Salem sets in vector spaces over finite fields (Q1749364)

For a finite field \(\mathbb{F}_p\), let \(\mathbb{F}_p^d\) be the \(d\)-dimensional vector space over this field. A subset \(E \subset \mathbb{F}_p^d\) is a Salem set if for all non-zero \(\xi \in \mathbb{F}_p^d\), \[ | \hat{E}(\xi) | = O(\sqrt{\# E}), \] in which \(\# E\) denotes the number of elements of \(E\) and \(\hat{E}\) denotes the discrete Fourier transform of the characteristic function of \(E\). The author proves that almost all random subsets of a finite vector space are weak Salem sets. Hayes derived such a result for random sets of fixed size. The main result of the paper under review (Theorem 1.3) extends this result to a different probability model; i.e. random sets \(E\) are obtained from picking each point of \(\mathbb{F}_p^d\) independently with probability \(\delta\) and removing it with probability \(1-\delta\).