The Setting
Conditioning affine words on their symbol types leaves a finite distribution of translations. The entropy lost when these translations are observed through a fine grid is compared with the mass that can be assigned to pairs of nearby translations. Words are retained according to a linear bound on their contraction depth.
A weighted affine iterated function system on a finite alphabet consists of maps and probabilities satisfying (1), with signed slopes and arbitrary real translations : (1)✦ The complete affine map of a word is its composition in address order.✦ Its signed slope , translation , word probability and contraction depth are given by (2): (2)
Conditional Laws and Entropy
For real weights on a finite set , their Shannon entropy in nats is in (3).✦ The fiber mass and entropy of an observable on are defined in (3), with zero contribution at zero mass. For the observable distribution below, the weights are strictly positive and have total mass one. (3) A finitely supported probability law on consists of a finite support and strictly positive point masses summing to one.✦ For a real-valued observable and strictly positive , the observable distribution has support and point masses .✦
A tag map assigns a type to each symbol. The word tag sequence of is .✦ The tag fiber of a word is the set of words of the same length whose tag sequence equals ✦, and the fiber mass is the sum of their word probabilities.✦ The type-conditioned translation law of a word is the conditional observable distribution of translation on the tag fiber , defined in (4): (4)✦
For a finite probability law and a mesh , the grid-cell mass is its mass in the cell indexed by . Its Shannon entropy and grid entropy in bits are defined by (5): (5) The conditional grid entropy deficit of the type-conditioned translation law at mesh is the loss in nats defined by (6): (6)✦ For and , the fine scale and first dyadic band are defined by (7): (7) The retained entropy loss for is the weighted sum of deficits on words satisfying the depth cutoff, as in (8): (8)✦
Pair Capacity
For a finite probability law and an integer , the pair set is the set of unordered pairs of atoms satisfying the distance bounds in (9).✦ The incident mass of weights on these pairs is one half of the total weight of pairs containing : (9)✦ Feasible pair weights on are nonnegative weights whose incident mass at each atom is at most its probability.✦ The fair-pair capacity of a finite probability law in band is the supremum of the total weights of feasible pair weights on , as in (10): (10)✦ The weighted pair budget on an integer band is the sum in (11), with the same depth cutoff as the retained loss: (11)✦
The Claim
Theorem 1. Let have constant signed slopes on its fibers: symbols with the same tag have the same . Fix . For and an integer , the estimate (12) holds simultaneously for all words with a common integer upper band index : (12)✦ If and satisfy for every , then there is such that, for each , an integer satisfies (13): (13)✦
In the paper, the result reads: under all the assumptions and notation of Lemma 4.1, there are and such that for each a deterministic finite integer satisfies , with independent of and and no asserted growth bound on .
What Is Checked
The claims are stated over finite weighted affine systems and tag maps that preserve the signed contraction ratio, using conditional translation laws and the depth cutoff in (8).