The Setting
Output-dependent truncations and energy-jump arguments in the analysis of the triangular Hilbert transform combine discrete maximal operators with scale-masked matrix energies. On the integer line, centered averages over finite sets are maximized over radii bounded by the distance from the center to the finite set. On the matrix side, a half-open scale window masks the zero-vertex component of a three-vertex matrix family before the resulting heat Hessian costs are integrated over a constant-sum center plane and summed across the vertices.
Discrete Support and Zero Extension
The support radius of a finite subset at an integer is the nonnegative integer , taken to be when is empty.✦ The extension by zero of a function on a finite subset is the real-valued function on equal to for and for .✦
Matrix Energy and Triangle Block Operators
Index the three vertices of the triangle by the triangle vertex set under cyclic addition modulo . For any index set , write and for the spaces of complex and real square matrices indexed by . The triangle block space of an index set is the Cartesian product , equipped with coordinate projections and .✦
The real trace of a matrix is the real number .✦ The real Hilbert-Schmidt inner product of matrices is the real trace of .✦ The Sylvester operator of a matrix is the map sending to .✦ The Sylvester penalty of matrices is the real number .✦ The Sylvester variational functional of matrices is the real functional .✦ The root Hessian cost of matrices is the real quantity .✦ The Hessian quadratic cost of matrices is the root Hessian cost evaluated at the matrix square root and .✦
The block matrix assembly of a -indexed array of complex square matrices on is the complex matrix on with entry .✦ The triangle matrix blocks of a family of matrices in form the array whose block is when , when , and when .✦ The triangle matrix of a family in is the block matrix assembly .✦ The vertex block projection of a vertex is the diagonal complex matrix on whose diagonal entry at is when and otherwise.✦ The vertex star row of a family in at a vertex is the matrix product , isolating the block row indexed by .✦ The zero component replacement of a matrix family in by a matrix is the -indexed family in whose component at equals when and otherwise.✦
Gaussian Weights, Heat Costs, and Plane Integration
The Gaussian density of a variance is the real function .✦ The heat weight of a step size , a point , a center , and a variance is the real number .✦ The Gaussian score of a point , center , and variance is the real quotient .✦ The vertex heat weights of a step size , node coordinates , centers , and variances form the family of real functions indexed by with .✦ The vertex root matrix of a family of real node weights is the diagonal complex matrix on with diagonal entry at .✦
The vertex star base row of a family in , real node weights on , and a vertex is the matrix product .✦ The vertex star base column of a family in , real node weights on , and a vertex is the matrix product .✦ The fixed base matrix of a family in and real node weights on at vertices is the complex matrix on equal to when and to otherwise.✦
The weighted Gram matrix of a matrix and a weight function on an index set is the matrix product .✦ For index sets and , a step size , coordinates , a group assignment , centers , and variances , the grouped heat weight is the real function on sending to .✦ The grouped first variation weight for index sets and , , , , , and a group is the real function on equal at to when and otherwise.✦ The local Gram matrix of a family in , a step size , coordinates , centers , variances , and vertices is the weighted Gram matrix .✦ The local first variation matrix of a family in , a step size , coordinates , centers , variances , and vertices is the weighted Gram matrix .✦ The heat Hessian cost of a family in , a step size , coordinates , centers , variances , and vertices is the Hessian quadratic cost evaluated at and .✦
The constant-sum center plane of a sum level is the plane . For each vertex and pair , the plane coordinate chart is the triple equal to when , to when , and to when .✦ The plane integral of a function at a sum level is the Lebesgue integral .✦ The strict diffusion rate of a vertex is the real function on taking the value at and for .✦
The Definition
Definition 1. For an integer and a nonnegative integer radius , the discrete interval is defined by (1): (1)✦ Given a finite subset , a function , an integer , and a radius , the discrete interval average is given by (2): (2)✦ The discrete maximal operator of a finite subset and a function maps each to the maximum over radii bounded by the support radius in (3): (3)✦ The line maximal operator of a function on a finite subset is defined at through the extension by zero via (4): (4)✦ With real bounds , a scale , and a complex number , the window mask is the half-open scale mask in (5): (5)✦ Given a matrix family in , real bound matrices and in , and a scale , the masked array is the -indexed family in formed by zero component replacement in (6): (6)✦ From a matrix family in , bound matrices , a step size , coordinates , a vertex , and a scale , the masked heat cost is the sum of plane integrals at sum level defined by (7): (7)✦
In the paper, the result reads: and , and for entrywise bounds the masked zero component is with and . Unlike the paper’s formulas, (2) through (4) average directly over a finite set up to its support radius rather than over all of , while (5) and (6) replace the open interval by a half-open interval before forming (7).
What Is Checked
The discrete definitions (1) through (4) are stated over integers , radii , finite subsets with support radius , and real functions on or on via extension by zero, while the masked definitions (5) through (7) are stated over real scale bounds and parameters, complex values, an index set , -indexed families in , matrices in , a step size , node coordinates , a vertex with strict diffusion rate , and plane integrals of heat Hessian costs on .