Non-planarity and metric Diophantine approximation for systems of linear forms (Q259484)

Let \(M_{m,n}\) denote the set of \(m \times n\) matrices over \({\mathbb R}\). A matrix \(Y \in M_{m,n}\) is said to be \textit{very well approximable (VWA)} if for some \(\varepsilon > 0\), \[ \| Y q - p \|^m < \| q \|^{-(1+\varepsilon)n} \] for infinitely many \(q \in {\mathbb Z}^n\) and \(p \in {\mathbb Z}^m\). It is \textit{multiplicatively very well approximable (MVWA)} if for some \(\varepsilon > 0\), \[ \Pi(Y q - p) < \Pi_+(q)^{-1-\varepsilon} \] for infinitely many \(q \in {\mathbb Z}^n\) and \(p \in {\mathbb Z}^m\), where for \(x = (x_1, \dots, x_k)\), \[ \Pi(x) = \prod_{i=1}^k | x_i | \quad \text{and} \quad \Pi_+(x) = \prod_{i=1}^k \max\{1,| x_i |\}. \] A measure \(\mu\) on \(M_{m,n}\) is said to be \textit{extremal} (resp. \textit{strongly extremal}) if \(\mu\)-almost no points are VWA (resp. MVWA). From the Borel-Cantelli lemma it follows that the Lebesgue measure is strongly extremal. Strong extremality implies extremality. It was shown by the last two authors and Wang [\textit{D. Kleinbock} et al., Int. J. Number Theory 6, No. 5, 1139--1168 (2010; Zbl 1211.11086)] that if \(\mu\) is a sufficiently regular measure supported on an open set \(U \subseteq {\mathbb R}^d\) and \(F : U \rightarrow M_{m,n}\) is a continuous map satisfying a certain decay condition with respect to the measure \(\mu\), as well as a geometric condition called strong non-planarity, then the push-forward of \(\mu\) under the map \(F\) is strongly extremal. The present paper strengthens this result by replacing strong non-planarity with weak non-planarity, a strictly weaker notion. The technical definition is to be found in the paper. As a consequence of the main theorem, it is shown that analytic, weakly non-planar submanifolds of \(M_{m,n}\) are strongly extremal. In addition to the main result, a number of stability properties of the notion of weak non-planarity are studied. Weighted and inhomogeneous variants of extremality are also derived, the latter using a transference principle due to the first author and \textit{S. Velani} [Proc. Lond. Math. Soc. (3) 101, No. 3, 821--851 (2010; Zbl 1223.11091)]. The paper ends with a nice collection of open questions.