New bounds for bilinear Calderón-Zygmund operators and applications (Q1989544)

The authors prove the pointwise control by sparse operators (with precise constants), of bilinear Calderón-Zygmund operators with Dini-continuous kernels. The case for linear operators was recently obtained by [\textit{M. T. Lacey}, Isr. J. Math. 217, 181--195 (2017; Zbl 1368.42016)] and [\textit{T.P. Hytönen}, \textit{L. Roncal} and \textit{O. Tapiola}, Isr. J. Math. 218, 133--164 (2017; Zbl 1368.42015)]. Under weaker assumptions on the kernels, the results were already known in the multilinear setting (see [\textit{A. Lerner} and \textit{F. Nazarov}, ``Intuitive dyadic calculus: the basics'', to appear in Expo. Math., \url{doi:10.1016/j.exmath.2018.01.001})] and [\textit{J.M. Conde-Alonso} and \textit{G. Rey}, Math. Ann. 365, No. 3-4, 1111--1135 (2016; Zbl 1353.42011)]). They also prove several mixed bounds for a general class of bilinear dyadic positive operators using the parallel stopping cubes technique. As a consequence they are able to obtain new precise weighted bounds for the commutators of these multilinear Calderón-Zygmund operators with BMO functions, square functions and Fourier multipliers. Their main result reads as follows. Let \(T\) be a bilinear Calderón-Zygmund operator with Dini continuous kernel \(K\), with respect to the modulus of continuity \(\omega\), and let \[ T_{\sharp}(f_1,f_2)(x)=\sup_{\varepsilon>0}|T_{\varepsilon }(f_1,f_2)(x)| \] be the maximal truncation of \(T\), where \(T_{\varepsilon}\) is a suitable truncation of \(T\). Assume that \(1\leq q_1,q_2<\infty\) and \(1/q=1/{q_1}+1/{q_2}\). Then, for every pair of compactly supported functions \(f_1,f_2\in L^1(\mathbb R^n)\), there exist sparse collections \(\mathcal{S}^u\subset \mathcal{D}^u\), \(u=1,2,\dots,3^n\), such that \[ T_{\sharp}(f_1,f_2)(x)\leq c_n(\|T\|_{L^{q_1}\times L^{q_2}\rightarrow L^q}+C_K+\|\omega\|_{\text{Dini}})\sum_{u=1}^{3^n}\mathcal{A}_{\mathcal{S}^u}(|f_1|,|f_2|)(x), \] for almost every \(x\in\mathbb R^n\), where \[ \mathcal{A}_{\mathcal{S}^u}(f_1,f_2)(x)=\sum_{Q\in \mathcal{S}^u}\bigg(\frac1{|Q|}\int_Qf_1(y)\,dy\bigg)\bigg(\frac1{|Q|}\int_Qf_2(y)\,dy\bigg)\mathbf{1}_Q(x). \]