Pith. sign in

REVIEW 1 major objections 4 minor 58 references

Convergence of Gaussian Process Regression with Estimated Hyper-parameters and Applications in Bayesian Inverse Problems

T0 review · 1 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Gaussian process regression with hyper-parameters estimated from data converges to the target function at the same rate as with fixed hyper-parameters, provided the estimates stay bounded, and the resulting error in Bayesian inverse…

desk verdict Solid, carefully written extension of GP emulator convergence to the empirical Bayes setting; the mean theorem holds, but the variance-rate proof has a real gap that needs repair. read the letter →

arxiv 1909.00232 v3 pith:JST3TNVD submitted 2019-08-31 math.NA cs.NAmath.STstat.TH

Aretha L Teckentrup This is my paper · ORCID
classification math.NAcs.NAmath.STstat.TH MSC 62G0862J0765D1565D4065J22
keywords GaussianprocessregressionhyperparameterestimationempiricalBayesconvergenceratesMatérnkernelssparsegridsBayesianinverseproblemsposteriorconsistency
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 asks whether Gaussian process regression still converges to the target function when the kernel's hyper-parameters are not fixed in advance but learned from the same data, the empirical Bayes setting. The answer it defends is yes: for Matérn and separable Matérn kernels, the predictive mean converges to any sufficiently smooth deterministic function $f$ as the number of design points grows, and the predictive variance contracts to zero, under very mild assumptions on the estimated hyper-parameters. When the estimates converge to a limiting value, the rate is exactly the rate one would get with those hyper-parameters fixed, so parameter learning does not degrade asymptotic accuracy. The same estimates bound the Hellinger error between the true Bayesian posterior of an inverse problem and the posterior built from the emulator, justifying the common practice of replacing an expensive forward model by a GP surrogate.

What carries the argument

The load-bearing object is the identification of the kernel's native space with a Sobolev space: for Matérn kernels with smoothness $\nu$, the native space is $H^{\nu+d_u/2}(U)$ with equivalent norms, and for separable Matérn kernels it is the tensor-product Sobolev space. On top of that sit two interpolation tools: the minimal-norm property, which identifies the predictive mean as the native-space interpolant of the data, and sampling inequalities that bound the interpolation error in $H^\beta$ by powers of the fill distance $h_{D_N,U}$ and the mesh ratio $\rho_{D_N,U}$. The proof of Theorem 3.5 combines these with a triangle-inequality decomposition of the predictive mean into an interpolation of $f$ and an interpolation of the prior mean $m$, then uses compactness of the hyperparameter set to make all constants uniform in $N$. The final bound separates the role of the true smoothness $\tilde\tau$ from that of the estimated smoothness $\tau(\hat\theta_N)$: the smaller one sets the rate in $h$, while any overshoot of $\tilde\tau$ by $\tau_+$ is penalised by a power of the mesh ratio.

What would settle it

Compute maximum-likelihood or cross-validation hyperparameter estimates for a Matérn GP emulator of a fixed smooth function, for example $f(u)=\sin(2\pi u)$ on $[0,1]$, on a sequence of nested design sets, and check whether $\hat\lambda_N$ or $\hat\sigma^2_N$ leaves any pre-specified compact interval as $N\to\infty$; if it does, or if the observed rate of $\|f-m^f_N(\hat\theta_N)\|_{L^2}$ drops below the theorem's prediction or fails to converge, that is direct evidence against the compactness-based claim. A sharper test uses a deliberately non-quasi-uniform design, points clustered at one end of the interval, with overestimated smoothness $\tau_+>\tilde\tau$, where the mesh-ratio term $\rho_{D_N,U}^{\tau_+-\tilde\tau}$ should grow and, for strong clustering, the bound and convergence should fail.

Watch

Extended reading notes

Core claim

The central result is Theorem 3.5: for a deterministic $f \in H^{\tilde\tau}(U)$ emulated with a Matérn kernel and estimated hyper-parameters $\hat\theta_N$ confined to a compact set, the predictive mean satisfies $$\|f - m^f_N(\hat\theta_N)\|_{H^\$\beta$(U)} \le C\, h_{D_N,U}^{\min\{\tilde\tau,\tau_-\}-\$\beta$}\, \rho_{D_N,U}^{\max\{\tau_+-\tilde\tau,0\}} \left(\|f\|_{$H^{{\tilde\tau}}$(U)} + \sup_{N\ge N^*}\|m(\hat\theta_N)\|_{$H^{{\tilde\tau}}$(U)}\right),$$ where $h_{D_N,U}$ is the fill distance of the design points, $\rho_{D_N,U}$ is the mesh ratio, $\tau_-$ and $\tau_+$ are the infimum and supremum of the estimated smoothness for large $N$, and $\beta \le \tilde\tau$. Since the fill distance decays like $N^{-1/d_u}$ for space-filling designs, this gives convergence in $N$ whenever $\tau_-$ has integer part larger than $d_u/2$. Theorems 3.8, 3.11, and 3.12 transfer the same conclusion to the predictive variance and to separable Matérn kernels on Smolyak sparse grids, where the rate is governed by mixed regularity and the dimension enters only through a logarithmic factor. Theorems 5.1 and 5.2 then bound the Hellinger distance between the true posterior and the emulator-based posterior by the GP predictive error, so all these rates carry over to Bayesian inverse problems.

Load-bearing premise

The load-bearing premise is that the estimated hyper-parameters remain in a fixed compact set, with the estimated smoothness never falling below a dimension-dependent threshold; if the estimates drift to the boundary, say correlation length or variance going to zero or infinity, the stated rates and even convergence are not guaranteed.

Editorial extensions

If this is right

  • Even if hyper-parameters are not identifiable and the estimates do not converge, the GP emulator still converges to $f$ as $N\to\infty$, provided the estimates stay in a compact set; this removes the need for a 'true' hyper-parameter value.
  • If $\hat\theta_N\to\theta_0$, the rate matches the fixed-hyperparameter rate exactly, so empirical Bayes does not slow down the emulator asymptotically.
  • For functions of mixed smoothness, separable Matérn kernels on sparse grids give rates that are essentially independent of dimension, up to logarithmic factors, unlike the $N^{-\tilde\tau/d_u}$ curse of dimensionality for isotropic kernels.
  • Overestimating the smoothness of $f$ is harmless for quasi-uniform designs but can destroy convergence for designs whose mesh ratio grows; underestimating smoothness only slows the rate.
  • In Bayesian inverse problems, the Hellinger distance between the true posterior and the emulator-based posterior is bounded by the GP predictive error, tying emulator convergence directly to posterior convergence.

Reading between the lines

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

  • A practical reading of the compactness assumption is that unconstrained hyperparameter optimisation is the main risk; adding a prior or box constraint that keeps the length scale and variance away from zero and infinity is not a mere regularisation convenience but is what makes the convergence theorem applicable.
  • The rate expression suggests an adaptive procedure: estimate $\nu$ (or the per-dimension $\nu_j$) and choose the kernel smoothness closest to the estimated regularity of $f$, since both under- and over-estimation degrade the exponent; this could be tested on functions with known Sobolev or mixed regularity.
  • The posterior bounds likely extend to non-Gaussian noise models and to log-likelihoods with the same Sobolev regularity, because the proof uses only smoothness of the misfit functional; the paper notes the Gaussian-noise assumption is for presentation only.
  • Combining these bounds with an asymptotic theory for maximum-likelihood estimates of hyper-parameters, for example for the marginal variance, would turn the convergence statement into a fully data-driven rate with explicit constants.
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

1 major / 4 minor

Summary. The paper studies the convergence of Gaussian process regression when the hyperparameters of the kernel and mean function are estimated from data in an empirical-Bayes (plug-in) fashion. For Matérn and separable Matérn kernels, it proves rates of convergence for the predictive mean and predictive variance as the number of design points grows, under assumptions that the estimated hyperparameters remain in a fixed compact set and that the effective kernel smoothness has a positive margin above du/2. The rates recover the fixed-hyperparameter rates when the estimates converge to a fixed value. The results are then applied to Bayesian inverse problems, where the GP emulator replaces the forward map or the log-likelihood, and bounds on the Hellinger distance between true and approximate posteriors are stated in Theorems 5.1 and 5.2.

Significance. If the results hold, this is a significant contribution: it provides a theoretical justification for the common practice of tuning GP hyperparameters from data, showing that, under mild boundedness assumptions, the asymptotic convergence of the emulator is not degraded by the estimation step. The proof of the main mean-convergence theorem (Theorem 3.5) is careful, tracks the constants, and builds on external scattered-data approximation results rather than on the author's own prior work. The application to posterior consistency in Bayesian inverse problems is valuable and goes beyond existing spatial-statistics results. The main caveat is a gap in the proof of the variance theorems, which is likely repairable and does not appear to invalidate the central conclusions.

major comments (1)
  1. [§3.1.2, proof of Theorem 3.8 (and §3.2.2, Theorem 3.12)] The proof of Theorem 3.8 reduces the variance to a supremum over g in the unit ball of H_{k(θ̂N)}(U) of ‖g − m_N^g(θ̂N)‖_{H^{du/2+ε}(U)}, and then invokes Theorem 3.5 together with the inequality ‖g‖_{H^{τ̃}(U)} ≤ C_up(θ̂N)‖g‖_{H_{k(θ̂N)}(U)}. This inequality is not a consequence of Proposition 3.3 unless τ̃ = τ(θ̂N). A function g in the unit ball of H_{k(θ̂N)}(U) is only guaranteed to lie in H^{τ(θ̂N)}(U); when τ̃ > τ(θ̂N), such g need not belong to H^{τ̃}(U) at all, so Theorem 3.5 cannot be applied with smoothness τ̃. The same gap appears in the proof of Theorem 3.12. This is load-bearing because the resulting bound on sup_u k_N(θ̂N; u, u) is used in assumption (b) of Theorem 5.2 for the posterior-consistency claims. The gap is repairable: since g is in the native space of the kernel k(θ̂N), one can use the fixed-hyperparameter native-space error estimate directly, giving ‖g − m_N^g(θ̂N)‖_{H^{du/2+ε}(U)} ≤ C(θ̂N) h^{τ(θ̂N)−du/2−ε} ‖g‖_{H_{k(θ̂N)}(U)} with C(θ̂N) uniform on S; the advertised h-rate min{τ̃,τ_-}−du/2−ε then follows, and this route in fact avoids the mesh-ratio factor. The authors should rewrite this proof.
minor comments (4)
  1. [Theorem 3.5, proof] The definition of h0 in the proof of Theorem 3.5, namely h0 := C_h(U) min_{θ̂N∈S′} min{⌊τ_+⌋−2, n−2}, is ambiguous: it is not clear whether the minimum is an exponent or a factor, and the expression can be negative when n = 1. Please clarify the condition inherited from [29].
  2. [Theorems 5.1 and 5.2] In the statements of both theorems, the hypothesis on the estimates is written as '{θ̂N}_{N=1}^∞ ⊆,' with the set S missing; it should read '⊆ S'.
  3. [Theorem 3.8] The theorem states the bound holds 'for any ε > 0', but the exponent min{τ̃,τ_-}−du/2−ε is negative for ε > min{τ̃,τ_-}−du/2, in which case the right-hand side does not converge to zero. The statement should specify that ε is a small positive number, or at least note that a vanishing bound requires ε below this threshold.
  4. [Lemma 5.5] The lemma assumes ν > 1 for the Matérn kernel, which is stronger than the lower bound on ν (or on τ(θ̂N) > du/2) used in Theorems 3.5 and 3.8. This mismatch between the smoothness needed for the emulator convergence and the smoothness needed for the posterior-consistency verification should be discussed explicitly, since it means the posterior results in Theorem 5.2 do not cover all cases covered by the emulator theorems.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the core convergence rates are derived from external scattered-data approximation results, and the self-citations to [49] are independently published prior theorems rather than inputs to the derivation.

full rationale

The derivation chain is not circular. Theorem 3.5 is proved by reducing the predictive-mean error to the external sampling inequalities of Narcowich, Ward, and Wendland [29, Lemma 4.1 and Theorem 4.2] together with Wendland's characterization of Matérn native spaces [54, Corollary 10.48]; the estimated hyperparameters enter only through the compactness assumption and through the quantities τ_- and τ_+ in assumption (e), not through any fitted value of the error itself. Theorem 3.8 obtains the variance bound via the standard power-function identity (Proposition 3.2, attributed to [49] but a known representer/kriging identity), followed by an application of Theorem 3.5 and Sobolev embedding; no fitted quantity is relabelled as a prediction. Theorem 3.11 likewise generalizes the external sparse-grid result of Nobile, Tempone, and Wolfers [32, Theorem 3], and Theorem 3.12 follows by the same argument as Theorem 3.8. In Section 5, Theorems 5.1 and 5.2 are explicitly adapted from the author's prior paper [49, Theorems 4.2, 4.9, and 4.11]; those are independently published theorems with their own proofs, not restatements of the present paper's assumptions or of its fitted values, and the present contribution is the combination of those posterior bounds with the Section 3 estimates. The self-citations are therefore not load-bearing in a circular sense. I also examined the flagged proof concern in Theorem 3.8: the step applying Theorem 3.5 to g in the unit ball of H_{k(θ̂_N)} and then using ‖g‖_{H^{τ̃}(U)} ≤ C_up(θ̂_N)‖g‖_{H_k(θ̂_N)(U)} is not justified when τ̃ > τ(θ̂_N), since H^{τ(θ̂_N)} need not embed into H^{τ̃}; this is a technical correctness gap in the proof as written, not a circularity, because the stated bound is not an input to the argument. The central claims are otherwise parameter-free and benchmarked against external results, so the circularity score is 0.

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

The paper introduces no free parameters fitted to data and no invented entities. The central claim rests on standard scattered-data approximation theory plus two modeling assumptions: the domain regularity (Lipschitz plus cone condition) and the boundedness of the hyperparameter estimates. The most fragile input is assumption (e), since practical estimators can leave any prescribed compact set; the theorems have nothing to say in that regime.

assumptions (6)
  • domain assumption U ⊆ R^{du} is compact with Lipschitz boundary and satisfies an interior cone condition (Theorem 3.5 assumption (a)).
    Required for the sampling inequalities from [28] that underpin the interpolation error bounds in Theorem 3.5. Excludes cusps and very thin domains.
  • standard math The native space of the Matérn kernel with smoothness ν equals the Sobolev space H^{ν+du/2}(U), with norm-equivalence constants bounded uniformly on compact hyperparameter sets (Proposition 3.3 and Lemma 3.4).
    Invoked throughout Section 3.1 to convert kernel-norm statements into Sobolev-norm rates. Relies on [54, Corollary 10.48] and Fourier analysis.
  • standard math Scattered-data interpolation error bounds for functions not necessarily in the kernel native space: Narcowich-Ward-Wendland [29, Lemma 4.1 and Theorem 4.2] provide the h and ρ exponents used in Theorem 3.5.
    The proof of Theorem 3.5 says the constants follow 'by inspection' of [29]; the bounding of C'(τ') and C''(θ) over the compact set S' is done here rather than re-derived.
  • standard math For tensor-product kernels, the sparse-grid interpolation bound of Nobile-Tempone-Wolfers [32, Theorem 3] extends to f not necessarily in the native space; Theorem 3.11's proof replaces [32, Proposition 4] with the one-dimensional bound (3.4).
    Underlies all Section 3.2 results and the dimension-independent (logarithmic) rates.
  • domain assumption The estimated hyperparameters {θ̂N} lie in a compact set S, and τ_- := inf_{N≥N*} τ(θ̂N) satisfies τ_- = n' + r' with n' > du/2 (Theorem 3.5 assumption (e)).
    This is the only assumption on the hyperparameter estimator; it rules out degenerate sequences such as σ̂²_N → 0 or λ̂_N → 0/∞, which standard estimators can produce. All theorems inherit it.
  • domain assumption For the posterior bounds in Section 5, the forward map G and log-likelihood Φ are continuous/bounded, the emulator error is bounded uniformly in N, and for Lemma 5.5 the Matérn smoothness satisfies ν > 1 (or ν_j > 1) so that the kernel is Lipschitz.
    Theorems 5.1 and 5.2 assume these to make Hellinger-distance constants independent of N; the ν > 1 condition is needed for the Sudakov-Fernique/Dudley argument bounding sup Φ_N.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Convergence of Gaussian Process Regression with Estimated Hyper-parameters and Applications in Bayesian Inverse Problems." pith.science (2026). https://pith.science/paper/JST3TNVD

@misc{pith2026190900232,
  author       = {Pith},
  title        = {Pith review of: Convergence of Gaussian Process Regression with Estimated Hyper-parameters and Applications in Bayesian Inverse Problems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JST3TNVD}},
  note         = {Machine review of arXiv:1909.00232}
}
abstract

This work is concerned with the convergence of Gaussian process regression. A particular focus is on hierarchical Gaussian process regression, where hyper-parameters appearing in the mean and covariance structure of the Gaussian process emulator are a-priori unknown, and are learnt from the data, along with the posterior mean and covariance. We work in the framework of empirical Bayes, where a point estimate of the hyper-parameters is computed, using the data, and then used within the standard Gaussian process prior to posterior update. We provide a convergence analysis that (i) holds for any continuous function $f$ to be emulated; and (ii) shows that convergence of Gaussian process regression is unaffected by the additional learning of hyper-parameters from data, and is guaranteed in a wide range of scenarios. As the primary motivation for the work is the use of Gaussian process regression to approximate the data likelihood in Bayesian inverse problems, we provide a bound on the error introduced in the Bayesian posterior distribution in this context.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 57 canonical work pages

  1. [1]

    http://functions.wolfram.com/bessel-typefunctions/besselk/20/showall.html, Accessed June 02, 2020

    work page 2020

  2. [2]

    R. A. Adams and J. J. Fournier , Sobolev spaces, Academic press, 2003. 25

    work page 2003

  3. [3]

    Anderes , On the consistent separation of scale and variance for Gauss ian random fields , The Annals of Statistics, (2010), pp

    E. Anderes , On the consistent separation of scale and variance for Gauss ian random fields , The Annals of Statistics, (2010), pp. 870–893

    work page 2010

  4. [4]

    Andrieu and G

    C. Andrieu and G. O. Roberts , The pseudo-marginal approach for efficient Monte Carlo computations, The Annals of Statistics, (2009), pp. 697–725

    work page 2009

  5. [5]

    Arcang ´ eli, M

    R. Arcang ´ eli, M. C. L. de Silanes, and J. J. Torrens , Extension of sampling inequal- ities to sobolev semi-norms of fractional order and derivat ive data , Numerische Mathematik, 121 (2012), pp. 587–608

    work page 2012

  6. [6]

    Bungartz and M

    H.-J. Bungartz and M. Griebel , Sparse grids , Acta numerica, 13 (2004), pp. 147–269

    work page 2004

  7. [7]

    R. H. Cameron and W. T. Martin , Transformations of Wiener integrals under transla- tions, Annals of Mathematics, (1944), pp. 386–396

    work page 1944

  8. [8]

    Choi and M

    T. Choi and M. J. Schervish , On posterior consistency in nonparametric regression prob - lems, Journal of Multivariate Analysis, 98 (2007), pp. 1969–198 7

    work page 2007

Show all 58 references
  1. [9]

    P. R. Conrad, M. Girolami, S. S ¨arkk¨a, A. Stuart, and K. Zygalakis , Statistical analysis of differential equations: introducing probabili ty measures on numerical solutions , Statistics and Computing, 27 (2017), pp. 1065–1082

    2017

  2. [10]

    Cotter, M

    S. Cotter, M. Dashti, and A. Stuart. , Variational data assimilation using targetted random walks, Int. J. Numer. Meth. Fluids., 68 (2012), pp. 403–421

    2012

  3. [11]

    T. Cui, C. Fox, and M. J. O’Sullivan , A posteriori stochastic correction of reduced mod- els in delayed-acceptance mcmc, with application to multip hase subsurface inverse problems , International Journal for Numerical Methods in Engineerin g, 118 (2019), pp. 578–605

    2019

  4. [12]

    W. Daqing , Fixed Domain Asymptotics and Consistent Estimation for Gaus sian Random Field Models in Spatial Statistics and Computer Experiments , PhD thesis, National University of Singapore, 2010

    2010

  5. [13]

    Diamond and M

    P. Diamond and M. Armstrong , Robustness of variograms and conditioning of kriging matrices, Mathematical Geology, 16 (1984), pp. 809–822

    1984

  6. [14]

    R. M. Dudley , The sizes of compact subsets of Hilbert space and continuity o f Gaussian processes, Journal of Functional Analysis, 1 (1967), pp. 290–330

    1967

  7. [15]

    D ˜ung, V

    D. D ˜ung, V. Temlyakov, and T. Ullrich , Hyperbolic cross approximation, Springer, 2018

    2018

  8. [16]

    M. M. Dunlop, T. Helin, and A. M. Stuart , Hyperparameter Estimation in Bayesian MAP Estimation: Parameterizations and Consistency , arXiv preprint arXiv:1905.04365, (2019)

    2019 arXiv

  9. [17]

    Fuglstad, D

    G.-A. Fuglstad, D. Simpson, F. Lindgren, and H. Rue , Constructing priors that penalize the complexity of Gaussian random fields , Journal of the American Statistical Association, 114 (2019), pp. 445–452

    2019

  10. [18]

    M. E. Johnson, L. M. Moore, and D. Ylvisaker , Minimax and maximin distance designs , Journal of statistical planning and inference, 26 (1990), p p. 131–148. 26

    1990

  11. [19]

    J. P. Kaipio and E. Somersalo , Statistical and Computational Inverse Problems , Springer, 2005

    2005

  12. [20]

    Karvonen, G

    T. Karvonen, G. Wynne, F. Tronarp, C. J. Oates, and S. S ¨arkk¨a, Maximum like- lihood estimation and uncertainty quantification for gauss ian process approximation of deter- ministic functions , arXiv preprint arXiv:2001.10965, (2020)

    2020 arXiv

  13. [21]

    M. C. Kennedy and A. O’Hagan , Bayesian calibration of computer models , Journal of the Royal Statistical Society: Series B (Statistical Methodol ogy), 63 (2001), pp. 425–464

    2001

  14. [22]

    B. T. Knapik, B. T. Szab ´o, A. W. V an Der V aart, and J. H. van Zanten , Bayes procedures for adaptive inference in inverse problems for t he white noise model , Probability Theory and Related Fields, 164 (2016), pp. 771–813

    2016

  15. [23]

    Ledoux and M

    M. Ledoux and M. Talagrand , Probability in Banach Spaces: isoperimetry and processes , Springer, 1991

    1991

  16. [24]

    H. C. Lie, T. Sullivan, and A. L. Teckentrup , Random forward models and log- likelihoods in Bayesian inverse problems , SIAM/ASA Journal on Uncertainty Quantification, 6 (2018), pp. 1600–1629

    2018

  17. [25]

    Loh , Fixed-domain asymptotics for a subclass of Mat´ ern-type Ga ussian random fields , The Annals of Statistics, 33 (2005), pp

    W.-L. Loh , Fixed-domain asymptotics for a subclass of Mat´ ern-type Ga ussian random fields , The Annals of Statistics, 33 (2005), pp. 2344–2394

    2005

  18. [26]

    Loh and T.-K

    W.-L. Loh and T.-K. Lam , Estimating structured correlation matrices in smooth Gaus sian random field models , The Annals of Statistics, 28 (2000), pp. 880–904

    2000

  19. [27]

    G. J. Lord, C. E. Powell, and T. Shardlow , An introduction to computational stochastic PDEs, Cambridge University Press, 2014

    2014

  20. [28]

    Narcowich, J

    F. Narcowich, J. W ard, and H. Wendland , Sobolev bounds on functions with scattered zeros, with applications to radial basis function surface fi tting, Mathematics of Computation, 74 (2005), pp. 743–763

    2005

  21. [29]

    F. J. Narcowich, J. D. W ard, and H. Wendland , Sobolev error estimates and a Bern- stein inequality for scattered data interpolation via radi al basis functions , Constructive Ap- proximation, 24 (2006), pp. 175–186

    2006

  22. [30]

    Niederreiter, Random Number Generation and quasi-Monte Carlo methods , SIAM, 1994

    H. Niederreiter, Random Number Generation and quasi-Monte Carlo methods , SIAM, 1994

    1994

  23. [31]

    Nobile, R

    F. Nobile, R. Tempone, and C. G. Webster , A sparse grid stochastic collocation method for partial differential equations with random input data , SIAM Journal on Numerical Analysis, 46 (2008), pp. 2309–2345

    2008

  24. [32]

    Nobile, R

    F. Nobile, R. Tempone, and S. Wolfers , Sparse approximation of multilinear problems with applications to kernel-based methods in UQ , Numerische Mathematik, (2017), pp. 1–34

    2017

  25. [33]

    Nobile and F

    F. Nobile and F. Tesei , A Multi Level Monte Carlo method with control variate for elli ptic PDEs with log-normal coefficients , Stochastic Partial Differential Equations: Analysis and Computations, 3 (2015), pp. 398–444. 27

    2015

  26. [34]

    Novak and H

    E. Novak and H. Triebel , Function spaces in Lipschitz domains and optimal rates of convergence for sampling , Constructive approximation, 23 (2006), pp. 325–350

    2006

  27. [35]

    O’Hagan, Bayesian analysis of computer code outputs: a tutorial , Reliability Engineering & System Safety, 91 (2006), pp

    A. O’Hagan, Bayesian analysis of computer code outputs: a tutorial , Reliability Engineering & System Safety, 91 (2006), pp. 1290–1300

    2006

  28. [36]

    G. D. Prato and J. Zabczyk. , Stochastic Equations in Infinite Dimensions , vol. 44 of Encyclopedia Math. Appl., Cambridge University Press, Cam bridge, 1992

    1992

  29. [37]

    Putter and G

    H. Putter and G. A. Young , On the effect of covariance function estimation on the accu- racy of kriging predictors , Bernoulli, 7 (2001), pp. 421–438

    2001

  30. [38]

    C. E. Rasmussen and C. K. Williams , Gaussian processes for machine learning , MIT Press, 2006

    2006

  31. [39]

    Reznikov and E

    A. Reznikov and E. Saff , The covering radius of randomly distributed points on a manif old, International Mathematics Research Notices, 2016 (2015), pp. 6065–6094

    2015

  32. [40]

    Rieger and H

    C. Rieger and H. Wendland , Sampling inequalities for sparse grids , Numerische Mathe- matik, 136 (2017), pp. 439–466

    2017

  33. [41]

    Sacks, W

    J. Sacks, W. J. Welch, T. J. Mitchell, and H. P. Wynn , Design and analysis of computer experiments, Statistical science, (1989), pp. 409–423

    1989

  34. [42]

    S ´andor, On new refinements of kobers and jordans trigonometric inequ alities, Notes Num- ber Theory Discrete Math, 19 (2013), pp

    J. S ´andor, On new refinements of kobers and jordans trigonometric inequ alities, Notes Num- ber Theory Discrete Math, 19 (2013), pp. 73–83

    2013

  35. [43]

    Scheuerer , Regularity of the sample paths of a general second order rand om field , Stochas- tic Processes and their Applications, 120 (2010), pp

    M. Scheuerer , Regularity of the sample paths of a general second order rand om field , Stochas- tic Processes and their Applications, 120 (2010), pp. 1879– 1897

    2010

  36. [44]

    Scheuerer, R

    M. Scheuerer, R. Schaback, and M. Schlather , Interpolation of spatial data–A stochas- tic or a deterministic problem? , European Journal of Applied Mathematics, 24 (2013), pp. 60 1– 629

    2013

  37. [45]

    Sch ¨olkopf, R

    B. Sch ¨olkopf, R. Herbrich, and A. J. Smola , A generalized representer theorem , in Computational learning theory, Springer, 2001, pp. 416–42 6

    2001

  38. [46]

    Sinsbeck and W

    M. Sinsbeck and W. Nowak , Sequential Design of Computer Experiments for the Solution of Bayesian Inverse Problems , SIAM/ASA Journal on Uncertainty Quantification, 5 (2017), pp. 640–664

    2017

  39. [47]

    M. L. Stein , Asymptotically efficient prediction of a random field with a mi sspecified covari- ance function , The Annals of Statistics, 16 (1988), pp. 55–63

    1988

  40. [48]

    , A simple condition for asymptotic optimality of linear pred ictions of random fields , Statistics & Probability Letters, 17 (1993), pp. 399–404

    1993

  41. [49]

    Stuart and A

    A. Stuart and A. Teckentrup , Posterior consistency for Gaussian process approximation s of Bayesian posterior distributions , Mathematics of Computation, 87 (2018), pp. 721–753

    2018

  42. [50]

    A. M. Stuart , Inverse problems , vol. 19 of Acta Numerica, Cambridge University Press, 2010, pp. 451–559. 28

    2010

  43. [51]

    A. W. V aart and J. A. Wellner , Weak convergence and empirical processes: with appli- cations to statistics , Springer, 1996

    1996

  44. [52]

    V an Der V aart and H

    A. V an Der V aart and H. V an Zanten , Information rates of nonparametric Gaussian process methods, Journal of Machine Learning Research, 12 (2011), pp. 2095– 2119

    2011

  45. [53]

    W ahba, Spline models for observational data , Siam, 1990

    G. W ahba, Spline models for observational data , Siam, 1990

    1990

  46. [54]

    Wendland , Scattered Data Approximation, Cambridge University Press, 2005

    H. Wendland , Scattered Data Approximation, Cambridge University Press, 2005

    2005

  47. [55]

    Wynne, F.-X

    G. Wynne, F.-X. Briol, and M. Girolami , Convergence guarantees for gaussian process approximations under several observation models , arXiv preprint arXiv:2001.10818, (2020)

    2020 arXiv

  48. [56]

    Ying , Maximum likelihood estimation of parameters under a spatia l sampling scheme , The Annals of Statistics, (1993), pp

    Z. Ying , Maximum likelihood estimation of parameters under a spatia l sampling scheme , The Annals of Statistics, (1993), pp. 1567–1590

    1993

  49. [57]

    Zhang , Inconsistent Estimation and Asymptotically Equal Interpola tions in Model-Based Geostatistics, Journal of the American Statistical Association, 99 (2004 ), pp

    H. Zhang , Inconsistent Estimation and Asymptotically Equal Interpola tions in Model-Based Geostatistics, Journal of the American Statistical Association, 99 (2004 ), pp. 250–261

    2004

  50. [58]

    Zhang and D

    H. Zhang and D. L. Zimmerman , Towards reconciling two asymptotic frameworks in spatial statistics, Biometrika, 92 (2005), pp. 921–936. 29

    2005

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.