Pith. sign in

REVIEW 3 major objections 4 minor 31 references

Tight bounds on recurrence time in closed quantum systems

T0 review · 3 major / 4 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read For any closed quantum system, the first recurrence time is bounded by the escape time times (4π/ε)^{d-1}, and random Hamiltonians saturate it.

desk verdict Theorems 1 and 2 give a clean, deterministic upper bound on recurrence time and are worth publishing; the claimed generic saturation is not established by the Supplemental proof, which yields only ε^{-(d-3)} rather than ε^{-d}. read the letter →

arxiv 2601.10409 v3 pith:JOC632NQ submitted 2026-01-15 quant-ph math-phmath.MP

Marcin Kotowski , Micha{l} Oszmaniec This is my paper · ORCID
classification quant-phmath-phmath.MP
keywords quantumrecurrencePoincaréspeedlimitpackingnumberescapetimeeffectivedimensionrandomHamiltonianunitarydynamics
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 proves a quantitative, deterministic upper bound on how long a closed quantum system takes to return near its initial state. The first recurrence time t_rec(ε) never exceeds t_exit(ε)(4π/ε)^{d-1}, where d is the Hilbert-space dimension, ε the size of the return neighborhood, and t_exit the time the system first leaves that neighborhood. The authors turn bounding t_exit into an 'inverse quantum speed limit' and show that, below a threshold set by the Hamiltonian's variance and fourth moment, t_exit ≤ ε/√Δ(H²) up to a correction factor. They also show that for random diagonal Hamiltonians with a uniform initial superposition the (1/ε)^d growth is saturated, so the general bound cannot be improved in its dependence on d and ε. Versions with effective support and with non-interacting particle Hamiltonians improve the exponent.

What carries the argument

The load-bearing geometric fact is that unitary evolution is an isometry on the state space. Proposition 1 shows that in any metric space with an isometric flow, the recurrence time is at most the exit time times the packing number N_pack(ε) — the largest number of points mutually ε-apart. Lemma 1 bounds this packing number for the orbit-reachable set T_{ψ0} (the torus of relative phases of the initial state) by (4π/ε)^{d-1}, using an explicit ε-net on the phase coordinates. The escape-time estimate follows from the identity F''(t)=tr(X_H ψ_t) with X_H=−[H,[H,ψ0]], together with the operator bound ∥X_H∥_∞ ≤ Δ(H²)+√Δ(H⁴); integrating F'' over the interval before escape yields the inverse quan

What would settle it

Search numerically for a violation of Eq. (3): for d=4 or 5, take a Hamiltonian with rationally independent eigenvalues and a generic initial state, and compute t_rec(ε) by dense time sampling; if for any ε < ε*/2 the observed first recurrence time exceeds t_exit(ε)(4π/ε)^{d-1}, the upper bound is false. To test the saturation claim, simulate many random diagonal Hamiltonians with uniform eigenvalues and uniform initial state for d=6 and ε=10^{-3}; if the typical t_rec falls below the theorem's claimed (1/ε)^d lower bound (with its pre-factors), the high-probability lower bound is false.

Watch

Extended reading notes

Core claim

For any Hamiltonian H and any initial pure state ψ0, the first time the evolved state returns to within trace distance ε of ψ0 (after first leaving that neighborhood) obeys t_rec(ε) ≤ t_exit(ε)·(4π/ε)^{d-1}, with d the Hilbert-space dimension and t_exit the first exit time. Treating t_exit as an inverse quantum speed limit, the paper shows that for ε < ε* = Δ(H²)/(Δ(H²)+√Δ(H⁴)), t_exit(ε) ≤ (ε/√Δ(H²))·(1−ε/ε*)^{-1/2}; hence for ε < ε*/2, t_rec(ε) ≤ (ε/(2√Δ(H²)))·(4π/ε)^{d-1}. For random diagonal Hamiltonians with independent uniform eigenvalues and a uniform superposition initial state, t_rec grows as (1/ε)^d with high probability, so the exponent cannot be improved in general.

Load-bearing premise

The upper bound requires only that the exit time be finite and that ε stay below ε* = Δ(H²)/(Δ(H²)+√Δ(H⁴)), which can fail for states nearly supported on one eigenstate; the tightness claim further assumes that random diagonal Hamiltonians with iid uniform eigenvalues and a uniform eigenbasis superposition represent 'generic' Hamiltonians — a transfer not proven for local or structured Hamiltonians.

Editorial extensions

If this is right

  • First recurrences occur no later than (exit time)·(4π/ε)^{d-1}; with Theorem 2 this yields the explicit bound t_rec ≤ (ε/(2√Δ(H²)))(4π/ε)^{d-1} for ε < ε*/2.
  • If ψ0 has δ-effective support d_supp(δ), the exponent improves to d_supp−1, giving t_rec ≤ t_exit·(8π/ε)^{d_supp(ε²/4)−1}; the same argument works for subsequent (kth) recurrences.
  • For non-interacting n-particle Hamiltonians, the single-particle dimension d replaces the full Hilbert-space dimension: t_rec ≤ t_exit·(4πn/ε)^{d−1}.
  • Random diagonal Hamiltonians with uniform initial superpositions saturate the (1/ε)^d growth with probability 1−o(1), so the general bound is tight in its dependence on d and ε.
  • Effective dimension is not the right measure: states with small effective dimension (∼log d) but large effective support (∼d) still have recurrence times exponential in d, so it is the effective support that controls the lower bound.

Reading between the lines

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

  • The packing-number argument is purely metric, so the same t_rec ≤ t_exit·N_pack bound should apply to any distance-invariant dynamics, including unitary channels under the diamond norm (which the paper treats) and possibly classical Hamiltonian flows on compact phase spaces.
  • For physical, local Hamiltonians the random-diagonal model is not a faithful proxy; the paper leaves open whether local non-commuting Hamiltonians recur faster, and the authors list this as an open problem. A numerical study of small spin chains could probe this gap.
  • The recurrence-time exponent (1/ε)^d is astronomical for macroscopic d; an immediate consequence is that observing recurrence in a many-body experiment requires engineering an effective low-dimensional subspace, which the effective-support version makes quantitative.
  • The threshold ε* depends on the initial state's variance and fourth moment; for states close to an eigenstate ε* becomes small, so the proven bound is vacuous for tiny ε — a refined treatment of that regime would be a natural next step.
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

3 major / 4 minor

Summary. The paper studies recurrence times for unitary evolutions of pure states in finite-dimensional Hilbert spaces. It proves a deterministic upper bound, t_rec(ε) ≤ t_exit(ε)·(4π/ε)^{d−1} (Theorem 1), where t_exit is the first time the evolution leaves the ε-neighborhood of the initial state. It then provides an upper bound on t_exit under a moment condition: t_exit(ε) ≤ ε/√Δ(H²) · (1−ε/ε*)^{-1/2} for ε<ε* (Theorem 2). Combining these gives an explicit bound t_rec(ε) = O(ε^{−(d−2)}) up to constants. The paper also claims a matching generic lower bound for random diagonal Hamiltonians (Theorem 3, Theorem S2): t_rec ≳ (1/ε)^d with high probability. Additional results concern effective support, non-interacting dynamics, unitary channels, and the role of effective dimension.

Significance. Theorems 1 and 2 are clean, self-contained, and appear correct; they give a rigorous deterministic upper bound on recurrence time that improves on Peres's heuristic argument and applies to arbitrary initial states and Hamiltonians. This is a valuable contribution. The claimed generic saturation, however, is a central part of the paper's advertised message ('tight bounds'), and the proof of that claim contains a concrete union-bound error and does not, as written, establish the stated lower bound. If the saturation claim can be repaired or appropriately weakened, the upper-bound part alone would still be a solid contribution; the present version overstates what is proven.

major comments (3)
  1. [Supplemental Material, Eqs. (83)–(88), Theorem S2] The union bound over grid points applies Proposition S5 with ε replaced by 2ε, but the factor 2^d is dropped. Writing (68) with ε→2ε gives P(D<2ε) ≤ (2π/ε²)(2Cε)^d, not (2π/ε²)(Cε)^d. Consequently the admissible T0 should be reduced by a factor 2^d. With the stated choice of T0 in Eq. (88), the claimed inequality T0 > 20(1/(10000ε))^d fails for the allowed parameter range (e.g., d=50, ε=5×10^{-5}). Theorem 3 and the abstract's 'generically saturated' claim are therefore not established by the proof as written.
  2. [§V, Eq. (17)–(18)] The effective-support reduction uses δ=ε²/4, but the displayed inequality D_tr(ψ0,ψt) ≤ D_tr(ψS0,ψSt)+√(2δ) is too weak: if D_tr(ψS0,ψSt)≤ε/2, it only gives D_tr(ψ0,ψt)≤(1/2+1/√2)ε > ε. With this displayed inequality one needs δ≤ε²/8. The argument can be rescued by instead using the tighter bound D_tr(ψ0,ψt) ≤ √(D_tr² + 2δ), but as written the derivation of (18) is incorrect.
  3. [Theorem 3 / SM Theorem S2, general tightness claim] Even apart from the missing 2^d, the proof's union bound yields a lower-bound horizon T0 of order ε^{3−d} up to constants, i.e. ε^{−(d−3)} for fixed d, whereas the upper bound from Theorems 1+2 is of order ε^{−(d−2)} (or ε^{−(d−1)} before using t_exit≈ε). The lower bound is therefore far smaller than the upper bound in the parameter range covered by Theorem S2, so the phrase 'generically saturated' is not supported by the quantitative estimates. The authors should either strengthen the lower-bound argument or weaken the saturation claim.
minor comments (4)
  1. [General] Typos and language issues: 'lack of of rigorous', 'arbirtary', 'theorem' for 'Theorem', and a few other grammatical errors should be corrected.
  2. [Eq. (8) and Lemma 2] There is a sign inconsistency in the definition of X_H: Theorem 2 defines X_H := −[H,[H,ψ0]], while Lemma 2 writes X_H = H²ψ0 − 2Hψ0H + ψ0H², which is the positive commutator. The norm bound is unaffected, but the notation should be made consistent.
  3. [Main text, Theorem 3] The statement 'for ε small enough' is imprecise. The SM theorem requires ε > (40π)^{1/3}(1/2)^{d/3}, so for fixed d, ε cannot be taken arbitrarily small; the parameter regime should be stated clearly in the main text.
  4. [Supplemental Material, proof of Theorem S2] The grid spacing is ε and the Lipschitz step uses a threshold 2ε; this is fine, but the choice of T0 in Eq. (88) should be derived after correctly including the 2^d factor from the ε→2ε substitution. The numerical constants in Theorem S2 should be rechecked.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: upper bound from packing/geometric arguments, lower bound from independent random Hamiltonian model.

full rationale

The derivation chain is self-contained. Theorem 1 follows from Proposition 1 (metric invariance plus packing numbers) and Lemma 1 (an explicit covering of the torus of relative phases); neither step assumes the conclusion. Theorem 2 is obtained by integrating the second derivative of fidelity, with Lemma 2 bounding ||X_H||_∞ algebraically in terms of Hamiltonian moments; t_exit is an independently defined quantity and is not fitted to t_rec. The random lower bound in the Supplemental Material is a separate probabilistic model: it concentrates ⟨H^k⟩, uses Theorem 2 only to sandwich t_exit, and then uses a geometric lattice/probability estimate (Proposition S5) plus a union bound. No parameter is fitted to data and then renamed as a prediction. The only apparent self-citation, [13], appears in background discussion of circuit complexity and is not load-bearing; [30] is cited for a standard diamond-norm inequality, an external parameter-free fact. The skeptical concern about the exponent/union-bound gap in the proof of Theorem S2 is a correctness issue about whether the stated lower bound is proven, not a circularity, because the lower-bound argument does not assume the upper-bound result it is compared with.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

The paper introduces no fitted parameters and no new entities. It relies on standard metric-space, probabilistic, and Diophantine facts; the random Hamiltonian model is a domain assumption used only for the lower-bound claim.

assumptions (7)
  • domain assumption Finite-dimensional Hilbert space and unitary Schrödinger evolution
    All theorems are for finite d and closed dynamics; recurrence definition relies on unitary isometries.
  • standard math Trace distance identity D_tr = sqrt(1-F) for pure states
    Used throughout proofs of Theorems 1 and 2.
  • standard math Packing/covering relation N_cov(X,2ε) ≤ N_pack ≤ N_cov(X,ε)
    Used in Lemma 1 and Lemma S4; cited to Vershynin.
  • standard math An isometry of a compact metric space is invertible
    Needed in Proposition 1 to shift indices via φ^{-i}.
  • standard math Hoeffding's inequality and volume estimates for random eigenvalues
    Used in the probability bounds of Theorems S2 and S3.
  • domain assumption Rational independence / Kronecker density of the phase torus
    Proposition S9 assumes rationally independent eigenvalues to characterize finite t_exit.
  • domain assumption Random Hamiltonian model: eigenvalues iid uniform on [-1,1]
    Used for the lower-bound tightness theorem; not justified as physical for local Hamiltonians.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Tight bounds on recurrence time in closed quantum systems." pith.science (2026). https://pith.science/paper/JOC632NQ

@misc{pith2026260110409,
  author       = {Pith},
  title        = {Pith review of: Tight bounds on recurrence time in closed quantum systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JOC632NQ}},
  note         = {Machine review of arXiv:2601.10409}
}
abstract

The evolution of an isolated quantum system inevitably exhibits recurrence: the state returns to the vicinity of its initial condition after finite time. Despite its fundamental nature, a rigorous quantitative understanding of recurrence has been lacking. We establish upper bounds on the recurrence time, $t_{\mathrm{rec}} \lesssim t_{\mathrm{exit}}(\epsilon)(1/\epsilon)^d$, where $d$ is the Hilbert-space dimension, $\epsilon$ the neighborhood size, and $t_{\mathrm{exit}}(\epsilon)$ the escape time from this neighborhood. For pure states evolving under a Hamiltonian $H$, estimating $t_{\mathrm{exit}}$ is equivalent to an inverse quantum speed limit problem: finding upper bounds on the time a time-evolved state $\psi_t$ needs to depart from the $\epsilon$-vicinity of the initial state $\psi_0$. We provide a partial solution, showing that under mild assumptions $t_{\mathrm{exit}}(\epsilon) \approx \epsilon /\sqrt{ \Delta(H^2)}$, with $\Delta(H^2)$ the Hamiltonian variance in $\psi_0$. We show that our upper bound on $t_{\mathrm{rec}}$ is generically saturated for random Hamiltonians. Finally, we analyze the impact of coherence of the initial state in the eigenbasis of $H$ on recurrence behavior.

Figures

Figures reproduced from arXiv: 2601.10409 by the authors.

Figure 1
Figure 1. The Hamiltonian evolution of a closed quan [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Main part of the proof of Proposition 1. If the number of points xk = φ k (x0) is greater than the packing number Npack(X, ε), there exists a point xj−i within distance ε of x0. The ball of radius ε around x0 is shown in red. some 1 ≤ k ≤ Npack(X, ε) we have: d(x0, φktexit (x0)) < ε (26) so a recurrence has occurred at time t = ktexit, which is indeed a recurrence since t ≥ texit. For an initial state ψ0 let Tψ0 be … view at source ↗

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 9 linked inside Pith

  1. [1]

    We will denote the variance⟨(H− ⟨H⟩)2⟩as ∆(H 2) and the fourth cen- tral moment⟨(H− ⟨H⟩)4⟩as ∆(H 4)

    For an operatorAwe will denote the expected value tr(ψ 0A) simply as⟨A⟩. We will denote the variance⟨(H− ⟨H⟩)2⟩as ∆(H 2) and the fourth cen- tral moment⟨(H− ⟨H⟩)4⟩as ∆(H 4) . Consider a metric space (X, d) and suppose we have a time evolutionx t starting from an initial pointx 0. We define theexit timet exit(ε) on scaleε as the first time when the time ev...

    2023

  2. [2]

    Poincar´ e, Sur le probl` eme des trois corps et les ´ equations de la dynamique, Acta Mathematica13, 1 (1890)

    H. Poincar´ e, Sur le probl` eme des trois corps et les ´ equations de la dynamique, Acta Mathematica13, 1 (1890)

  3. [3]

    I. P. Cornfeld, S. V. Fomin, and Y. G. Sinai, Ergodic Theory, Grundlehren der mathematischen Wissenschaften, Vol. 245 (Springer-Verlag, 1982)

    1982

  4. [4]

    Bocchieri and A

    P. Bocchieri and A. Loinger, Quantum recurrence theorem, Phys. Rev.107, 337 (1957)

    1957

  5. [5]

    Peres, Recurrence phenomena in quantum dy- namics, Phys

    A. Peres, Recurrence phenomena in quantum dy- namics, Phys. Rev. Lett.49, 1118 (1982)

    1982

  6. [7]

    Riddell, N

    J. Riddell, N. Pagliaroli, and A. Alhambra, Concen- tration of quantum equilibration and an estimate of the recurrence time, SciPost Physics15(2023)

    2023

  7. [8]

    Karimi, X

    B. Karimi, X. Wu, A. N. Cleland, and J. P. Pekola, A blueprint for experiments exploring the poincar´ e quantum recurrence theorem (2025), arXiv:2508.13489 [quant-ph]

    arXiv 2025

  8. [9]

    Kac, On the distribution of values of trigono- metric sums with linearly independent frequencies, American Journal of Mathematics65, 609 (1943)

    M. Kac, On the distribution of values of trigono- metric sums with linearly independent frequencies, American Journal of Mathematics65, 609 (1943)

    1943

Show all 31 references
  1. [10]

    Rauer, S

    B. Rauer, S. Erne, T. Schweigler, F. Catal- dini, M. Tajik, and J. Schmiedmayer, Re- currences in an isolated quantum many- body system, Science360, 307 (2018), https://www.science.org/doi/pdf/10.1126/science.aan7938

    2018 doi

  2. [11]

    D’Alessio, Y

    L. D’Alessio, Y. Kafri, A. Polkovnikov, and M. Rigol, From quantum chaos and eigenstate ther- malization to statistical mechanics and thermo- dynamics, Advances in Physics65, 239 (2016), https://doi.org/10.1080/00018732.2016.1198134

    2016

  3. [12]

    A. R. Brown and L. Susskind, Second law of quan- tum complexity, Phys. Rev. D97, 086015 (2018), arXiv:1701.01107 [hep-th]

    2018 arXiv

  4. [13]

    Baiguera, V

    S. Baiguera, V. Balasubramanian, P. Caputa, S. Chapman, J. Haferkamp, M. P. Heller, and N. Yunger Halpern, Quantum complexity in grav- ity, quantum field theory, and quantum information science, arXiv e-prints , arXiv:2503.10753 (2025), arXiv:2503.10753 [hep-th]

    2025

  5. [14]

    Oszmaniec, M

    M. Oszmaniec, M. Kotowski, M. Horodecki, and N. Hunter-Jones, Saturation and recurrence of quantum complexity in random local quantum dy- namics, Phys. Rev. X14, 041068 (2024)

    2024

  6. [15]

    Susskind, Black Holes at Exp-time, arXiv e-prints , arXiv:2006.01280 (2020), arXiv:2006.01280 [hep- th]

    L. Susskind, Black Holes at Exp-time, arXiv e-prints , arXiv:2006.01280 (2020), arXiv:2006.01280 [hep- th]

    2006 arXiv

  7. [16]

    Mandelstam and I

    L. Mandelstam and I. Tamm, The uncertainty re- lation between energy and time in non-relativistic quantum mechanics, Journal of Physics (USSR)9, 249 (1945), reprinted in Selected Papers, edited by B. M. Bolotovskii et al. (Moscow: Nauka, 1972)

    1945

  8. [17]

    Campos Venuti, The recurrence time in quantum mechanics, arXiv e-prints , arXiv:1509.04352 (2015), arXiv:1509.04352 [quant-ph]

    L. Campos Venuti, The recurrence time in quantum mechanics, arXiv e-prints , arXiv:1509.04352 (2015), arXiv:1509.04352 [quant-ph]

    2015 arXiv

  9. [18]

    P. L. Krapivsky, J. M. Luck, and K. Mallick, Quantum return probability of a system of N non- interacting lattice fermions, Journal of Statisti- cal Mechanics: Theory and Experiment2, 023104 (2018), arXiv:1710.08178 [cond-mat.mes-hall]

    2018 arXiv

  10. [19]

    Laba and Tkachuk, Geometric characteristics of quantum evolution: curvature and torsion, Condensed Matter Physics20, 13003 (2017), arXiv:1006.4447 [quant-ph]

    2017 arXiv

  11. [20]

    R. J. Banks, E. Haque, F. Nazef, F. Fethallah, F. Ruqaya, H. Ahsan, H. Vora, H. Tahir, I. Ahmad, I. Hewins, I. Shah, K. Baranwal, M. Arora, M. Asad, M. Khan, N. Hasan, N. Azad, S. Fedaiee, S. Majeed, S. Bhuyan, T. Tarannum, Y. Ali, D. E. Browne, and P. A. Warburton, Continuous...

    2024

  12. [21]

    Blais, C

    E. Blais, C. L. Canonne, and T. Gur, Distribution testing lower bounds via reductions from communi- cation complexity, ACM Transactions on Computa- tion Theory11, 10.1145/3316374 (2019)

    2019 doi

  13. [22]

    Reimann, Foundation of statistical mechanics un- der experimentally realistic conditions, Phys

    P. Reimann, Foundation of statistical mechanics un- der experimentally realistic conditions, Phys. Rev. Lett.101, 190403 (2008)

    2008

  14. [23]

    Linden, S

    N. Linden, S. Popescu, A. J. Short, and A. Win- ter, Quantum mechanical evolution towards thermal equilibrium, Phys. Rev. E79, 061103 (2009)

    2009

  15. [24]

    A. J. Short and T. C. Farrelly, Quantum equili- bration in finite time, New Journal of Physics14, 5 013063 (2012)

    2012

  16. [25]

    E. J. Torres-Herrera and L. F. Santos, Quench dy- namics of isolated many-body quantum systems, Phys. Rev. A89, 043620 (2014), arXiv:1310.5153 [cond-mat.stat-mech]

    2014 arXiv

  17. [26]

    E. J. Torres-Herrera, M. Vyas, and L. F. Santos, General features of the relaxation dynamics of in- teracting quantum systems, New Journal of Physics 16, 063010 (2014), arXiv:1402.3299 [cond-mat.stat- mech]

    2014 arXiv

  18. [27]

    ´A. M. Alhambra, A. Anshu, and H. Wilming, Re- vivals imply quantum many-body scars, Phys. Rev. B101, 205107 (2020), arXiv:1911.05637 [quant-ph]

    2020 arXiv

  19. [28]

    She and H

    A. She and H. Yuen, Unitary Property Testing Lower Bounds by Polynomials, in14th Innova- tions in Theoretical Computer Science Conference (ITCS 2023), Leibniz International Proceedings in Informatics (LIPIcs), Vol. 251, edited by Y. Tau- man Kalai (Schloss Dagstuhl – Leibniz-Z...

    2023

  20. [29]

    M. A. Nielsen and I. L. Chuang,Quantum Compu- tation and Quantum Information, 10th ed. (Cam- bridge University Press, Cambridge, 2010)

    2010

  21. [30]

    D. C. Brody and L. P. Hughston, Geometry of quan- tum statistical inference, Phys. Rev. Lett.77, 2851 (1996)

    1996

  22. [31]

    Oszmaniec, A

    M. Oszmaniec, A. Sawicki, and M. Horodecki, Epsilon-nets, unitary designs, and random quantum circuits, IEEE Trans. Inf. Theory68, 989 (2022), arXiv:2007.10885 [quant-ph]

    2022 arXiv

  23. [32]

    R. Vershynin,High-Dimensional Probability: An In- troduction with Applications in Data Science, Cam- bridge Series in Statistical and Probabilistic Math- ematics (Cambridge University Press, 2018). 6 Appendix For a metric space (X, d) we formally define packing and covering nu...

    2018

Pith tools

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