An Inverse Theorem for Certain Directional Gowers Uniformity Norms
Abstract
Let G be a finite-dimensional vector space over a prime field Fp with some subspaces H1, …, Hk. Let f G C be a function. Generalizing the notion of Gowers uniformity norms, Austin introduced directional Gowers uniformity norms of f over (H1, …, Hk) as \[\|f\|U(H1, …, Hk)2k = Ex ∈ G,h1 ∈ H1, …, hk ∈ Hk ∂h1 … ∂hk f(x)\] where ∂u f(x) = f(x + u) f(x) is the discrete multiplicative derivative. Suppose that G is a direct sum of subspaces G = U1 U2 … Uk. In this paper we prove the inverse theorem for the norm \[\|·\|U(U1, …, Uk, [b] G, …, G ),\] which is the simplest interesting unknown case of the inverse problem for the directional Gowers uniformity norms. Namely, writing \|·\|U for the norm above, we show that if f G C is a function bounded by 1 in magnitude and obeying \|f\|U ≥ c, provided < p, one can find a polynomial α G Fp of degree at most k + - 1 and functions gi j ∈ [k] \i\ Gj \z ∈ C |z| ≤ 1\ for i ∈ [k] such that \[|Ex ∈ G f(x) ωα(x) Πi ∈ [k] gi(x1, …, xi-1, xi+1, …, xk)| ≥ ((Op,k,(1))(Op,k,(c-1)))-1.\] The proof relies on an approximation theorem for the cuboid-counting function that is proved using the inverse theorem for Freiman multi-homomorphisms.
Turn this paper into a lesson
ArcXiv compiles a structured reading guide from this paper's metadata: plain-English importance, contributions, prerequisite concepts, which sections to read first, flashcards, and a quiz. Grounded in the abstract, never invented.