A generalization of Komlós's theorem on random matrices (Q2746302)

Let \(\Delta\) be an arbitrary finite subset of \(\mathbb{Z}\), \(\Delta^n\) the set of all vectors \((x_1,x_2, \dots, x_n)^t\), \(x_i\in\Delta\), \(\Delta_{m \times n}\) the set of all \(m\times n\) matrices with entries from \(\Delta\) and \(q=|\Delta|\) the cardinality of \(\Delta\), assumed \(2\leq q<\infty\). The author proves a generalization of Komlós's theorem on random matrices: Let \(A\) be a random matrix from \(\Delta_{n\times n}\). Then as \(n\to \infty\) \(\text{Prob}(r(A) <n)=0(1/n)\). To prove this theorem he uses some rather involved techniques for random matrices.