Arithmetic progressions in certain subsets of finite fields (Q6136712)

Szemerédi-type theorems have been studied over finite fields also. One studies the question of how many arithmetic progressions of a particular length exist in subsets like squares or cubes. Usually, the study involves estimation of character sums, using techniques involved in the Riemann hypothesis over finite fields. In the paper under review, the authors study two problems. They also involve character sums but these are singular. They need to use recent estimates proved in this set-up by \textit{A. Rojas-León} [Int. Math. Res. Not. 2005, No. 20, 1221--1234 (2005; Zbl 1078.11045)] which extended the work of N. Katz. The authors consider the set \(S_p\) of squares, and the set \(C_p\) of cubes in \(\mathbb{F}_p\), for an odd prime \(p\). In the case of squares, they prove that the number of \(k\)-term arithmetic progressions in \(S_p\) is \(p^2/2^k + R\), where the error \(R\) satisfies \(|R| \leq \bigg( \frac{k-2}{4} - \frac{k-2}{2^{k-1}}\bigg) p^{3/2} + c_k p\), where \(c_k\) is an effectively computable constant, depending only on \(k\). They also obtain some exact formulae in the case of arithmetic progressions of small length \(\leq 5\), which involve studying certain elliptic curves, and they show that the error is optimal as predicted by Sato-Tate conjecture (now a theorem). Similar to the use of Gauss sums in estimating the number of 3-term arithmetic progressions in \(S_p\), the authors use Kummer sums to estimate the analogous number in the cube set \(C_p\). The basic techniques involved in the proofs are finite Fourier analysis, arithmetic of varieties over finite fields and estimation of exponential sums.