The Setting
A Boolean function on a finite biased product has a Fourier spectrum indexed by coordinate sets. The question is how its mass on small nonempty sets depends on the distribution of its pivotal influences.
Let be a finite coordinate set and let . The Bernoulli mass on and its product probability mass on are defined by (1): (1) Write and for probability and expectation with this mass. The real indicator of a Boolean value is for true and for false; a Boolean function is also regarded as its real indicator.✦
For , let denote with coordinate set to . The pivotal influence of coordinate is the probability that the two choices at give different values of , and the total influence is their sum, as in (2): (2)✦
Set . The biased coordinate character and the product character for are the real functions in (3): (3) The biased Fourier coefficient of a real function at is its expectation against , as in (4): (4)✦
The Claim
Theorem 1. Let be Boolean and let be real. Suppose that for every , and put . Then the full-sum estimate (5) holds: (5)✦ The restricted-sum estimate is (6): (6)✦
In the paper, the result reads: for finite , , , real , and a Boolean function with for every , the cutoff gives , where the cutoff applies literally to integer cardinalities, so the sum is empty when .
What Is Checked
The claims (5) and (6) are stated over Boolean functions on finite Bernoulli products, with the pivotal influences of (2) and the Fourier coefficients of (4) taken at the same interior bias .