Pith. sign in

REVIEW 2 major objections 3 minor 31 references

On the multivariate multifractal formalism: examples and counter-examples

T0 review · 2 major / 3 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read For two measures, the joint singularity spectrum and its Legendre bound can have disjoint supports, and even for a pair of correlated random cascades the Legendre spectrum's support is strictly smaller than the true joint spectrum.

desk verdict A concrete counterexample to the simultaneous-maximizer claim in Lemma 5.3 breaks the proof of Theorem 1.3 as written, though the paper's questions and constructions are still worthwhile. read the letter →

arxiv 2411.13959 v1 pith:ZF3XMQVL submitted 2024-11-21 math.MG

Stéphane Seuret (UPEC UP12) This is my paper · ORCID
classification math.MG MSC 28A7828A8028C1511K5560G5737D35
keywords multifractalformalismbivariateanalysisLegendrespectrumBernoullicascadesrandommeasuresHausdorffdimensionlocaldyadicintervals
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 the bivariate version of the Legendre multifractal formalism holds: in the univariate case the multifractal spectrum is always bounded above by the Legendre transform of the Lq-spectrum, so it is natural to ask whether the same is true for pairs of measures. The paper answers negatively in a strong way by constructing two measures whose bivariate singularity spectrum and the Legendre transform of their bivariate Lq-spectrum have disjoint supports. It then analyzes a random model that switches between two Bernoulli parameters p1 and p2 at random generations, producing a pair (µp1, νη). When p1 and p2 lie on the same side of 1/2, the bivariate formalism holds and the support is a parallelogram; when they lie on opposite sides, the support of the multifractal spectrum is a pentagon strictly larger than that of the Legendre spectrum, so the formalism fails. These are among the first examples where the two supports are proved to differ, and they caution against the common practice of estimating a joint multifractal spectrum as the Legendre transform of an estimated Lq-spectrum.

What carries the argument

The central objects are the bivariate Lq-spectrum τµ,ν(q1,q2) = liminf_{j→∞} -1/j log2 Σ_{I∈D_j} µ(3I)^{q1} ν(3I)^{q2} and its Legendre transform τ*µ,ν. For the random model the key mechanism is the factorization νη(I) = µp1(I^A)µp2($I^{{A^c}}$), where A is the random set of generations in which the Bernoulli parameter switches, combined with the affine relation G1,2 between the local dimensions of µp1 and µp2 (Lemma 4.1) and a concentration lemma (Lemma 4.3) for the cardinality of A in short windows. The crucial difference between the two regimes is Lemma 5.3: when p1 and p2 are on the same side of 1/2, the same neighbouring dyadic word e_w maximizes both µp1(3I_w) and νη(3I_w), so the dilated sum is comparable to the undilated sum and τ = T^η; when they are on opposite sides, this joint domination fails, and a three-range split of the generation index yields τ = min(T^η, eT^η) with a phase transition. The dimension computations rest on Proposition 4.4, which gives the Hausdorff dimension of level sets defined through the random subwords A and A^c.

What would settle it

For the random model with p1=0.27, p2=0.8, η=0.5, compute the partial sums defining τµp1,νη,j(q1,q2) for large j on a grid of (q1,q2), and check that they converge to min(T^η, eT^η); then compute the Legendre transform numerically and verify that its support equals the pentagon Pη_1. The central claim of Theorem 1.4 would be refuted if any finite-generation estimate deviated systematically from the closed form, or if a point inside Pη_2 \ Pη_1 were found with no x attaining dim(µp1,x)=H1 and dim(νη,x)=H2.

Watch

Extended reading notes

Core claim

The central claim is that the natural bivariate analogue of the Legendre upper bound fails: there are probability measures µ and ν supported on [0,1] for which Supp(Dµ,ν) ∩ Supp(τ*µ,ν) = ∅ (Theorem 1.2). The construction alternates between a Lebesgue-like scheme and a Cantor-like scheme, forcing both measures to have only local dimensions 1/2 and +∞ while the bivariate Lq-spectrum is the minimum of three affine functions, whose Legendre transform is supported on a triangle that misses the four atoms of the singularity spectrum. For the random model (µp1, νη), the paper gives the complete bivariate analysis: when p1 and p2 are on the same side of 1/2, the formalism holds everywhere on a deterministic parallelogram Pη, with τ = T^η and D = (T^η)*; when 0 < p1 < 1/2 < p2 < 1, the Lq-spectrum is min(T^η, eT^η), the support of the Legendre spectrum is a pentagon Pη_1, and the support of D is a strictly larger pentagon Pη_2, so the formalism fails. The underlying reason is that the lower local dimensions of the two measures may be attained at different scales, while the Lq-spectrum samples the measures at the same scale.

Load-bearing premise

The computation of the Lq-spectrum in the same-side regime relies on the fact that, for every dyadic word w, the same neighbouring interval e_w maximizes both µp1 and νη among the three intervals contributing to 3I_w; if joint maximization failed at some generations, τµp1,νη would not equal T^η and the verification of the formalism in that regime would break down.

Editorial extensions

If this is right

  • For a pair of measures, the bivariate Legendre transform of the Lq-spectrum is not, in general, an upper bound for the bivariate multifractal spectrum; even the supports can be disjoint.
  • For the random cascade pair with p1 and p2 on opposite sides of 1/2, the support of the joint singularity spectrum is strictly larger than that of the Legendre spectrum, so Legendre-based estimation misses an entire region of possible joint scaling behaviors.
  • When p1 and p2 are on the same side of 1/2, the same model does satisfy the bivariate multifractal formalism, with the joint spectrum equal to the Legendre transform on a deterministic parallelogram.
  • The gap between the two pentagons depends on the switching probability η and on the parameters, so the area ratio encodes information about the correlation between the two measures.
  • The counter-example of Theorem 1.2 extends to higher-dimensional measures and to more than two measures by the same construction.

Reading between the lines

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

  • If the paper is right, the disjoint-support phenomenon suggests that any successful multivariate multifractal formalism must either restrict the class of measures (for example by a joint doubling or neighbouring condition) or replace the Lq-spectrum by an object that can sample several scales independently for each coordinate.
  • A testable consequence for applications is that in bivariate signal or image analysis, Legendre-based estimates of joint multifractal spectra may systematically under-report rare joint events when the two signals are correlated only at some scales; this could be checked by comparing Legendre estimates with direct box-counting estimates on synthetic cascades with known switching probabilities.
  • One could extend the random switching construction to more than two parameters or to a Markov-switching sequence; the phase-transition curve T^η = eT^η would then be expected to become a fractal set, and the support gap between the two spectra would presumably grow with the number of regimes.
  • The explicit dimension formula D = min(D1,D2) for the random model provides an exact benchmark against which numerical estimators of joint fractal dimensions can be tested, especially on the non-convex pentagonal support.
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

2 major / 3 minor

Summary. The paper studies the bivariate multifractal formalism for pairs of probability measures. Theorem 1.2 constructs two measures for which the support of the bivariate multifractal spectrum and the support of the Legendre transform of the bivariate Lq-spectrum are disjoint, disproving the natural multivariate analogue of the univariate upper bound. Theorems 1.3 and 1.4 then analyze the pair (μp1, νη), where νη is a random binomial cascade that changes parameter at random generations. The paper claims that when p1, p2 lie on the same side of 1/2 the bivariate Lq-spectrum is T^η and the multifractal formalism holds with parallelogram support, while for p1 < 1/2 < p2 the support is a pentagon strictly larger than the Legendre pentagon and the formalism fails. The proofs use explicit step constructions, large-deviation estimates, and a dimension formula for intersections of level sets.

Significance. If the results are correct, Theorem 1.2 is a striking and valuable counterexample, and the random pair (μp1, νη) is a natural test case for bivariate multifractal behavior. The paper provides explicit constructions with quantitative estimates, and it reproduces the needed computation from [19] in Lemma 5.1 rather than merely citing it. However, the proof of Theorem 1.3 currently rests on false statements in Section 5, so the advertised same-side positive result is not established as written. The disjoint-support counterexample and the opposite-side analysis may still be sound, but the manuscript needs substantive repair before the central claims can be accepted.

major comments (2)
  1. [§5.1, Lemma 5.3] The simultaneous-maximizer assertion in Lemma 5.3 is false, and the counterexample is not marginal. Take p1 = 0.1, p2 = 0.45, A = {1,2,3,6}, and w = 001000 of length 6. For μp1 the three neighboring masses are μp1(I_{w^-}) = 0.000729, μp1(I_w) = 0.000009, μp1(I_{w^+}) = 0.000081, so e_w = w^- is the unique μp1-maximizer. For νη the masses are νη(I_{w^-}) = 0.00027225, νη(I_w) = 0.00018225, νη(I_{w^+}) = 0.00164025, so the νη-maximizer is w^+, not w^-. Moreover νη(3I_w) = 0.00209475 > 3νη(I_{w^-}) = 0.00081675, so the inequality νη(3I_w) ≤ 3νη(I_{e_w}) used in the proof fails at this word. Since A has positive probability, such words occur almost surely at infinitely many generations. Consequently the comparison between the dilated sum defining τμp1,νη and the undilated sum 2^{-jT^η(q1,q2)} of Lemma 5.1 is not justified. The proof of τμp1,νη = T^η, and with it the verification of the multifractal formalism in Section 5.3, has a genuine gap. I have not been able to determine whether Theorem 1.3(1) itself is false; the provided proof is invalid.
  2. [§5.2, Lemma 5.5] The slope assertion in Lemma 5.5 is reversed. The proof states that for 0 < p1 < p2 < 1/2 the slope of G1,2 is δ1,2 > 1. But δ1,2 = (H_{2,max} - H_{2,min}) / (H_{1,max} - H_{1,min}), and for p < 1/2 the width H_{p,max} - H_{p,min} = log2((1-p)/p) decreases as p increases toward 1/2. Hence 0 < p1 < p2 < 1/2 gives δ1,2 < 1, not > 1. The later sentence 'when 0 < p2 < p1 < 1/2, then δ1,2 < 1' is also reversed: for p2 < p1 < 1/2 one has δ1,2 > 1. This is not a harmless typo: the four-case analysis and the exclusion argument for the lower endpoint a3 = a1 explicitly use the sign of δ1,2 - 1 to decide whether the minimum of H2 corresponds to the maximum or the minimum of α. With the correct sign, the case labels and the extremal computations must be interchanged. The parallelogram support may still be correct, but the proof as written does not establish it for both orderings of p1 and p2 covered by Theorem 1.3.
minor comments (3)
  1. [Theorem 1.4] The hypothesis '0 < p1 < 1/2 < p2 < 1/2' is contradictory; it should read '0 < p1 < 1/2 < p2 < 1'.
  2. [Title and Theorem 1.3] The title in the text appears as 'ON THE MULTIV ARIATE MULTIFRACTAL FORMALISM', with an unwanted space; also Theorem 1.3's phrase 'both greater or both larger than 1/2' should be 'both smaller than 1/2 or both larger than 1/2'.
  3. [§5.2, Lemma 5.5] The proof assumes 0 < p1 < p2 < 1/2 'without loss of generality', but Theorem 5.4 also covers the opposite ordering; since the sign of δ1,2 - 1 changes with the ordering, the reduction should be stated explicitly.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the bivariate spectra are computed from explicit constructions with free parameters, and the self-citations to [1,19] are contextual or reproduced; the possible Lemma 5.3 gap is a correctness issue, not circularity.

full rationale

The derivation chain is self-contained: no claim reduces by definition to its own input. Theorem 1.2 is a constructive counterexample: the measures ν1 and ν2 are built by alternating schemes (P1) and (P2), and the two spectra are afterwards derived — D_{ν1,ν2} from the scale bounds (16),(28) giving local dimension 1/2, and τ_{ν1,ν2} from the explicit estimates (13),(19),(25),(31) that yield the three-affine-piece function min(q1+q2−1, q1/2+q2−1/2, q1+q2/2−1/2). The disjointness of Supp(D) and Supp(τ*) is a computed consequence of the construction, not an input to it. Theorem 1.3(1) is also derived: Lemma 5.1 computes the undilated sum exactly by factoring over the random generations and using Lemma 4.3 (a Hoeffding bound proved in the paper), obtaining T^η; Lemma 5.3 then supplies the 3I-versus-I comparison that passes from eτ to τ. T^η is defined before the theorem, and τ is defined by (5), so the equality τ = T^η is substantive. The formalism check in Section 5.3 is an independent verification that the Legendre transform of this derived T^η reproduces the spectrum D computed in Theorem 5.4 from Proposition 4.4 (proved in-paper, explicitly not imported from [27,28,2]). Theorem 1.4 follows the same pattern: the pentagons P^η_1, P^η_2 are geometric outputs of the two systems (S1),(S2) derived from the computed local-dimension constraints. No parameter is fitted to data; p1, p2, η are free variables over which the theorems quantify uniformly. Self-citations to [1,19] (shared authorship) are used for background, for the η = 0 special case, and for credit; where a cited computation is genuinely needed (eτ_{μp1,μp2} = T in Lemma 5.1), it is reproduced in the paper, so the citation is not load-bearing. The reviewer's counterexample against Lemma 5.3, if valid, would be a correctness gap in a proof step, not a circularity: the lemma asserts a checkable property of the binomial construction and is not equivalent by construction to the conclusion it supports. Circumstantial self-citation of contextual background is present, hence score 1 rather than 0, but the central results carry independent content.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The only new objects are mathematical constructions, namely the measures ν1 and ν2 in Theorem 1.2 and the random measure νη, used to prove theorems rather than to explain empirical data. The model parameters p1, p2 and eta are arbitrary inputs, not fitted constants; the construction lengths in Section 2 are chosen generically and disappear from the conclusions.

assumptions (4)
  • domain assumption Standard large-deviation and multifractal formalism for binomial measures: Proposition 3.3 items (1) through (6) and the Legendre relations.
    Imported from [10,26,3] and used throughout Sections 4 to 6; cited, not reproved in this paper.
  • standard math Hoeffding's inequality and the Borel-Cantelli lemma for the concentration estimate (38).
    Used in Lemma 4.3 to control the random set A uniformly; standard probability tools.
  • standard math Billingsley's lemma for converting pointwise local dimension of a measure into a lower bound on Hausdorff dimension.
    Used in Proposition 4.4 and Proposition 6.6 to prove lower bounds for the dimension of level sets.
  • domain assumption The dyadic-neighbour equivalence between ball masses and masses of I, I+, I- together with the doubling property of the constructed measures.
    Used throughout; for binomial measures it follows from the product structure, and for the Theorem 1.2 measures it is built in via the inequalities around equations (11) to (17).

how reviews work

0 comments
Cite this review

Pith. "Pith review of On the multivariate multifractal formalism: examples and counter-examples." pith.science (2026). https://pith.science/paper/ZF3XMQVL

@misc{pith2026241113959,
  author       = {Pith},
  title        = {Pith review of: On the multivariate multifractal formalism: examples and counter-examples},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZF3XMQVL}},
  note         = {Machine review of arXiv:2411.13959}
}
read the original abstract

In this article, we investigate the bivariate multifractal analysis of pairs of Borel probability measures. We prove that, contrarily to what happens in the univariate case, the natural extension of the Legendre spectrum does not yield an upper bound for the bivariate multifractal spectrum. For this we build a pair of measures for which the two spectra have disjoint supports. Then we study the bivariate multifractal behavior of an archetypical pair of randomly correlated measures, which give new, surprising, behaviors, enriching the narrow class of measures for which such an analysis is achieved.

Figures

Figures reproduced from arXiv: 2411.13959 by the authors.

Figure 1
Figure 1. Case 0 < p1, p2 < 1/2 or 1/2 < p1, p2 < 1: Left: Support of Dµp1 ,νη . Right: Bivariate multifractal spectrum Dµp1 ,νη of (µp1 , νη). Theorem 1.3 extends the results of [19, 1] which correspond to η = 0. Calling δ1,2 = − log2 (p2) + log2 (1 − p2) − log2 (p1) + log2 (1 − p1) (9) , we will see that the support of Dµp1 ,νη is the parallelogram Pη generated by four straight lines, those with slopes 1 or δ1,2 passing by … view at source ↗
Figure 2
Figure 2. Case 0 < p1 = 0.27 < 1/2 < p2 = 0.8 and η = 0.5. Top left: Support of the Legendre spectrum τ ∗ µp1 ,νη Top right: Legendre spectrum τ ∗ µp1 ,νη . Bottom left: Support of the multifractal spectrum Dµp1 ,νη Bottom right: Bivariate multifractal spectrum Dµp1 ,νη . The supports are distinct (but not disjoint) pentagons. usually write w = w1 · · ·w|w| and w ∈ Σ|w| , and w = w1 · · ·wn · · · when w ∈ Σ. Then if j ≤ |w|, … view at source ↗
Figure 3
Figure 3. Legendre transform of τν1,ν2 . One now collects the information: • Let i ∈ {1, 2}, ε > 0, and x ∈ supp(νi). By construction, (16) and (28), there exist infinitely many generations j = 2j 1 n ot j = 2j 3 n such that εn ≤ ε, 2−j/2(1+ε) ≤ νi(B(x, 2 −j )) = νi(Ij (x)) ≤ 2 −j/2(1−ε) . Also, for every j large, 2−j/2(1+ε) ≤ νi(B(x, 2 −j )) ≤ 2 −j(1−ε) by (15), (17), (21), (23) and the same bounds hold at Step 3 and 4. Henc… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Two views of the Legendre spectrum when 0 < p1 = 0.2 < 1/2 < p2 = 0.8 with η = 0.5. ε > 0, if ε −2 ≤ ℓ ≤ j − j ε , then ℓ + ℓ 1/ √ log ℓ ≤ ℓ + ℓ ε ≤ j and (38) (applied with ℓ + 2 instead of J and j − 1 instead of J + j − 1) gives for j large that (51) |#Aℓ+2,j−1 − η(j…
Figure 5
Figure 5. Figure 5: Top: Plot of D1 and D2 (equations (55) and (56)) when η = 0.7, p1 = 0.4 et p2 = 0.8. Bottom: Superposition of D1 and D2, and bivariate multifractal spectrum of (µp1 , νη) (i.e. minimum of D1 and D2). Consider the interval Ie∈ {I, I+, I−} such that νη(Iej (x)) is maxima…
Figure 6
Figure 6. Figure 6: Two views of the superposition of the Legendre (in black) and multifractal (in blue) spectrum of the pair (µp1 , νη) for p1 = 0.2, p2 = 0.8, η = 0.4. The two spectra coincide on a region of the plane, and Dµp1 ,νη has a wider support than τ ∗ µp1 ,νη . References [1] P…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 31 canonical work pages

  1. [19]

    Jaffard, S

    S. Jaffard, S. Seuret, H. Wendt, R. Leonarduzzi, S. Roux, and P. Abry. Multivariate multifractal analysis. Applied and Computational Harmonic Analysis , 46(3):653–663, 2019

    work page 2019

  2. [1]

    P. Abry, S. Jaffard, R. Leonarduzzi, S. Seuret, and H. Wendt. Multifractal formalisms for multivariate analysis. Proceedings of the Royal Society A , 2229(475), September 2019

    work page 2019

  3. [2]

    Attia and J

    N. Attia and J. Barral. Hausdorff and packing spectra, large deviations and free energy for branching random walks in Rd. Comm. Math. Phys. , 331(139-187), 2014

    work page 2014

  4. [3]

    Barral, F

    J. Barral, F. Ben Nasr, and J. Peyri` ere. Comparing multifractal formalisms: the neighboring condition. Asian J. Math. , 7:149–166, 2003

    work page 2003

  5. [4]

    Barral and D.-J

    J. Barral and D.-J. Feng. On multifractal formalism for self-similar measures with overlaps. Math. Z. , 298(1-2):359–383, 2021

    work page 2021

  6. [5]

    Barral and D.J

    J. Barral and D.J. Feng. Multifractal formalism for almost all self-affine measures. Comm. Math. Phys., 318:473–504, 2013

    work page 2013

  7. [6]

    Barral and B

    J. Barral and B. Mandelbrot. Random multiplicative multifractal measures. Proc. Symp. Pure Math., M. L. Lapidus and M. van Frankenhuysen eds , 72:1–52, 2004

    work page 2004

  8. [7]

    Barreira and B

    L. Barreira and B. Saussol. Variational principles and mixed multivariate spectra. Trans. A. M. S., 353 (10):3919–3944, 2001

    work page 2001

Show all 31 references
  1. [8]

    Barreira, B

    L. Barreira, B. Saussol, and J. Schmeling. Higher-dimensional multifractal analysis. J. Maths Pures Appl., 81:67–91, 2002

    2002

  2. [9]

    F. Bayart. Multifractal spectra of typical and prevalent measures. Nonlinearity, 26:353–367, 2013

    2013

  3. [10]

    Brown, G

    G. Brown, G. Michon, and J. Peyri` ere. On the multifractal analysis of measures.J. Stat. Phys. , 66:775– 790, 1992

    1992

  4. [11]

    Buczolich and S

    Z. Buczolich and S. Seuret. Typical Borel measures on [0 , 1]d satisfy a multifractal formalism. Nonlin- earity., 23(11):7–13, 2010

    2010

  5. [12]

    Falconer

    K. Falconer. Fractal Geometry: Mathematical Foundations and Applications. John Wiley & Sons, West Sussex, England, 1993

    1993

  6. [13]

    A.-H. Fan. Multifractal analysis of weighted ergodic averages. Adv. Math., 377, 2021

    2021

  7. [14]

    Fan, D.-J

    A.-H. Fan, D.-J. Feng, and J.J. Wu. Recurrence, dimension and entropy. J. London Math. Soc., 64:229– 244, 2001

    2001

  8. [15]

    Feng and K.S

    D.-J. Feng and K.S. Lau. Multifractal formalism for self-similar measures with weak separation condi- tion. J. Math. Pures Appl. , 92(3):407–428, 2009

    2009

  9. [16]

    D.J. Feng. Gibbs properties of self-conformal measures and the multifractal formalism. Ergodic Theory Dynam. Systems , 27(3):787–812, 2007. 38 ST ´EPHANE SEURET

    2007

  10. [17]

    Enrico Fermi

    U. Frisch and D. Parisi. Fully developed turbulence and intermittency in turbulence, and predictability in geophysical fluid dynamics and climate dynamics. International school of Physics “Enrico Fermi”, course 88, edited by M. Ghil, North Holland , pages 84–88, 1985

    1985

  11. [18]

    Halsey, M.H

    T.C. Halsey, M.H. Jensen, L.P. Kadanoff, I. Procaccia, and B.I. Shraiman. Fractal measures and their singularities: the characterisation of strange sets. Phys. Rev. A , 33:1141–1151, 1986

    1986

  12. [20]

    Joint multifractal analysis based on wavelet leaders

    Zhi-Qiang Jiang, Yan-Hong Yang, Gang-Jin Wang, and Wei-Xing Zhou. Joint multifractal analysis based on wavelet leaders. Frontiers of Physics, 12, 2017

    2017

  13. [21]

    Lau and S.-M

    K.S. Lau and S.-M. Ngai. Multifractal measures and a weak separation condition. Adv. Math. , 141(1):45–96, 1999

    1999

  14. [22]

    Marstrand

    J.M. Marstrand. Some fundamental geometrical properties of plane sets of fractional dimensions. Lon- don Math. Soc. Proc. , 3(4):257–302, 1954

    1954

  15. [23]

    P. Mattila. Hausdorff dimension, orthogonal projections and intersections with planes. Ann. Acad. Sci. Fenn. Math., 1:227–244, 1975

    1975

  16. [24]

    P. Mattila. Geometry of Sets and Measures in Euclidean Spaces . Cambridge University Press, 1995

    1995

  17. [25]

    Meneveau, K.R

    C. Meneveau, K.R. Sreenivasan, P. Kailasnath, and M.S. Fan. Joint multifractal measures - theory and applications to turbulence. Physical Review A , 41(2):894–913, January 1990

    1990

  18. [26]

    L. Olsen. A multifractal formalism. Adv. Math., 116:92–195, 1995

    1995

  19. [27]

    L. Olsen. Mixed divergence points of self-similar measures. Indiana University Mathematics Journa , 52:1343–1372, 2003

    2003

  20. [28]

    L. Olsen. Mixed generalized dimensions of self-similar measures. J. Math. Anal. Appl. , 306:516–539, 2005

    2005

  21. [29]

    Shmerkin

    P. Shmerkin. On Furstenberg’s intersection conjecture, self-similar measures, and the lq norms of convolutions. Ann. Math., 189(2):319–391, 2019

    2019

  22. [30]

    Wendt, P

    H. Wendt, P. Abry, and S. Jaffard. Bootstrap for empirical multifractal analysis. IEEE Signal Process. Mag., 24(4):38–48, 2007

    2007

  23. [31]

    M. Wu. A proof of Furstenberg’s conjecture on the intersections of ×p and ×q-invariant sets. Ann. Math., 189(3):707–751, 2019. St´ephane Seuret, Univ Paris Est Creteil, Univ Gustave Eiffel, CNRS, LAMA UMR8050, F-94010 Creteil, France Email address : seuret@u-pec.fr

    2019

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.