The Setting
Theorem 1.1 of the preprint, the log-Brunn-Minkowski inequality, states that in every dimension , for origin-symmetric convex bodies and in and every , the volume of the log combination , the Wulff body of the function with the support function of , is at least , all volumes being Lebesgue volumes taken in . Section 2 of the preprint, from a variance bound to volume, works with finitely many directions in place of the whole unit sphere.
How are a Wulff body and its volume recovered from the bodies cut out by finitely many symmetric slabs, as the directions fill the sphere? Which inequality for such bodies, or which variance bound, gives Theorem 1.1, and what does Theorem 1.1 say on the line? Every finite set of directions considered contains the coordinate vectors of .
Slab Bodies and Wulff Bodies
The Euclidean space is the space of real -tuples with the inner product and the norm .✦ Its unit sphere is , and are the coordinate vectors.✦ Volumes are Lebesgue volumes.
A convex body in is a compact convex subset of with nonempty interior.✦ A set is origin-symmetric if exactly when .✦ The support function of a set is the function on given by (1)✦
The Wulff body of a real function on is the set (2)✦ The interpolated support function of sets with a real parameter is the function on sending to .✦ The log combination of and with parameter is the Wulff body of the interpolated support function, (3)✦
The slab body of a finite set and a real function on is the intersection of one symmetric slab for each vector of , (4)✦ For a sequence in and an integer , the set of test directions is the finite subset of made of the coordinate vectors and the first terms of the sequence, (5)✦
The Soft Variance Inequality
The feature polynomial of a vector , a width and an exponent is the polynomial on .✦ Vectors and widths give the feature family , with the feature polynomial of , and a common exponent .✦
With a regularization , the soft slab law of the feature family is the probability measure on , where the regularized potential is (6)✦ is the Gibbs density✦ and is the partition function.✦
The mean and the variance of a real function on under are (7)
Let and , let be even and , let be nonzero vectors and positive widths, and let be the soft slab law of the corresponding feature family. The soft variance inequality, Proposition 2.1 of the preprint, says that every choice of real coefficients satisfies (8)✦
The Claim
Theorem 1. (a) Let be a dense sequence in with test directions , and let be an even real function on whose restriction to is continuous. Then (9)✦ and (10)✦
(b) Let , let and be origin-symmetric convex bodies in , and let . Suppose that every finite set containing satisfies (11) Then (12)✦
(c) Let and be origin-symmetric convex bodies in , and let . Then (13)✦
(d) Suppose the soft variance inequality (8) holds. Let , let and be origin-symmetric convex bodies in , and let . Then (14)✦
What Is Checked
The claims (9) to (14) are stated over the Euclidean space and its unit sphere, the test directions and the slab bodies , Wulff bodies, support functions and log combinations of origin-symmetric convex bodies, and the soft variance inequality (8) for the soft slab law of a feature family.