arxiv: 2604.11273 · v1 · submitted 2026-04-13 · 🧮 math.FA · math.PR

Recognition: unknown

A twosided linear estimate and a dyadic reduction of the UMD Conjecture

Komla Domelevo , Stefanie Petermichl

Authors on Pith no claims yet

Pith reviewed 2026-05-10 15:40 UTC · model grok-4.3

classification 🧮 math.FA math.PR

keywords UMD conjectureHilbert transformdyadic shiftBanach space valued functionsL^p boundednessmartingale transformsantiinvolution

0 comments

The pith

The Hilbert transform on Banach-valued L^p functions is bounded exactly when a certain dyadic shift operator is, with linear norm control in both directions.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The authors introduce a time-faithful dyadic shift operator of complexity one that is antisymmetric and an antiinvolution. They prove that this operator is bounded on L^p if and only if the Hilbert transform taking values in the same Banach space is bounded, and the equivalence carries linear two-sided estimates on the operator norms. Because the UMD conjecture is already known to be equivalent to Hilbert-transform boundedness, the result collapses the conjecture to the question of whether two elementary dyadic operators are bounded. A reader interested in functional analysis would care because the continuous integral operator is replaced by a discrete, finite-complexity object that may be easier to analyze directly.

Core claim

We define a time faithful dyadic shift operator of complexity one, that is an antisymmetric antiinvolution. We show that the Hilbert transform with values in a Banach space is L^p bounded if and only if the dyadic shift is—with a linear two sided norm dependence. The results reduce the famous UMD conjecture to a pair of simple dyadic operators.

What carries the argument

The time-faithful dyadic shift operator of complexity one, an antisymmetric antiinvolution whose L^p boundedness is proved equivalent to that of the Hilbert transform.

If this is right

Boundedness of the dyadic shift implies the UMD property for the underlying Banach space.
Linear norm estimates give explicit comparison between the constants for the two operators.
The UMD conjecture now reduces to checking boundedness of two fixed dyadic operators rather than the continuous Hilbert transform.
Any proof or counterexample for the dyadic shift immediately settles the corresponding statement for the Hilbert transform.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Numerical checks on large but finite dyadic grids could provide evidence for or against the conjecture before a full proof is found.
The same reduction technique might apply to other singular integral operators valued in Banach spaces.
If the dyadic shift fails to be bounded on some space, it would immediately give a concrete counterexample to UMD for that space.

Load-bearing premise

The newly defined dyadic shift is truly an antisymmetric antiinvolution whose boundedness transfers exactly to the Hilbert transform under the usual L^p and Banach-space hypotheses.

What would settle it

Exhibit a Banach space X and p such that the Hilbert transform is bounded on L^p(X) but the dyadic shift is not, or vice versa.

Figures

Figures reproduced from arXiv: 2604.11273 by Komla Domelevo, Stefanie Petermichl.

**Figure 1.** Figure 1: Standard dyadic tosses (left) versus dyadic tosses adapted to S0 (right) 4. First L p estimates We have considered so far S0 acting on either functions defined on R or functions defined on a I0. In this section, we note specifically S R 0 : hI± → ±hI∓ , ∀I ∈ D. Further, let us denote S J 0 the operator acting on functions defined on J, defined by S J 0 : hI± → ±hI∓ , ∀I ⊆ J, S J 0 (1J ) = 0, S J 0 (hJ ) = … view at source ↗

**Figure 2.** Figure 2: Standard one–dimensional discrete random walk on the left. Each toss produces a left or right step. Two– dimensional discrete random walk with memory. Each toss outcome decides the direction, horizontal or vertical, of the next step of the random walk. we evaluate now, for any l ⩾ 1, S0dB 1 l (x) =√ 2δ X I∈D+ l |I| 1/2S0hI (x) = √ 2δ X I∈D− l |I| 1/2 (+hI (x)) = +dB 2 l (x) S0dB 2 l (x) =√ 2δ X I∈D− l |I| … view at source ↗

**Figure 3.** Figure 3: Trigonometric tosses adapted to S0 We will use expectation operators to denote integrals against probability measures. For a function f = f(x), with x ∈ I0 = [0, 1), and for a function F = F( ⃗θ) = F(θ0, . . . , θn), we note E x f := Z 1 0 f(x)dx, E θ⃗ F := Z 2π 0 . . . Z 2π 0 F(θ0, . . . , θn) dθ0 2π . . . dθn 2π . Remark 6.1. Using this notation, we have for the functions F( ⃗θ) and f(x) related like abo… view at source ↗

**Figure 4.** Figure 4: The constant c0. corresponding L 2 -normalized Haar functions and by S α,r 0 the dyadic Hilbert transform associated to Dα,r. Since in the usual sense, S α,r 0 : f(x) 7→ X I∈Dα,r [PITH_FULL_IMAGE:figures/full_fig_p031_4.png] view at source ↗

**Figure 5.** Figure 5: sign distribution Following the strategy of the second author in [21], the average of the kernel by dilation and translation we consider is: ErEαK α,r 0 (t, x) = 1 log 2 Z 2 1 lim R→∞ 1 2R Z R −R K α,r 0 (t, x)dα dr r . Via the explanations in [21], the resulting average depends upon t − x, is antisymmetric and of homogeneity −1, such as uniquely satisfied by the function c x−t for some c ∈ R. This can be … view at source ↗

read the original abstract

We define a time faithful dyadic shift operator of complexity one, that is an antisymmetric antiinvolution. We show that the Hilbert transform with values in a Banach space is $L^p$ bounded if and only if the dyadic shift is -- with a linear two sided norm dependence. The results reduce the famous UMD conjecture to a pair of simple dyadic operators.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This paper gives a direct linear-norm equivalence between the Hilbert transform and a new time-faithful dyadic shift that is an antisymmetric antiinvolution, reducing the UMD conjecture to boundedness of two simple dyadic operators.

read the letter

The central contribution is the explicit construction of the time-faithful dyadic shift of complexity one together with the if-and-only-if statement that its L^p(X) boundedness is equivalent to that of the Hilbert transform, with constants linear in both directions. The proof proceeds by direct kernel comparison and by using the antiinvolution property to transfer boundedness one way and the other. No extra logarithmic factors or hidden assumptions on X appear. This is new; earlier reductions to dyadic models did not isolate an operator with these exact algebraic features or achieve linear dependence. The algebraic verification that the shift is an antisymmetric antiinvolution is clean and self-contained. The reduction itself is therefore sharp and lands exactly as claimed. The remaining task is to decide boundedness of the dyadic shift, which is formally simpler but still nontrivial. The paper does not claim to settle that question, only to move it to a dyadic setting where combinatorial arguments may be easier to apply. Readers working on UMD spaces, martingale transforms, or Banach-valued harmonic analysis will find the equivalence useful because it isolates the essential difficulty without extra overhead. The citation pattern is appropriate and the arguments are internally consistent. I would bring this to a reading group and would cite it if I were working on related operator bounds. It deserves peer review because the reduction is precise and the details check out.

Referee Report

0 major / 2 minor

Summary. The paper defines a time-faithful dyadic shift operator of complexity one that is an antisymmetric antiinvolution. It proves that the L^p(X)-boundedness of the Hilbert transform on a Banach space X is equivalent to the boundedness of this dyadic shift, with linear two-sided norm dependence in both directions. The equivalence is obtained by direct kernel comparison and by using the antiinvolution property to transfer boundedness. This reduces the UMD conjecture to the boundedness of a pair of simple dyadic operators.

Significance. If the claimed equivalence holds, the result is significant because it supplies an explicit, parameter-free reduction of the UMD property to dyadic operators with linear norm control and no hidden logarithmic factors. The manuscript provides explicit definitions, verifies the algebraic properties (antisymmetry and antiinvolution), and derives both directions of the equivalence. These features make the reduction falsifiable and potentially useful for further study of UMD spaces.

minor comments (2)

The abstract refers to reduction 'to a pair of simple dyadic operators,' but the main text centers on a single shift operator; clarify whether the second operator is the adjoint, the complex conjugate, or another explicitly defined object, and state the corresponding norm equivalence.
Notation for the time-faithful dyadic shift (e.g., its kernel or action on intervals) should be introduced with a short comparison to the standard dyadic shift to highlight the 'time-faithful' and 'complexity one' modifications.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive summary, significance assessment, and recommendation of minor revision. No specific major comments were provided in the report, so we have no points to address point-by-point. We will incorporate any minor editorial changes in the revised version.

Circularity Check

0 steps flagged

No significant circularity; derivation self-contained via explicit definitions and direct equivalences

full rationale

The paper introduces a newly defined time-faithful dyadic shift operator of complexity one with explicit algebraic properties (antisymmetric antiinvolution) and derives the if-and-only-if L^p boundedness equivalence to the Hilbert transform through direct kernel comparisons and operator manipulations. No fitted parameters, self-referential definitions, or load-bearing self-citations reduce the central claim to its inputs; the reduction is constructed from the paper's own definitions and standard Banach space assumptions without circular collapse. This is the expected non-circular outcome for a direct equivalence proof.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The claim rests on the definition of a new operator and standard background assumptions in vector-valued harmonic analysis; no free parameters or additional invented entities beyond the defined operator are indicated.

axioms (1)

domain assumption Standard properties of the Hilbert transform and L^p spaces over Banach spaces hold in the usual way.
The boundedness statements presuppose the conventional setup for vector-valued operators.

invented entities (1)

time faithful dyadic shift operator of complexity one no independent evidence
purpose: Provides a discrete antisymmetric antiinvolution that serves as an equivalent simpler model for the Hilbert transform.
Newly defined in the paper; no independent evidence outside the definition is given.

pith-pipeline@v0.9.0 · 5353 in / 1295 out tokens · 48243 ms · 2026-05-10T15:40:50.449893+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

31 extracted references · 1 canonical work pages

[1]

Ba˜ nuelos, M

R. Ba˜ nuelos, M. Kwa´ snicki.On theℓp norm of the discrete Hilbert transform, Duke Math. J. 168(3): pp. 471–504, 2019

2019
[2]

Ba˜ nuelos, G

R. Ba˜ nuelos, G. Wang.Orthogonal martingales under differential subordination and applications to Riesz transforms, I llinois J. Math., 40(4): pp. 678–691, 1996

1996
[3]

Bourgain.Some remarks on Banach spaces in which martingale difference sequences are unconditional, Ark

J. Bourgain.Some remarks on Banach spaces in which martingale difference sequences are unconditional, Ark. Mat. 21(2): pp. 163–168, 1983

1983
[4]

D. L. Burkholder.A geometric condition that implies the existence of certain singular integrals of Banach-space-valued functions, Conference on harmonic analysis in honor of Antoni Zygmund, Vol. I, II (Chicago, Ill., 1981), Wadsworth Math. Ser., 270–286. Wadsworth, Belmont, CA, 1983

1981
[5]

D. L. Burkholder.Boundary value problems and sharp inequalities for martingale transforms, Ann. Probab., 12(3): pp. 647–702, 1984

1984
[6]

D. L. Burkholder.Martingales and singular integrals in Banach spaces, Handbook of the geometry of Banach spaces, Vol. I, North-Holland, Amsterdam, 2001

2001
[7]

Domelevo, S

K. Domelevo, S. Petermichl.SharpL p estimates for discrete second order Riesz trans- forms, Adv. Math., 262: pp. 932–952, 2014

2014
[8]

Domelevo, S

K. Domelevo, S. Petermichl, S. Treil, A. Volberg.The matrixA 2 conjecture fails, i.e. 3/2>1, arXiv:2402.06961, pp. 1–46, 2024

work page arXiv 2024
[9]

Ess´ en.A superharmonic proof of the M

M. Ess´ en.A superharmonic proof of the M. Riesz conjugate function theorem, Ark. Mat., 22(2): pp. 241–249, 1984

1984
[10]

Figiel.Singular integral operators: a martingale approach, London Math

T. Figiel.Singular integral operators: a martingale approach, London Math. Soc. Lecture Note Ser., 158, Cambridge Univ. Press, Cambridge, 1990

1990
[11]

Geiss, S

S. Geiss, S. Montgomery-Smith, E. Saksman.On singular integral and martingale transforms, Trans. Amer. Math. Soc., 362(2): pp. 553–575, 2010

2010
[12]

Hyt¨ onen.The sharp weighted bound for general Calder´ on–Zygmund operatorsAnn

T. Hyt¨ onen.The sharp weighted bound for general Calder´ on–Zygmund operatorsAnn. Math. 175(3): pp. 1473–1506, 2012. DYADIC REDUCTION OF THE UMD CONJECTURE 33

2012
[13]

Hyt¨ onen, J

T. Hyt¨ onen, J. van Neerven, M. Veraar, L. Weis.Analysis in Banach spaces, Vol. I. Martingales and Littlewood-Paley theory, volume 63 of Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. Springer, Cham, 2016

2016
[14]

Small step

S. Kakaroumpas, S. Treil.“Small step” remodeling and counterexamples for weighted estimates with arbitrarily “smooth” weights, Adv. Math., 376: pp. 1–52, 2021

2021
[15]

P. E. Kloeden, E. Platen.Numerical solution of stochastic differential equations, vol- ume 23 of Applications of Mathematics (New York). Springer-Verlag, Berlin, 1992

1992
[16]

M. T. Lacey.Two-weight inequality for the Hilbert transform: A real variable char- acterization, II, Duke Math. J. 163(15): pp. 2821–2840, 2014

2014
[17]

G. F. Lawler.Random walk and the heat equation, volume 55 of Student Mathematical Library. American Mathematical Society, Providence, RI, 2010

2010
[18]

Nazarov.A counterexample to Sarason’s conjecture, unpublished manuscript, avail- able athttp://users.math.msu.edu/users/fedja/prepr.html

F. Nazarov.A counterexample to Sarason’s conjecture, unpublished manuscript, avail- able athttp://users.math.msu.edu/users/fedja/prepr.html
[19]

Nazarov, G

F. Nazarov, G. Pisier, S. Treil, A. Volberg.Sharp estimates in vector Carleson imbed- ding theorem and for vector paraproducts, J. reine angew. Math. 542: pp. 147–171, 2002

2002
[20]

Nazarov, S

F. Nazarov, S. Treil, A. Volberg.The Tb-theorem on non-homogeneous spaces, Acta Math. 190(2): pp. 151–239, 2003

2003
[21]

Petermichl.Dyadic shift and a logarithmic estimate for Hankel operators with ma- trix symbol, C

S. Petermichl.Dyadic shift and a logarithmic estimate for Hankel operators with ma- trix symbol, C. R. Acad. Sci. Paris S´ er. I Math., 330(6): pp. 455–460, 2000

2000
[22]

Petermichl, S

S. Petermichl, S. Pott.A version of Burkholder’s theorem for operator-weighted spaces, Proc. Amer. Math. Soc., 131(11): pp. 3457–3461, 2003

2003
[23]

Petermichl.The sharp bound for the Hilbert transform on weighted Lebesgue spaces in terms of the classicalA p characteristic, Amer

S. Petermichl.The sharp bound for the Hilbert transform on weighted Lebesgue spaces in terms of the classicalA p characteristic, Amer. J. Math. 129, no. 5, pp. 1355–1375, 2007

2007
[24]

Petermichl, S

S. Petermichl, S. Treil, A. Volberg.Why the Riesz transforms are averages of the dyadic shifts?Publ. Mat. 46, pp. 209–228, 2002

2002
[25]

Petermichl, A

S. Petermichl, A. Volberg.Heating of the Ahlfors-Beurling operator: weakly quasireg- ular maps on the plane are quasiregular, Duke Math. J. 112, no. 2: pp. 281–305, 2002

2002
[26]

S. K. Pichorides.On the best values of the constants in the theorems of M. Riesz, Zygmund and Kolmogorov, Studia Math., 44: pp. 165–179. (errata insert), 1972

1972
[27]

Pisier.Martingales in Banach spaces, Cambridge Studies in Advanced Mathemat- ics, vol

G. Pisier.Martingales in Banach spaces, Cambridge Studies in Advanced Mathemat- ics, vol. 155, Cambridge University Press, Cambridge, 2016

2016
[28]

S. Pott, A. Stoica.Linear bounds for Calder´ on-Zygmund operators with even kernel on UMD spaces, J. Funct. Anal., 266(5): pp. 3303–3319, 2014

2014
[29]

Talay.Discr´ etisation d’une ´ equation diff´ erentielle stochastique et calcul approch´ e d’esp´ erances de fonctionnelles de la solution, RAIRO Mod´ el

D. Talay.Discr´ etisation d’une ´ equation diff´ erentielle stochastique et calcul approch´ e d’esp´ erances de fonctionnelles de la solution, RAIRO Mod´ el. Math. Anal. Num´ er., 20(1): pp. 141–179, 1986

1986
[30]

Treil.SharpA 2 estimates of Haar shifts via Bellman function, Recent trends in analysis

S. Treil.SharpA 2 estimates of Haar shifts via Bellman function, Recent trends in analysis. Proceedings of the conference in honor of Nikolai Nikolski on the occasion of his 70th birthday, Bucharest: The Theta Foundation, pp. 187–208, 2013

2013
[31]

Treil, A

S. Treil, A. Volberg.Wavelets and the angle between past and future, J. Funct. Anal., 143(2): pp. 269–308, 1997. Institute of Mathematics, University of W ¨urzburg, Germany Institute of Mathematics, University of W ¨urzburg, Germany

1997