Pith. sign in

REVIEW 3 major objections 5 minor 50 references

Quantum Key Distribution Beyond Stationary Channels

T0 review · 3 major / 5 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read For non-stationary QKD channels, mixture-martingale bounds keep finite-key statistics sharp even when the channel model is wrong, cutting required pulses by over 70% in satellite simulations.

desk verdict The concentration inequalities are correct and the binomial benchmark result is elegant, but the >70% reduction is demonstrated only for a fixed mis-estimated loss, not for the fluctuating channel the abstract advertises. read the letter →

arxiv 2607.17690 v2 pith:2NNBSZEU submitted 2026-07-20 quant-ph

Vaisakh Mannalath , Víctor Zapatero , Kiyoshi Tamaki , Marcos Curty This is my paper · ORCID
classification quant-ph MSC 81P9460G4260E15 PACS 03.67.Dd
keywords quantumkeydistributionsatelliteQKDconcentrationinequalitiesmixturemartingalesnon-IIDfinitestatisticsdecoy-stateBB84channelmismatchfinite-keysecurity
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

Satellite QKD links fluctuate strongly, and the statistical bounds used to turn sparse detection events into a secure key are typically calibrated to a specific channel loss that can be off by tens of decibels. The paper introduces mixture-martingale concentration inequalities for adaptive, non-IID data that average over a range of calibration parameters instead of committing to one guess. These bounds stay nearly as tight as the ideal binomial benchmark while remaining valid across a wide range of channel losses. In decoy-state BB84 simulations, they reduce the minimum number of transmitted pulses needed for positive secure key rate by more than 70% under a 10 dB loss mismatch.

What carries the argument

The mixture martingale M̃_n = (1/(B−A))∫_A^B M_n(λ)dλ, where M_n(λ) = ∏_{i=1}^n e^{−λξ_i}/E[e^{−λξ_i} | F_{i−1}] is the per-parameter exponential martingale. Its terminal lower bound K_{A,B}(s,u) = (1/(B−A))∫_A^B e^{−λs} [1−(1−e^{−λ}) u/N]^{−N} dλ, inverted via monotonicity, yields the one-sided confidence functions U_μ, L_Λ, U_Λ, L_μ through Proposition 1. A Gaussian-integral estimate around the optimizer λ*(x,y) shows the mixing penalty is only (1+N(B−A)²/(2π))^{−1/2}.

What would settle it

Run the satellite simulation with a time-varying transmittance process (e.g., log-normal fading or a random walk in dB) over a pass, using the same [A,B] range, and check whether the 70% pulse reduction shrinks below, say, 20%; if it does, the practical claim fails to extend to the motivating dynamic scenario.

Watch

Extended reading notes

Core claim

The central claim is that averaging a tunable martingale over a prior interval of its parameter yields a concentration inequality that, for any adaptive [0,1]-valued process, satisfies P[K_{A,B}(Λ_N, μ_N) ≥ 1/ε] ≤ ε, and when the KL optimizer λ*(x,y) = ln[y(1−x)/x(1−y)] lies inside [A,B], K_{A,B}(Nx,Ny) ≥ e^{N D(x∥y)} (1+N(B−A)²/(2π))^{−1/2}. This shows the mixture bound has the same leading exponential rate as the binomial tail—the best achievable by any general non-IID bound—while adding only a constant penalty inside the exponent, and it does so without committing to a single channel estimate. Inserting these bounds into a decoy-state BB84 security analysis reduces the minimum required ro

Load-bearing premise

The headline 70% reduction is demonstrated in a simulation that treats channel mismatch as a single fixed observed loss differing from the design loss, not as a genuinely time-varying random transmittance process.

Editorial extensions

If this is right

  • Replacing Kato bounds with the mixture bounds in satellite QKD finite-key analyses lowers the pulse count needed for positive key rate, with reductions up to 71% at 10 dB loss mismatch in the paper's simulations.
  • The interval width [A,B] becomes a design knob trading robustness against channel variation against tightness when the channel is well characterized.
  • Because the bounds match the Clopper–Pearson binomial benchmark in leading exponent, they are essentially as sharp as any general non-IID bound can be.
  • The construction applies to any adaptive [0,1]-valued process, so it can be used in other quantum communication protocols with sparse non-IID data and uncertain operating conditions.
  • A union of the fixed-λ and mixture bounds can outperform Kato even when the channel model is accurate.

Reading between the lines

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

  • If the bounds remain tight under a genuinely time-varying transmittance process (e.g., log-normal fading or a random walk in dB) rather than a static mismatch, they could enable key extraction from shorter satellite overpasses; the paper's simulations only use a fixed observed loss.
  • The analytic penalty factor suggests a principled way to choose the interval: set A and B from extreme expected channel losses and verify λ* stays inside, refining the range per concentration bound rather than using a common range.
  • The martingale construction could be adapted to non-uniform priors or multi-parameter mixtures, potentially yielding even tighter adaptive bounds for channels with drift or correlated noise.
  • A natural testable extension is to run the protocol with a real fading model and measure how the 70% improvement degrades with fade depth and correlation time.
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. The paper develops concentration inequalities for adaptive [0,1]-valued processes, where the conditional probabilities p_i = E[ξ_i | F_{i-1}] may vary with the past. It introduces a mixture-martingale statistic K_{A,B}(Λ_N, μ_N) obtained by averaging fixed-λ martingale bounds over λ ∈ [A,B], proves P[K_{A,B}(Λ_N, μ_N) ≥ 1/ε] ≤ ε (Proposition 1), and shows that when the KL optimizer λ*(x,y) lies in [A,B], K_{A,B}(Nx,Ny) is at least e^{ND(x∥y)} times a polynomial factor in N(B−A)^2. The authors benchmark these bounds against Clopper-Pearson binomial inversions, claiming that the latter are uniformly tighter than any general non-IID bound. They then apply the bounds to finite-key analysis of a decoy-state BB84 satellite protocol, reporting that under 10 dB loss mismatch the mixture-martingale bounds reduce the minimum number of transmitted rounds by about 71% (LEO) and 44% (GEO) relative to Kato bounds.

Significance. If the claims hold, this is a useful contribution to finite-key QKD and, more broadly, to non-IID concentration. The martingale construction is clean, Proposition 1 and the inversion bounds in SM I–II are correct, and the closed-form analytical relaxations in SM III–IV make the method easy to insert into existing security proofs. The comparison with the binomial-tail benchmark is conceptually valuable, though as discussed below it needs a monotonicity caveat. The reported static-mismatch improvement is impressive, and the proposed union of fixed-λ and mixture bounds is a sensible practical device. The main weakness is that the headline applied claim — a >70% reduction for strongly fluctuating satellite channels — is demonstrated only for a static, mis-estimated transmittance, not for a time-varying channel; this is a significant gap between the advertised regime and the simulations.

major comments (3)
  1. [Simulations, Eq. (S131)] The abstract states that the >70% reduction is obtained in 'realistic simulations of satellite QKD with fluctuating loss', but the simulations in the main text and SM VIII model the channel as a single fixed transmittance η_sys = 10^{−L_obs/10}; no time-varying transmittance process is generated. The only stochasticity is the i.i.d. choice of intensity and basis in the idealized count formulas (S137)–(S142). The non-IID capability of the bound is therefore never exercised in the headline demonstration. Either add a genuinely time-varying channel simulation (e.g., η_i drawn from a fading model with a specified distribution and correlation over the pass) and report the threshold reduction, or amend the abstract and discussion to claim reduction under static model mismatch only. This is load-bearing because the paper's motivation is non-stationary satellite channels.
  2. [Main text, Eq. (11) and SM III] The analytical guarantee requires λ*(x,y) ∈ [A,B]. For a genuinely fluctuating process, λ* will range over an interval, and the penalty factor (1+N(B−A)^2/(2π))^{−1/2} depends on its width. The simulations set [A,B]=[0,10] and assert that λ* remains inside for all simulated cases, but they verify this only for the fixed L_obs mismatch cases. A time-varying simulation should track λ* over the channel extremes and either adapt [A,B] as suggested in the text or compute the resulting penalty explicitly. Without this, the >70% reduction is an extrapolation from a best-case static mismatch scenario.
  3. [SM V, Proposition S6] The main text claims that binomial-tail inversions are 'uniformly tighter than any one-sided bound valid for general non-IID processes'. The proposition, however, assumes U_μ and L_Λ are nondecreasing functions. Without monotonicity, the proof's event inclusion fails and the statement is not true for arbitrary one-sided bounds. Since the paper's own boundaries are monotone, the benchmark is valid for the relevant family, but the wording in the main text and the proposition statement should be qualified to 'any monotone one-sided bound'.
minor comments (5)
  1. [SM III, Eqs. (S63)–(S65)] The step from the log-quadratic Gaussian-tail bound to the final factor (1+Nw^2/(2π))^{−1/2} is compressed, and the displayed form of Eq. (S64) appears to have a typo in the denominator (it should likely be sqrt(π+4z^2) rather than π+4z^2). A short derivation would help readers verify the factor.
  2. [Simulation count formulas, Eq. (S137)] The simulated counts are deterministic expected values (S137), not random samples. This is standard practice for QKD feasibility simulations, but it should be stated explicitly in the main text so readers do not interpret the threshold reduction as arising from observed random data.
  3. [Figure 1 and SM VI] The statement that 'at this scale, the fixed-λ bound is barely distinguishable from Kato's' is specific to the plotted guess value; Fig. S1 shows visible differences at other calibration points. Please add a sentence in the main text clarifying that the comparison is for a particular guess and that behavior away from that guess is shown in the Supplemental Material.
  4. [SM VIII, Eqs. (S139)–(S140)] The variable N_Error_sift is used in Eq. (S139) before it is defined in Eq. (S140). Reorder the definitions or add a parenthetical reference.
  5. [Data availability] The code is 'available from the authors upon reasonable request'. Given the field's current reproducibility norms, making the simulation code publicly available would strengthen the paper.

Circularity Check

0 steps flagged · score 2.0 of 10

Central mixture-martingale derivation is self-contained; the only self-citations are minor and non-load-bearing.

full rationale

The main derivation chain is not circular. The paper starts from the martingale M_n(λ) defined in Eq. (2), proves it is a nonnegative mean-one martingale, applies the inequality e^{-λz} ≤ 1 - (1-e^{-λ})z and Jensen's inequality to reach K_λ in Eq. (3), and then obtains P[K_λ ≥ 1/ε] ≤ ε directly from Markov's inequality in Eq. (4). The mixture extension in Eqs. (6)-(8) integrates over a prior on λ and again uses Markov's inequality; Proposition 1 and the Supplemental Material I proof show the inversion is valid by monotonicity. Equation (11), relating K_{A,B} to the KL exponent, is derived in SM III by Taylor expansion and Gaussian-mass estimates; no fitted constant enters. The binomial benchmark is also derived in SM V by an explicit contradiction argument, not by matching the paper's curves. The simulations use [A,B]=[0,10] as a fixed, a priori integration range, not a parameter fitted to the reported 70% reduction, and the numerical mixture bound is valid regardless of whether λ* lies in that interval; the analytical guarantee merely adds a condition. The paper does cite the authors' own prior work, notably Ref. [3] for rational KL approximations and Ref. [11] for GEO feasibility, but these are background/reusable algebraic tools rather than premises that force the central bounds. The advertised 'fluctuating loss' claim is demonstrated with static loss mismatch L_exp ≠ L_obs, which is an extrapolation/scope concern, not a circularity: nothing is predicted from data that also defines the prediction. Overall, no load-bearing step reduces to its own input by construction.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The central claim depends on standard probability tools (martingales, Markov, Jensen, Tonelli) plus the adaptive Bernoulli model for QKD detection records. The only non-classical inputs are the authors' own rational-KL bounds (Ref. [3]), used for closed-form relaxations. The free parameters are the mixture interval [A,B] and the simulation/optimization settings; none are fitted to the headline reduction.

free parameters (3)
  • Mixture interval [A,B] = [0,10] in simulations
    Hand-chosen range for the uniform prior over λ. The analytical KL guarantee (Eq. 11) only holds when λ*∈[A,B]; the authors assert this holds in all simulations. Width w trades robustness against tightness.
  • Protocol optimization parameters = not listed
    Decoy intensities, intensity probabilities, basis probabilities, and beam-splitter ratio q are optimized to maximize key rate at L_exp; exact values are not tabulated, so the headline 70% numbers depend on this optimization.
  • Security/error-correction parameters = ε=10^-20/144, f_ec=1.16, Δ_PA=71, d=10^-7 or 10^-9, δ_mis=0.03 or 0.01
    Simulation inputs defining the scenario; they are standard for QKD analyses and not fitted to the bounds.
assumptions (6)
  • standard math Martingale convergence and Markov's inequality applied to nonnegative martingales with mean 1.
    Used in Proposition S1 to convert the mixture martingale tail into a probability bound.
  • standard math Convexity inequality e^{-λz} ≤ 1−(1−e^{-λ})z for z∈[0,1], and Jensen's inequality on the concave log function.
    Bridges the martingale product to the closed-form K_λ in Eq. (3).
  • standard math Tonelli's theorem for exchanging expectation and integration over λ.
    Justifies E[∫M_N(λ)dλ] = ∫E[M_N(λ)]dλ in SM II.
  • domain assumption Adaptive Bernoulli model: ξ_i∈[0,1], p_i=E[ξ_i|F_{i-1}] is the standard framework for finite-key security against coherent attacks.
    The central statistical problem (Eq. 1) rests on this model; it is standard in QKD proofs but is an assumption about how detection records behave.
  • standard math Clopper-Pearson binomial tail inversion is the pointwise-optimal uniform bound over the adaptive class (Prop S6).
    Proved in SM V; used as the benchmark for tightness.
  • standard math Rational KL approximations from Ref. [3] (same authors' prior work) and Topsøe bounds are valid for the lifted martingale context.
    Used in SM III-IV to turn the KL sufficient condition into closed-form bounds.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Quantum Key Distribution Beyond Stationary Channels." pith.science (2026). https://pith.science/paper/2NNBSZEU

@misc{pith2026260717690,
  author       = {Pith},
  title        = {Pith review of: Quantum Key Distribution Beyond Stationary Channels},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2NNBSZEU}},
  note         = {Machine review of arXiv:2607.17690}
}
abstract

Quantum key distribution (QKD) over non-stationary channels, such as satellite links, is characterized by short, high-loss, and strongly fluctuating transmission windows that produce sparse detection events. In many QKD protocols, these data must be analyzed using non-IID statistical inequalities, yet existing methods either become loose for small sample sizes or heavily rely on fine-tuning, yielding poor estimates when the optical channel is mis-modeled. Using mixture martingale techniques, we introduce tight concentration inequalities that retain sharpness when the channel model is accurate, while remaining robust to model mismatch. In realistic simulations of satellite QKD with fluctuating loss, the resulting bounds can reduce the minimum required number of transmitted signals by more than $70\%$.

Figures

Figures reproduced from arXiv: 2607.17690 by the authors.

Figure 1
Figure 1. FIG. 1. Comparison of upper bounds on [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Finite secret key rate versus number of rounds [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Finite secret key rate versus number of rounds for an [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 5 linked inside Pith

  1. [3]

    If0< x≤y <1, then the unique stationary point ofϕ x,y(λ)onRis λ∗(x, y) := ln y(1−x) x(1−y) ,(S45) and this point satisfiesλ ∗(x, y)≥0

  2. [1]

    The first and second derivatives ofϕ x,y(λ)are ϕ′ x,y(λ) =−N x+N ye−λ (1−y) +ye −λ ,(S43) ϕ′′ x,y(λ) =−Nbpλ(1−bpλ),bp λ := ye−λ (1−y) +ye −λ . (S44)

  3. [2]

    The functionϕ x,y(λ)is strictly concave on[0,∞)

  4. [4]

    N√ N+ 2a Λ −1(gµ) u√ N +a Λ −1(gµ)−b −1 aΛ −1(gµ) # , √ N+ 2a Λ −1(gµ)>0, 0,otherwise, U Λ Kato(gµ)(u) :=    ΠN

    At that maximizer, ϕx,y[λ∗(x, y)] =N D(x∥y),(S46) where D(x∥y) :=xln x y + (1−x) ln 1−x 1−y .(S47) Proof.Item 1 follows by direct differentiation. Since 0< y <1, we havebpλ ∈(0,1) for everyλ≥0, and therefore Item 2 follows from (S44). For Item 3, solvingϕ′ x,y(λ) = 0 using (S43) gives x= ye−λ (1−y) +ye −λ ,(S48) 9 which rearranges to (S45); the conditionx...

  5. [5]

    Portmann and R

    C. Portmann and R. Renner, Security in quantum cryp- tography, Rev. Mod. Phys.94, 025008 (2022)

    2022

  6. [6]

    C. C.-W. Lim, F. Xu, J.-W. Pan,et al., Security Analysis of Quantum Key Distribution with Small Block Length and Its Application to Quantum Space Communications, Phys. Rev. Lett.126, 100501 (2021)

    2021

  7. [7]

    Mannalath, V

    V. Mannalath, V. Zapatero, and M. Curty, Sharp Fi- nite Statistics for Quantum Key Distribution, Phys. Rev. Lett.135, 020803 (2025)

    2025

  8. [8]

    Liao, W.-Q

    S.-K. Liao, W.-Q. Cai, W.-Y. Liu,et al., Satellite-to- ground quantum key distribution, Nature549, 43 (2017)

    2017

Show all 50 references
  1. [9]

    J. Yin, Y. Cao, Y.-H. Li,et al., Satellite-to-Ground Entanglement-Based Quantum Key Distribution, Phys. Rev. Lett.119, 200501 (2017)

    2017

  2. [10]

    Y.-A. Chen, Q. Zhang, T.-Y. Chen,et al., An integrated space-to-ground quantum communication network over 4,600 kilometres, Nature589, 214 (2021)

    2021

  3. [11]

    Li, W.-Q

    Y. Li, W.-Q. Cai, J.-G. Ren,et al., Microsatellite- based real-time quantum key distribution, Nature640, 47 (2025)

    2025

  4. [12]

    J. S. Sidhu, T. Brougham, D. McArthur,et al., Finite key effects in satellite quantum key distribution, npj Quan- tum Inf.8, 18 (2022)

    2022

  5. [13]

    Islam, J

    T. Islam, J. S. Sidhu, B. L. Higgins,et al., Finite-Resource Performance of Small-Satellite-Based Quantum-Key-Distribution Missions, PRX Quantum5, 030101 (2024)

    2024

  6. [14]

    P. V. Trinh, A. Carrasco-Casado, H. Takenaka,et al., Statistical verifications and deep-learning predictions for satellite-to-ground quantum atmospheric channels, Com- mun. Phys.5, 225 (2022)

    2022

  7. [15]

    Mannalath, V

    V. Mannalath, V. Zapatero, and M. Curty, Finite-key feasibility of geostationary quantum key distribution, preprint arXiv:2605.29706 (2026)

    2026 arXiv

  8. [16]

    Boileau, K

    J.-C. Boileau, K. Tamaki, J. Batuwantudawe,et al., Un- conditional Security of Three State Quantum Key Dis- tribution Protocols, Phys. Rev. Lett.94, 040503 (2005)

    2005

  9. [17]

    Tamaki, N

    K. Tamaki, N. L¨ utkenhaus, M. Koashi,et al., Un- conditional security of the Bennett 1992 quantum-key- distribution scheme with strong reference pulse, Phys. Rev. A80, 032302 (2009)

    1992

  10. [18]

    Curr´ as-Lorenzo,´A

    G. Curr´ as-Lorenzo,´A. Navarrete, K. Azuma,et al., Tight finite-key security for twin-field quantum key distribu- tion, npj Quantum Inf.7, 22 (2021)

    2021

  11. [19]

    Mizutani, S

    A. Mizutani, S. Kawakami, and G. Kato, Finite-key secu- rity analysis of the decoy-state BB84 QKD with passive measurement, Quantum Sci. Technol.11, 015010 (2026)

    2026

  12. [20]

    Azuma, Weighted sums of certain dependent random variables, Tohoku Math

    K. Azuma, Weighted sums of certain dependent random variables, Tohoku Math. J. (2)19, 357 (1967)

    1967

  13. [21]

    Kato, Concentration inequality using unconfirmed knowledge, preprint arXiv:2002.04357 (2020)

    G. Kato, Concentration inequality using unconfirmed knowledge, preprint arXiv:2002.04357 (2020)

    2002 arXiv

  14. [22]

    Curr´ as-Lorenzo, ´A

    G. Curr´ as-Lorenzo, ´A. Navarrete, M. Pereira,et al., Finite-key analysis of loss-tolerant quantum key distri- bution based on random sampling theory, Phys. Rev. A 104, 012406 (2021)

    2021

  15. [23]

    Zapatero and M

    V. Zapatero and M. Curty, Finite-key security of passive quantum key distribution, Phys. Rev. Appl.21, 014018 (2024)

    2024

  16. [24]

    Curr´ as-Lorenzo, M

    G. Curr´ as-Lorenzo, M. Pereira, G. Kato,et al., Security framework for quantum key distribution with imperfect sources, Optica Quantum3, 525 (2025)

    2025

  17. [25]

    Navarrete, G

    ´A. Navarrete, G. Curr´ as-Lorenzo, M. Pereira, and M. Curty, Numerical security analysis for practical quan- tum key distribution, preprint arXiv:2605.12984 (2026)

    2026 arXiv

  18. [26]

    Liorni, H

    C. Liorni, H. Kampermann, and D. Bruß, Satellite-based links for quantum key distribution: beam effects and weather dependence, New J. Phys.21, 093055 (2019)

    2019

  19. [27]

    Scarfe, F

    L. Scarfe, F. Hufnagel, M. F. Ferrer-Garcia,et al., Fast adaptive optics for high-dimensional quantum commu- nications in turbulent channels, Commun. Phys.8, 79 (2025)

    2025

  20. [28]

    C.-Y. Lu, Y. Cao, C.-Z. Peng, and J.-W. Pan, Micius quantum experiments in space, Rev. Mod. Phys.94, 035001 (2022)

    2022

  21. [29]

    Ecker, B

    S. Ecker, B. Liu, J. Handsteiner,et al., Strategies for achieving high key rates in satellite-based QKD, npj Quantum Inf.7, 5 (2021)

    2021

  22. [30]

    Takenaka, A

    H. Takenaka, A. Carrasco-Casado, M. Fujiwara,et al., Satellite-to-ground quantum-limited communication us- ing a 50-kg-class microsatellite, Nat. Photonics11, 502 (2017)

    2017

  23. [31]

    Robbins and D

    H. Robbins and D. Siegmund, Confidence Sequences and Interminable Tests, Bull. Inst. Internat. Statist.43, 379 (1969)

    1969

  24. [32]

    Robbins, Statistical Methods Related to the Law of the Iterated Logarithm, Ann

    H. Robbins, Statistical Methods Related to the Law of the Iterated Logarithm, Ann. Math. Statist.41, 1397 (1970)

    1970

  25. [33]

    T. L. Lai, On Confidence Sequences, Ann. Statist.4, 265 (1976)

    1976

  26. [34]

    Kaufmann and W

    E. Kaufmann and W. M. Koolen, Mixture Martingales Revisited with Applications to Sequential Tests and Con- fidence Intervals, J. Mach. Learn. Res.22, 1 (2021)

    2021

  27. [35]

    J. L. W. V. Jensen, Sur les fonctions convexes et les in´ egalit´ es entre les valeurs moyennes, Acta Math.30, 175 (1906)

    1906

  28. [36]

    E. M. Stein and R. Shakarchi,Real Analysis: Measure Theory, Integration, and Hilbert Spaces, 1st ed., Prince- ton Lectures in Analysis, Vol. 3 (Princeton University Press, Princeton, NJ, 2005)

    2005

  29. [37]

    T. M. Apostol,Mathematical Analysis, 2nd ed., Addison- Wesley series in mathematics (Addison-Wesley Pub. Co., 1974)

    1974

  30. [38]

    A. S. Kechris,Classical Descriptive Set Theory, Graduate Texts in Mathematics, Vol. 156 (Springer New York, New York, 1995)

    1995

  31. [39]

    Klenke,Probability Theory: A Comprehensive Course, Universitext (Springer London, London, 2008)

    A. Klenke,Probability Theory: A Comprehensive Course, Universitext (Springer London, London, 2008)

    2008

  32. [40]

    Mastin and P

    A. Mastin and P. Jaillet, Log-Quadratic Bounds for the Gaussian Q-function (2013), preprint arXiv:1304.2488 [math.PR]

    2013 arXiv

  33. [41]

    Tonelli, Sull’integrazione per parti, Atti Accad

    L. Tonelli, Sull’integrazione per parti, Atti Accad. Naz. Lincei, Ser. 518, 246 (1909)

    1909

  34. [42]

    Chernoff, A Measure of Asymptotic Efficiency for Tests of a Hypothesis Based on the Sum of Observations, Ann

    H. Chernoff, A Measure of Asymptotic Efficiency for Tests of a Hypothesis Based on the Sum of Observations, Ann. Math. Statist.23, 493 (1952)

    1952

  35. [43]

    Topsøe, Some bounds for the logarithmic function, Inequality Theory and Applications4, 137 (2007)

    F. Topsøe, Some bounds for the logarithmic function, Inequality Theory and Applications4, 137 (2007)

    2007

  36. [44]

    C. J. Clopper and E. S. Pearson, The Use of Confidence or Fiducial Limits Illustrated in the Case of the Binomial, Biometrika26, 404 (1934). 22

    1934

  37. [45]

    R. R. Bahadur and R. Ranga Rao, On Deviations of the Sample Mean, Ann. Math. Statist.31, 1015 (1960)

    1960

  38. [46]

    Arratia and L

    R. Arratia and L. Gordon, Tutorial on Large Deviations for the Binomial Distribution, Bulletin of Mathematical Biology51, 125 (1989)

    1989

  39. [47]

    H. Zhu, Z. Li, and M. Hayashi, Nearly tight univer- sal bounds for the binomial tail probabilities, preprint arXiv:2211.01688 (2022)

    2022 arXiv

  40. [48]

    Koashi, Simple security proof of quantum key dis- tribution based on complementarity, New J

    M. Koashi, Simple security proof of quantum key dis- tribution based on complementarity, New J. Phys.11, 045018 (2009)

    2009

  41. [49]

    Avesani, L

    M. Avesani, L. Calderaro, M. Schiavon,et al., Full day- light quantum-key-distribution at 1550 nm enabled by in- tegrated silicon photonics, npj Quantum Inf.7, 93 (2021)

    2021

  42. [50]

    Navarrete and M

    ´A. Navarrete and M. Curty, Improved finite-key secu- rity analysis of quantum key distribution against Trojan- horse attacks, Quantum Sci. Technol.7, 035021 (2022)

    2022

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.