Pith. sign in

REVIEW 2 major objections 2 minor 33 references

Nonlinear approximation with adaptive dictionaries

T0 review · 2 major / 2 minor · reviewed 2026-07-14 · grok-4.5

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

Pith's one-line read Sparse approximation of a kernel in a kernel-dependent adaptive dictionary controls the errors of optimal linear sampling recovery for the associated integral-operator function classes.

desk verdict Abstract-only Temlyakov note extending his widths/entropy program to adaptive-dictionary sparse approx of kernels for sampling recovery; program is clear, body uncheckable. read the letter →

arxiv 2607.10052 v1 pith:NEHWIPFQ submitted 2026-07-11 math.NA cs.NAmath.FA

V. Temlyakov This is my paper · ORCID
classification math.NAcs.NAmath.FA MSC 41A4641A2565D1546B2894A20
keywords nonlinearapproximationadaptivedictionarysparsesamplingrecoveryintegraloperatorskernelKolmogorovwidthsfunctionclasses
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 establishes that the right nonlinear-approximation problem for optimal linear sampling recovery of function classes that arise as images of L_q unit balls under an integral operator J_K is sparse approximation of the kernel K itself with respect to an adaptive dictionary determined by K. This replaces the classical bilinear dictionary that governs Kolmogorov widths of the same classes. The paper develops the approximation theory for that adaptive dictionary. It also continues a collection-level program: rather than treating one smoothness class at a time, it derives asymptotic sampling-recovery error behavior for the whole family of classes obtained when the kernels range over a fixed function class, extending earlier work of the same kind for Kolmogorov widths and entropy numbers.

What carries the argument

The adaptive dictionary determined by the kernel K (as opposed to the classical bilinear dictionary). Sparse approximation rates of K in this K-dependent dictionary carry the sampling-recovery error bounds for the images J_K(B(L_q)).

What would settle it

Exhibit a concrete kernel K for which the optimal linear sampling recovery error of J_K(B(L_q)) fails to be asymptotically equivalent to the sparse approximation error of K with respect to the adaptive dictionary determined by K.

Watch

Extended reading notes

Core claim

Sparse approximation of an integral kernel K with respect to an adaptive dictionary determined by K is the correct nonlinear-approximation object that controls the errors of optimal linear sampling recovery of the classes J_K(B(L_q)). Moreover, asymptotic recovery-error rates can be obtained for the entire collection of such classes as the kernels themselves range over a given function class.

Load-bearing premise

The recently discovered correspondence that optimal linear sampling recovery of J_K-images is governed by sparse approximation of K in a K-dependent adaptive dictionary, rather than the classical bilinear dictionary, is taken as the structural starting point.

Editorial extensions

If this is right

  • Optimal linear sampling recovery rates for images under J_K are controlled by the sparse approximation rates of K in its adaptive dictionary.
  • Asymptotic recovery-error behavior can be read off for whole families of integral-operator classes at once when kernels come from a fixed function class.
  • The collection-based asymptotic program previously used for Kolmogorov widths and entropy numbers now extends to sampling recovery.
  • Techniques of nonlinear approximation with adaptive dictionaries become direct tools for sampling recovery theory.

Reading between the lines

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

  • Adaptive dictionaries may systematically replace fixed bilinear dictionaries whenever the recovery functionals are restricted to point evaluations rather than free linear functionals.
  • The collection approach points toward a uniform theory of recovery rates for operator-generated classes parameterized solely by the smoothness of the kernel class.
  • Explicit rates for standard kernels would follow by computing adaptive-dictionary approximation numbers of those kernels and comparing them to bilinear ones, quantifying the gap between widths and sampling recovery.
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

2 major / 2 minor

Summary. Based solely on the abstract, the manuscript studies nonlinear sparse approximation of an integral kernel K with respect to an adaptive dictionary determined by K (rather than the classical bilinear dictionary). It asserts that this adaptive-dictionary approximation is the correct object for controlling the errors of optimal linear sampling recovery of the classes J_K(B_{L_q}), in contrast to Kolmogorov widths, which connect to bilinear sparse approximation. The paper further develops a collection-level program: asymptotic sampling-recovery error behavior for the family of such classes when kernels range over a given function class, continuing earlier work on Kolmogorov widths and recent work on entropy numbers.

Significance. If the claimed correspondence between adaptive sparse approximation of K and optimal linear sampling recovery holds with matching (or order-equivalent) rates for the kernels under study, the work would supply the appropriate nonlinear-approximation framework for sampling recovery of integral-operator images and would extend the collection-level asymptotic program from widths and entropy numbers to sampling recovery. That would be a meaningful contribution in approximation theory. Credit is due for framing the problem at the collection level rather than only for individual smoothness classes. The assessment is provisional: only the abstract is available, so theorems, proofs, rate statements, and the precise form of the correspondence cannot be checked.

major comments (2)
  1. [Abstract] The abstract's second sentence takes as structural starting point a 'recently discovered' correspondence: optimal linear sampling recovery of J_K-images is governed by sparse approximation of K in a K-dependent adaptive dictionary. The abstract states neither the precise form of that correspondence, its hypotheses, nor whether it supplies two-sided (or order-equivalent) bounds for the kernels under study. Without the body, one cannot verify that adaptive sparse rates control recovery errors as claimed; if the prior result is only one-sided or holds only under restrictions not met by a positive portion of the kernel class, the collection asymptotics do not transfer. This is load-bearing for the central claim.
  2. [Abstract] The collection-level asymptotic recovery-error results (final sentences of the abstract) rest on the same correspondence transferring rates when K ranges over a given function class. The abstract does not indicate the range of kernels, the form of the rates, or any comparison with classical bilinear rates. Full verification requires the statements and proofs in the body; on the abstract alone the transfer cannot be assessed.
minor comments (2)
  1. [Abstract] The abstract is readable but dense; a brief explicit definition or pointer to the adaptive dictionary (even at the level of 'dictionary of the form {K(x,·) : x in a discrete set}') would help non-specialists.
  2. [Abstract] The phrase 'this important problem of nonlinear approximation with respect to an adaptive dictionary' is evaluative; a more neutral formulation is preferable in the abstract.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: abstract-only theoretical program with no definitional reduction or fitted predictions.

full rationale

Only the abstract is available. It states a known connection between Kolmogorov widths of J_K(B_Lq) and sparse approximation of K in the classical bilinear dictionary, then notes a recently discovered parallel for optimal linear sampling recovery that instead requires sparse approximation of K in a K-dependent adaptive dictionary. The paper proposes to study that nonlinear-approximation problem and to obtain collection-level asymptotics for sampling-recovery errors when kernels range over a given class (an approach previously used for widths and entropy numbers). No equations, definitions, fitted parameters, uniqueness theorems, or explicit self-citations appear in the provided text. Nothing is claimed to equal its input by construction; the adaptive-dictionary object is presented as the natural counterpart of the recovery problem rather than as a tautological redefinition of the recovery error. Ordinary self-citation risk for the author's prior line cannot be verified or scored as load-bearing circularity from the abstract alone. Per the hard rules, an abstract-only theoretical announcement with no exhibited reduction receives score 0 and empty steps.

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

Abstract-only pure-math paper: no numerical free parameters are introduced. Background is standard approximation-theory structure (integral operators, L_q balls, sparse approximation, sampling recovery). The load-bearing domain assumption is the recently discovered link between adaptive sparse approximation of K and optimal linear sampling recovery. No new physical entities are postulated.

assumptions (3)
  • domain assumption Kolmogorov widths of J_K-images of L_q unit balls are closely connected to sparse approximation of K in the classical bilinear dictionary.
    Stated as well-known background in the opening sentence; used to contrast with the adaptive-dictionary setting.
  • domain assumption Optimal linear sampling recovery of the same classes is governed by sparse approximation of K with respect to an adaptive dictionary determined by K.
    Abstract attributes this to recent discovery and takes it as the reason to study adaptive dictionaries; it is the structural premise of the paper.
  • domain assumption Asymptotic analysis of recovery (or width/entropy) quantities over collections of classes generated by kernels from a given function class is a meaningful and comparable program.
    Abstract says this approach was already realized for Kolmogorov widths and recently for entropy numbers; the paper continues it for sampling recovery.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Nonlinear approximation with adaptive dictionaries." pith.science (2026). https://pith.science/paper/NEHWIPFQ

@misc{pith2026260710052,
  author       = {Pith},
  title        = {Pith review of: Nonlinear approximation with adaptive dictionaries},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NEHWIPFQ}},
  note         = {Machine review of arXiv:2607.10052}
}
abstract

It is well known that the study of the Kolmogorov widths of a function class, which is the image of the unit ball of the $L_q$ space of an integral operator $J_K$ with the kernel $K$, is closely connected with the study of sparse approximations of the kernel $K$ with respect to the classical bilinear dictionary. Recently, it was discovered that if instead of the Kolmogorov widths we study the errors of optimal linear sampling recovery of the same classes, then we need to study sparse approximations of the kernel $K$ with respect to an adaptive dictionary, which is determined by the kernel $K$. In this paper we study this important problem of nonlinear approximation with respect to an adaptive dictionary. Also, in this paper we continue to develop the following general approach, which is related to the above nonlinear approximation problem. We study asymptotic behavior of the errors of sampling recovery not for an individual smoothness class, how it is usually done, but for the collection of classes, which are defined by integral operators with kernels coming from a given class of functions. Earlier, such approach was realized for the Kolmogorov widths and very recently for the entropy numbers.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 6 linked inside Pith

  1. [1]

    Batson, D.A

    J. Batson, D.A. Spielman, and N. Srivastava, Twice-Ramanujan Spar- sifiers,SIAM Review,56(2014), 315–334

    2014

  2. [2]

    Belinsky, Estimates of entropy numbers and Gaussian measures for classes of functions with bounded mixed derivative,J

    E.S. Belinsky, Estimates of entropy numbers and Gaussian measures for classes of functions with bounded mixed derivative,J. Approx. Theory,93(1998), 114-127

    1998

  3. [3]

    Borodin, Yu.A

    P.A. Borodin, Yu.A. Skvortsov, Approximation of functions on two variables by linear combinations of its crosses, Mat. Sb.217:4(2026), 42–65

    2026

  4. [4]

    F. Dai, A. Prymak, V.N. Temlyakov, and S.U. Tikhonov, Integral norm discretization and related problems,Russian Math. Surveys 74:4(2019), 579–630. Translation fromUspekhi Mat. Nauk74:4(448) (2019), 3–58; arXiv:1807.01353v1. 27

    arXiv 2019

  5. [5]

    F. Dai, A. Prymak, A. Shadrin, V. Temlyakov, and S. Tikhonov, Entropy numbers and Marcinkiewicz-type discretization theorem, J. Functional Analysis281(2021) 109090; arXiv:2001.10636v1 [math.CA] 28 Jan 2020

    arXiv 2021

  6. [6]

    Dissertation(1985), Moscow, MGU

    Dinh Dung, Approximation of multivariate functions by means of har- monic analysis,Hab. Dissertation(1985), Moscow, MGU

    1985

  7. [7]

    Temlyakov, and T

    Ding Dung, V.N. Temlyakov, and T. Ullrich, Hyperbolic Cross Ap- proximation, Advanced Courses in Mathematics CRM Barcelona, Birkh¨ auser, 2018; arXiv:1601.03978v2 [math.NA] 2 Dec 2016

    arXiv 2018

  8. [8]

    Kashin, E

    B.S. Kashin, E. Kosov, I. Limonova, and V.N. Temlyakov, Sampling discretization and related problems,J. Complexity71(2022), Paper No. 101653

    2022

Show all 33 references
  1. [9]

    Kiefer and J

    J. Kiefer and J. Wolfowitz, The equivalence of two extremum prob- lems, Canadian J. Math.,12(1960), 363–366

    1960

  2. [10]

    Kolmogorov, Uber die beste Ann¨ aherung von Funktionen einer Funktionklasse, Ann

    A.N. Kolmogorov, Uber die beste Ann¨ aherung von Funktionen einer Funktionklasse, Ann. Math.,37(1936), 107–111

    1936

  3. [11]

    Krieg, K

    D. Krieg, K. Pozharska, M. Ullrich, and T. Ullrich, Sampling pro- jections in the uniform norm, Journal of Mathematical Analysis and Applications553(2)(2025), 129873

    2025

  4. [12]

    Limonova, Yu

    I. Limonova, Yu. Malykhin, and V. Temlyakov, One-sided discretiza- tion inequalities and sampling recovery,Uspekhi Mat. Nauk,79 (2024), no. 3(477), 149–180

    2024

  5. [13]

    Limonova and V

    I. Limonova and V. Temlyakov, On sampling discretization inL 2,J. Math. Anal. Appl.515(2022), Paper No. 126457; arXiv:2009.10789v1 [math.FA] 22 Sep 2020

    2022 arXiv

  6. [14]

    Schmidt, Zur Theorie der linearen und nichtlinearen Integralgle- ichungen

    E. Schmidt, Zur Theorie der linearen und nichtlinearen Integralgle- ichungen. I,Math. Annalen,63(1906), 433–476

    1906

  7. [15]

    Smolyak, Quadrature and interpolation formulas for tensor prod- ucts of certain classes of functions, Dokl

    S.A. Smolyak, Quadrature and interpolation formulas for tensor prod- ucts of certain classes of functions, Dokl. Akad. Nauk SSSR,148 (1963), 1042–1045; English transl. in Soviet Math. Dokl.,4(1963). 28

    1963

  8. [16]

    Temlyakov, Approximate recovery of periodic functions of sev- eral variables, Mat

    V.N. Temlyakov, Approximate recovery of periodic functions of sev- eral variables, Mat. Sb.128(170)(1985), 256–268; English transl. in Math. USSR Sb.56(1987)

    1985

  9. [17]

    Temlyakov, On best bilinear approximations of periodic func- tions of several variables, Dokl

    V.N. Temlyakov, On best bilinear approximations of periodic func- tions of several variables, Dokl. Akad. Nauk SSSR,286(1986), 301– 304; English transl. in Soviet Math. Dokl.,33(1986), 96–99

    1986

  10. [18]

    Temlyakov, Approximation of Periodic Functions of Several Vari- ables by Bilinear Forms, Izvestiya AN SSSR, Ser

    V.N. Temlyakov, Approximation of Periodic Functions of Several Vari- ables by Bilinear Forms, Izvestiya AN SSSR, Ser. Mat.,50(1986), 137–155; English transl. in Mathematics of the USSR-Izvestia,28 (1987), 133–150

    1986

  11. [19]

    Temlyakov, On widths of function classes, Dokl

    V.N. Temlyakov, On widths of function classes, Dokl. Akad. Nauk SSSR,294(1987), 1069–1072; English transl. in Soviet Math. Dokl., 35(1987), 639–642

    1987

  12. [20]

    Temlyakov, Estimates of the best bilinear approximations of functions of two variables and some of their applications, Mat

    V.N. Temlyakov, Estimates of the best bilinear approximations of functions of two variables and some of their applications, Mat. Sb., 134(1987), 93–107; English transl. in Math. USSR-Sb62(1989), 95–109

    1987

  13. [21]

    Temlyakov, Approximation of functions with bounded mixed derivative, Trudy MIAN,178(1986), 1–112

    V.N. Temlyakov, Approximation of functions with bounded mixed derivative, Trudy MIAN,178(1986), 1–112. English transl. in Proc. Steklov Inst. Math.,1(1989)

    1986

  14. [22]

    Temlyakov, Estimates of best bilinear approximations of periodic functions, Trudy Mat

    V.N. Temlyakov, Estimates of best bilinear approximations of periodic functions, Trudy Mat. Inst. Steklov,181(1988), 250–267; English transl. in Proc. Steklov Inst. of Math.4(1989), 275–293

    1988

  15. [23]

    Temlyakov, Bilinear approximation and related questions, Trudy MIAN im

    V.N. Temlyakov, Bilinear approximation and related questions, Trudy MIAN im. Steklova,194(1992), 229–248; English transl. in Proceed- ings of the Steklov Inst. Math., 1993, Issue 4, 245–265

    1992

  16. [24]

    Temlyakov, On Approximate Recovery of Functions with Bounded Mixed Derivative, J

    V.N. Temlyakov, On Approximate Recovery of Functions with Bounded Mixed Derivative, J. Complexity,9(1993), 41–59

    1993

  17. [25]

    Temlyakov,Approximation of periodic functions, Nova Science Publishes, Inc., New York., 1993

    V.N. Temlyakov,Approximation of periodic functions, Nova Science Publishes, Inc., New York., 1993. 29

    1993

  18. [26]

    Temlyakov, Greedy Approximation, Cambridge University Press, 2011

    V.N. Temlyakov, Greedy Approximation, Cambridge University Press, 2011

    2011

  19. [27]

    Temlyakov,Multivariate Approximation, Cambridge University Press, 2018

    V. Temlyakov,Multivariate Approximation, Cambridge University Press, 2018

    2018

  20. [28]

    Temlyakov, The Marcinkiewicz-Type Discretization Theorems, Constructive Approximation,48(2018), 337–369

    V.N. Temlyakov, The Marcinkiewicz-Type Discretization Theorems, Constructive Approximation,48(2018), 337–369

    2018

  21. [29]

    Temlyakov, Connections between numerical integration, discrep- ancy, dispersion, and universal discretization, SMAI-JCM SMAI Jour- nal of Computational Mathematics,S5(2019), p

    V. Temlyakov, Connections between numerical integration, discrep- ancy, dispersion, and universal discretization, SMAI-JCM SMAI Jour- nal of Computational Mathematics,S5(2019), p. 185-209

    2019

  22. [30]

    Temlyakov, On optimal recovery inL 2, J

    V.N. Temlyakov, On optimal recovery inL 2, J. Complexity65(2021), 101545; arXiv:2010.03103v1 [math.NA] 7 Oct 2020

    2021 arXiv

  23. [31]

    Temlyakov, Entropy numbers of classes defined by integral opera- tors, arXiv:2505.08572v1 [math.NA] 13 May 2025

    V. Temlyakov, Entropy numbers of classes defined by integral opera- tors, arXiv:2505.08572v1 [math.NA] 13 May 2025

    2025 arXiv

  24. [32]

    Temlyakov, Sampling recovery on classes defined by integral oper- ators and sparse approximation with adaptive dictionaries, Journal of Complexity,96(2026), 102066

    V. Temlyakov, Sampling recovery on classes defined by integral oper- ators and sparse approximation with adaptive dictionaries, Journal of Complexity,96(2026), 102066

    2026

  25. [33]

    Trigub and E.S

    R.M. Trigub and E.S. Belinsky, Fourier Analysis and Approximation of Functions, Kluwer Academic Publishers, 2004. V.N. Temlyakov,University of South Carolina, USA, Steklov Mathematical Institute of Russian Academy of Sciences, Russia; Lomonosov Moscow State University, Russia;...

    2004

Pith tools

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