Multiple correlation sequences and nilsequences (Q894234)

The author studies the structure of multiple correlations sequences which are sequences of the form \(\int T_1^{n_1}f_1\dots T_l^{n_l}f_l \, d\mu\) where \((X,\chi,\mu, T_1,\dots T_l)\) is a system, \(f_1,\dots, f_l \in L^{\infty} (\mu)\), \(n_1,\dots, n_l \in \mathbb{Z}\). The study of the limiting behavior of averages of such sequences, where the iterates are restricted to certain subsets of \(\mathbb{Z}^l\), has been an indispensable tool in ergodic Ramsey theory. When the iterates of the transformations are integer polynomials it is proved that any such correlation sequence is the sum of nilsequences and an error term that is small in uniform density. The author uses this decomposition result to give convergence for multiple averages and deduce some results to infer convergence for actions of commuting transformations for the special case of actions of a single transformation. The proof of the decomposition result differs from previous works. It consists of a simple orthogonality argument and the main tool is an inverse theorem of \textit{B. Host} and \textit{B. Kra} [J. Anal. Math. 108, 219--276 (2009; Zbl 1183.37011)] for general bounded sequences.