The Setting
Let be random variables with values in a standard Borel space , indexed by a finite set , with joint law on . The item concerns two kinds of dependence among them. One is linear: which square-integrable functions of a single coordinate the other coordinates reach in the mean, the closed subspace of such functions being the active space at . The other is multiplicative: which circle-valued functions of the coordinates multiply to an almost surely constant product, such a tuple of functions being an exact tuple, and the tuple with each function taken up to a constant phase an exact line.
The linear question is also asked of arrays: for an array object, an array of order is a square of cells, and the cells are grouped into blocks of rows and columns, all defined below. The joinings are then conditional laws given a block of cells and the parameter array, which records the parameters of the entries of the array, so that a block carries a measurable family of Markov operators rather than a single one. The library defines the active space of the conditional law of an array of order one at a slot, that is, at an index , the row, column and common active spaces of a block of cells, and the exact tuples and exact lines of a law on . Square-integrable functions are complex-valued throughout.
Arrays and Array Objects
An array of order over assigns an entry in to each index tuple in ; the arrays form the product ✦, and an array of order zero is a point of ✦. The plane restriction of an array of order that fixes the axis at the index is the array of order whose entry at is the entry of at the tuple obtained by inserting into at position .✦
Initial array laws of a measurable automorphism of a standard Borel space are probability measures on , one for each , each invariant under applied entrywise, consistent, in that every plane restriction carries to , and proper: for every subset of other than , the planes of an array of law that fix the last axis at the indices in are independent, each of law .✦
An array object over the initial array laws of consists of a standard Borel space with an automorphism , a standard Borel space of parameters with an automorphism , measurable maps and , with intertwining with and intertwining with , and probability measures on , one for each .✦ The parameter array of an array is applied entrywise. Each is invariant under applied entrywise and is carried to by applied entrywise, and every plane restriction carries to . For every subset of other than , the planes of an array of law that fix the last axis at the indices in are independent, each of law , and they stay independent given the parameter array, each with the conditional law of an array of order given the matching plane of the parameter array; under , a law on , the maps and are independent.
For a parameter array , is the conditional law under of an array given that its parameter array is , a Markov kernel from to .✦
The Grouped Square
An array of order is a square of cells. The row of at is the plane restriction of fixing the first axis at , and the column of at is the plane restriction of fixing the last axis at ; both are arrays of order . The cell of at is the plane restriction of fixing the first axis at and the last axis at , so that its entries are . For a set of rows and a set of columns, is the block of cells on ✦, and the grouped data of on is the pair ✦.
For a parameter array , the row family kernel draws arrays of order , one for each , independently, from .✦ The law of cells is the law of the block on whose cell at is the plane of that fixes the last axis at .✦
For a set of rows and sets , , of columns, the row joining at grouped data , with a block on , is the row pair kernel there: the conditional law under of the pair of blocks given that and .✦
The parent rows of an array of order on a set are its rows , ✦, and the posterior kernel sends a parameter array and parents to the conditional law of under given that its parent rows on are ✦. For a set of columns and sets , , of rows, the column group kernel at grouped data , with a block on , draws arrays of order , one for each , independently, from the posterior given the parameter column and the parents .✦ The column joining at , also called the column pair kernel, is the law of the pair of blocks on and on whose cell at is the row of at .✦
The Definition
Definition 1. Let and be measure-preserving maps from to and to , and let , , be the pullback isometry of .✦ The joining operator of and , the Markov operator of the joining of and , is (1)✦ For a kernel from a space of parameters to , the family of joining operators assigns to each parameter the Markov operator from to of the joining of and .✦
Let be a measure on a space and let and be Hilbert spaces. The essential kernel of a family of bounded operators from to is the subspace of the vectors with for -almost every .✦ The essential range of a family of bounded operators from to is the orthogonal complement of the essential kernel of the adjoints: (2)✦
For a parameter tuple , let , a law on , and let and . The raw active space at the slot is (3)✦ The slot active space of at is the closure of the range of the joining operator of and , where is the block on whose single cell is : (4)✦
Fix . For a set of rows and sets , , of columns, the row joining operator at grouped data on is the Markov operator from to of the row joining there✦; at it is the fibre row operator ✦, and the row active space is (5)✦ For a set of columns and sets , , of rows, the column joining operator at grouped data on is the Markov operator from to of the column joining there✦; at it is the fibre column operator ✦, and the column active space is (6)✦ For sets , , of rows and , , of columns, the common active space of the block is (7)✦
Let be the group of measurable maps from to the circle under pointwise multiplication. For a law on , the set of exact tuples is (8)✦ For a law on , the constant phase subgroup consists of the maps -almost everywhere equal to a constant✦, and the local projective group is ✦, whose elements are the lines. For laws on , , with , write for the line of in . The tuple projectivization is the map from to that sends to .✦ The set of exact lines of is the image of under it: (9)✦
In the paper, the result reads: for the law on of a tuple , with the law of , the active space at an index is the closed span of the conditional expectations in , where runs over the functions of the other coordinates, an exact tuple is a tuple of maps with almost surely for some constant , and an exact line is a line occurring in an exact tuple, up to constants of modulus one.
What Is Checked
The active spaces (3) to (7) are stated over an array object with a finite index set , its laws and its conditional laws given parameter arrays, through the joining operators (1) and the essential ranges (2); the exact tuples (8) and the exact lines (9) are stated over a law on and laws on .
The library proves that the essential kernel is a linear subspace and that the measurable circle maps form a subgroup of all circle-valued maps on , with a subgroup of it. It proves that is a subgroup of and that the tuple projectivization is a homomorphism, whose image is a subgroup of .