Searcharxiv⌕ Search

arXiv subjects

Sergei Treil

Publications and source records attributed to Sergei Treil.

At least 37 records · Page 2Linked to original sources
arXiv2017

General Clark model for finite rank perturbations

All unitary perturbations of a given unitary operator $U$ by finite rank $d$ operators with fixed range can be parametrized by $(d\times d)$ unitary matrices $Γ$; this generalizes unitary rank one ($d=1$) perturbations, where the Aleksandrov--Clark family of unitary perturbations is parametrized by the scalars on the unit circle $\mathbb{T}\subset\mathbb{C}$. For a purely contractive $Γ$ the resulting perturbed operator $T_Γ$ is a contraction (a completely non-unitary contraction under the natural assumption about cyclicity of the range), so they admit the functional model. In this paper we investigate the Clark operator, i.e. a unitary operator that intertwines $T_Γ$ (presented in the spectral representation of the non-perturbed operator $U$) and its model. We make no assumptions on the spectral type of the unitary operator $U$; absolutely continuous spectrum may be present. We find a representation of the adjoint Clark operator in the coordinate free Nikolski--Vasyunin functional model. This representation features a special version of the vector-valued Cauchy integral operator. Regularization of this singular integral operator yield representations of the adjoint Clark operator in the Sz.-Nagy--Foias transcription. In the special case of inner characteristic functions (purely singular spectral measure of $U$) this representation gives what can be considered as a natural generalization of the normalized Cauchy transform (which is a prominent object in the Clark theory for rank one case) to the vector-valued settings.

math.FA↗
arXiv2017

Convex body domination and weighted estimates with matrix weights

We introduce the so called convex body valued sparse operators, which generalize the notion of sparse operators to the case of spaces of vector valued functions. We prove that Calderón--Zygmund operators as well as Haar shifts and paraproducts can be dominated by such operators. By estimating sparse operators we obtain weighted estimates with matrix weights. We get two weight $A_2$-$A_\infty$ estimates, that in the one weight case give us the estimate $$ \|T\|_{L^2(W)\to L^2 (W)} \le C [W]_{A_2}^{1/2} [W]_{A_\infty} \le C[W]_{A_2}^{3/2} $$ where $T$ is either a Calderon--Zygmund operator (with modulus of continuity satisfying the Dini condition), or a Haar shift or a paraproduct.

math.CA↗
arXiv2016

Two weight estimates with matrix measures for well localized operators

In this paper, we give necessary and sufficient conditions for weighted $L^2$ estimates with matrix-valued measures of well localized operators. Namely, we seek estimates of the form: \[ \| T(\mathbf{W} f)\|_{L^2(\mathbf{V})} \le C\|f\|_{L^2(\mathbf{W})} \] where $T$ is formally an integral operator with additional structure, $\mathbf{W}, \mathbf{V}$ are matrix measures, and the underlying measure space possesses a filtration. The characterization we obtain is of Sawyer-type; in particular we show that certain natural testing conditions obtained by studying the operator and its adjoint on indicator functions suffice to determine boundedness. Working in both the matrix weighted setting and the setting of measure spaces with arbitrary filtrations requires novel modifications of a T1 proof strategy; a particular benefit of this level of generality is that we obtain polynomial estimates on the complexity of certain Haar shift operators.

math.FA↗
arXiv2015

Two weight $L^p$ estimates for paraproducts in non-homogeneous settings

We give a necessary and sufficient condition for the two weight $L^p$-estimates for paraproducts in non-homogeneous settings, $1 2$ both "direct" and "adjoint" testing conditions are needed. An interesting feature is that the "adjoint" testing condition is that it is a testing condition not for the adjoint of the paraproduct, but for some auxiliary operator.

math.CA↗
arXiv2015

Singular integrals, rank one perturbations and Clark model in general situation

We start with considering rank one self-adjoint perturbations $A_α= A+α(\,\cdot\,,φ)φ$ with cyclic vector $φ\in \mathcal{H}$ on a separable Hilbert space $\mathcal H$. The spectral representation of the perturbed operator $A_α$ is realized by a (unitary) operator of a special type: the Hilbert transform in the two-weight setting, the weights being spectral measures of the operators $A$ and $A_α$. Similar results will be presented for unitary rank one perturbations of unitary operators, leading to singular integral operators on the circle. This motivates the study of abstract singular integral operators, in particular the regularization of such operator in very general settings. Further, starting with contractive rank one perturbations we present the Clark theory for arbitrary spectral measures (i.e. for arbitrary, possibly not inner characteristic functions). We present a description of the Clark operator and its adjoint in the general settings. Singular integral operators, in particular the so-called normalized Cauchy transform again plays a prominent role. Finally, we present a possible way to construct the Clark theory for dissipative rank one perturbations of self-adjoint operators. These lecture notes give an account of the mini-course delivered by the authors at the Thirteenth New Mexico Analysis Seminar and Afternoon in Honor of Cora Sadosky. Unpublished results are restricted to the last part of this manuscript.

math.FA↗
arXiv2014

Entropy conditions in two weight inequalities for singular integral operators

The new type of "bumping" of the Muckenhoupt $A_2$ condition on weights is introduced. It is based on bumping the entropy integral of the weights. In particular, one gets (assuming mild regularity conditions on the corresponding Young functions) the bump conjecture, proved earlier by A. Lerner and independently by Nazarov--Reznikov--Treil--Volberg, as a corollary of entropy bumping. But our entropy bumps cannot be reduced to the bumping with Orlicz norms in the solution of bump conjecture, they are effectively smaller. Henceforth we get somewhat stronger result than the one that solves the bump conjecture. New results concerning one sided bumping conjecture are obtained. All the results hold in the general non-homogeneous situation.

math.CA↗
arXiv2014

A Bellman function proof of the $L^2$ bump conjecture

We approach the problem of finding the sharp sufficient condition of the boundedness of all two weight Calderon--Zygmund operators. We solve this problem in $L^2$ by writing a formula for a Bellman function of the problem.

math.CA↗
arXiv2013

Clark model in general situation

For a unitary operator the family of its unitary perturbations by rank one operators with fixed range is parametrized by a complex parameter $γ, |γ|=1$. Namely all such unitary perturbations are $U_γ:=U+(γ-1) (., b_1)_{\mathcal H} b$, where $b\in\mathcal H, \|b\|=1, b_1=U^{-1} b, |γ|=1$. For $|γ|<1$ operators $U_γ$ are contractions with one-dimensional defects. Restricting our attention on the non-trivial part of perturbation we assume that $b$ is cyclic for $U$. Then the operator $U_γ$, $|γ|<1$ is a completely non-unitary contraction, and thus unitarily equivalent to its functional model $\mathcal M_γ$, which is the compression of the multiplication by the independent variable $z$ onto the model space $\mathcal K_{θ_γ}$, where $θ_γ$ is the characteristic function of the contraction $U_γ$. The Clark operator $Φ_γ$ is a unitary operator intertwining $U_γ, |γ|<1$ and its model $\mathcal M_γ$, $\mathcal M_γΦ_γ= Φ_γU_γ$. If spectral measure of $U$ is purely singular (equivalently, $θ_γ$ is inner), operator $Φ_γ$ was described from a slightly different point of view by D. Clark. When $θ_γ$ is extreme point of the unit ball in $H^\infty$ was treated by D. Sarason using the sub-Hardy spaces introduced by L. de Branges. We treat the general case and give a systematic presentation of the subject. We find a formula for the adjoint operator $Φ^*_γ$ which is represented by a singular integral operator, generalizing the normalized Cauchy transform studied by A. Poltoratskii. We present a "universal" representation that works for any transcription of the functional model. We then give the formulas adapted for the Sz.-Nagy--Foias and de Branges--Rovnyak transcriptions, and finally obtain the representation of $Φ_γ$.

math.FA↗
arXiv2012

Carleson--Buckley measures beyond the scope of $A_\infty$ and their applications

Carleson measures are ubiquitous in Harmonic Analysis. In the paper of Fefferman--Kenig--Pipher in 1991 an interesting class of Carleson measures was introduced for the need of regularity problems of elliptic PDE. These Carleson measures were associated with $A_\infty$ weights. In discrete setting (we need exactly discrete setting here) they were studied by Buckley's, where they were associated with dyadic $A\infty^d$. Our goal here is to show that such Carleson--Buckley measures (in discrete setting) exists for virtually any positive function (weight). Of course some modification is needed, because it is known that Carleson property of Buckley's measure are equivalent to the weight to be in $A_\infty^d$. However a very natural generalization of those facts exist for weights more general $A_\infty$, and of course, in a special case of $A_\infty^d$ it gives Buckley's results. Our generalization of Buckley's inequality beyond the scope of $A_\infty$ allows us to prove the so-called bump conjecture.

math.CA↗
arXiv2012

Corona Solutions Depending Smoothly on Corona Data

In this note we show that if the Corona data depends continuously (smoothly) on a parameter, the solutions of the corresponding Bezout equations can be chosen to have the same smoothness in the parameter.

math.CA↗
arXiv2012

Similarity of Operators in the Bergman Space Setting

We give a necessary and sufficient condition for an n-hypercontraction to be similar to the backward shift operator in a weighted Bergman space. This characterization serves as a generalization of the description given in the Hardy space setting, where the geometry of the eigenvector bundles of the operators is used.

math.FA↗
arXiv2011

Sharp $A_2$ estimates of Haar shifts via Bellman function

We use the Bellman function method to give an elementary proof of a sharp weighted estimate for the Haar shifts, which is linear in the $A_2$ norm of the weight and in the complexity of the shift. Together with the representation of a general Calderón--Zygmund operator as a weighted average (over all dyadic lattices) of Haar shifts, (cf. arXiv:1010.0755v2[math.CA], arXiv:1007.4330v1[math.CA]) it gives a significantly simpler proof of the so-called the $A_2$ conjecture. The main estimate is a very general fact about concave functions, which can be very useful in other problems of martingale Harmonic Analysis. Concave functions of such type appear as the Bellman functions for bounds on the bilinear form of martingale multipliers, thus the main estimate allows for the transference of the results for simplest possible martingale multipliers to more general martingale transforms. Note that (although this is not important for the $A_2$ conjecture for general Calderón--Zygmund operators) this elementary proof gives the best known (linear) growth in the complexity of the shift.

math.CA↗
arXiv2011

A sharp estimate of weighted dyadic shifts of complexity 0 and 1

A simple shortcut to proving sharp weighted estimates for the Martingale Transform and for the dyadic shift of order 1 (and so for the Hilbert transform) is presented. It is a unified proof for these both transforms. Key words: Calder\on--Zygmund operators, $A_2$ weights, $A_1$ weights, Carleson embedding theorem, Bellman function, dyadic shifts, nonhomogeneous Harmonic Analysis.

math.CA↗
arXiv2011

Regularizations of general singular integral operators

In the theory of singular integral operators significant effort is often required to rigorously define such an operator. This is due to the fact that the kernels of such operators are not locally integrable on the diagonal, so the integral formally defining the operator or its bilinear form is not well defined (the integrand is not in L^1) even for nice functions. However, since the kernel only has singularities on the diagonal, the bilinear form is well defined say for bounded compactly supported functions with separated supports. One of the standard ways to interpret the boundedness of a singular integral operators is to consider regularized kernels, where the cut-off function is zero in a neighborhood of the origin, so the corresponding regularized operators with kernel are well defined (at least on a dense set). Then one can ask about uniform boundedness of the regularized operators. For the standard regularizations one usually considers truncated operators. The main result of the paper is that for a wide class of singular integral operators (including the classical Calderon-Zygmund operators in non-homogeneous two weight settings), the L^p boundedness of the bilinear form on the compactly supported functions with separated supports (the so-called restricted L^p boundedness) implies the uniform L^p-boundedness of regularized operators for any reasonable choice of a smooth cut-off of the kernel. If the kernel satisfies some additional assumptions (which are satisfied for classical singular integral operators like Hilbert Transform, Cauchy Transform, Ahlfors--Beurling Transform, Generalized Riesz Transforms), then the restricted L^p boundedness also implies the uniform L^p boundedness of the classical truncated operators.

math.CA↗