On the Bogolyubov-Ruzsa lemma (Q1951735)

Let \(G\) be an abelian group. A nonempty subset of \(G\) is called coset progression of dimension \(d\) if it can be written in the form \(P + H\), where \(H\) is a subgroup of \(G\) and \(P\) is a proper progression of dimension \(d\). In this paper the author proved that if \(A\) and \(S\) two nonempty subsets of \(G\), with \(|A + S| \leq K \min \{|A|, |S|\}\), then \((A - A) + (S - S)\) contains a proper symmetric coset progression \(M\) of dimension \(O(\log^{6}2K)\) with \(|M| = e^{-h(K)}|A + S|\), where \(h(K) = O(\log^{6}2K \cdot \log 2 \cdot \log 2K)\). The proof is based on the Fourier analysis on groups. The main tool is the method of \textit{E. Croot} and \textit{O. Sisask} [ Geom. Funct. Anal. 20, No. 6, 1367--1396 (2010; Zbl 1234.11013)] developed recently. Roughly speaking, this is a probabilistic method to find convolutions of subsets of groups which are almost periodic in some sense. The last section of the paper contains some interesting corollaries of the main result, such as the Freiman's theorem for abelian groups and an improvement of a result about the long arithmetic progressions in sumsets.