The Setting
The question is whether a bound on logarithmic entropy growth controls the entropy increment at every dyadic scale. A weighted affine iterated function system on a finite alphabet consists of real similarities and probability weights satisfying (1); the slopes may have either sign, and the translations are arbitrary.✦ (1)
For a measure on , its weighted component is the weighted pushforward in (2).✦ A measure on is self-similar for when it satisfies the second identity in (2).✦ (2)
For a measure on a measurable space and a map , the fiber mass is the real value of the mass of ;✦ the partition entropy is the sum in (3).✦ The subscript denotes the real value of an extended nonnegative quantity, and has value zero. (3)
The shifted grid cell index assigns a point to an integer cell at scale and shift . For a measure on , the shifted grid entropy is the partition entropy of the shifted grid cell index , as specified in (4).✦ (4)
For a measure on , the averaged grid entropy is the Lebesgue average of the shifted grid entropy over shifts in .✦ For a measure on , the dyadic entropy deficit at scale is the difference between and the averaged entropy increment in (5).✦ $$\overline H(\eta,s)=\int_0^1 H_{s,u}(\eta)\,du,\qquad \Delta_s(\eta)=\ln 2-igl(\overline H(\eta,s)-\overline H(\eta,2s)\bigr).$$(5) These entropies and deficits use natural logarithms; division by gives their base-two normalization.
The Coding Exponent
The symbol probability measure of is the discrete measure on with masses , written in (6) using the unit point mass at .✦ The Bernoulli address measure on is the countable product of the symbol probability measure , as in (6).✦ (6)
For an address and a nonnegative integer , the prefix contraction product is the signed product in (7).✦ The scaled translation term for an address and a nonnegative integer is the product of the prefix contraction product and the translation ;✦ the coding map assigns to an address the sum of its scaled translation terms, as in (7).✦ (7)
The self-similar coding measure is the pushforward of the Bernoulli address measure under the coding map: (8)✦ Its weighted components are those of (2), applied to the measure in (8).
The rational radii are the positive rational numbers at most one.✦ For measures on , the maximal ball mass ratio is the supremum of their ratios on closed balls with radii in , as in (9), where .✦ (9)
For measures on , the positive logarithmic maximal envelope is the function on whose value at is the nonnegative part of the logarithm in (10).✦ (10)
The symbolic logarithmic maximal envelope on is the function whose value at an address is the positive logarithmic maximal envelope of relative to at , as in (11).✦ The logarithmic Radon-Nikodym derivative on is the real-valued function obtained by taking the logarithm of the real value of the Radon-Nikodym derivative of with respect to at , as in (11).✦ (11)
The clipped logarithmic density correction on is the function whose value at an address is truncated between zero and according to (12).✦ The symbol information observable on is the function whose value at an address is the negative logarithm of the initial symbol’s weight.✦ The projection information rate of is the expectation of the symbol information observable under minus the expectation of the clipped logarithmic density correction under .✦ (12)
The positive Lyapunov exponent of is the weighted average of the negative base-two logarithms of the absolute slopes, as in (13).✦ The coding dimension exponent of is the ratio in (13).✦ (13)
The Claim
Theorem 1. Let be a finite weighted affine system and let be a self-similar probability measure for . For every real , there is a real constant such that (14)✦ Suppose real numbers and satisfy for every . There is a real such that (15)✦
In the paper, the result reads: a self-similar probability measure for a finite affine system with , exact dimensional of dimension , satisfies , without a separation or orientation assumption.
What Is Checked
The logarithmic growth bound (14) is stated for a self-similar probability measure of a finite weighted affine system, and the deficit assertion (15) assumes a logarithmic entropy bound with a coefficient below one.