vol. 3). Cambridge university press

    2000

  39. [47]

    Wasserman, L., & Zhou, S. (2010). A statistical framework for differential privacy. Journal of the American Statistical Association, 105(489), 375-389. 51

    2010

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.