On Beloshapka's rigidity conjecture for real submanifolds in complex space (Q2080643)

Let \(M\in\mathbb{C}^{n+d}\) be a smooth generic submanifold of codimension \(d\) at a given point \(p\) of \textit{finite type} \(m=(m_1, \dots,m_k),\) in the sense of Kohn and Bloom-Graham, where \(2 \le m_1< \dots<m_k\) are the Hörmander numbers. Using a result of Bloom-Graham, we may assume that \(M\) is given in suitable holomorphic weighted coordinates \((z,w)\in\mathbb{C}^{n+d}\) by \[ M= \{ \Im w= P(z, \bar z, \Re w)+ \dots \}, \] where \(P=(P_1,\dots, P_d),\) and \(P_j\) is a (non-zero) weighted homogeneous polynomial of degree \(m_j\) with no pluriharmonic terms, and the dots corresponding to \(P_j\) are weighted terms of order \(> {m_j}.\) One associates to \(M\) the polynomial model \(M_0\) given by \[ M_0= \{ \Im w= P(z, \bar z , \Re w)\}. \] When the only Hörmander number is \(2\), then it is known that nonlinear automorphisms do exist for nondegenerate models. In the present paper, the author studies the case of general Hörmander numbers and answers the following Beloshapka's Rigidity Conjecture in its full generality in the affirmative: \begin{itemize} \item[] All totally nondegenerate models of length \(l>2\) (where \(l\) is the (maximal) weighted degree of \(P_d\)) are rigid, that is, admit no nonlinear origin preserving automorphisms. \end{itemize} See Definition 1 and Definition 2 pages 429 and 430 for the notion of totally nondegenerate models. We note that the case \(l=3\) was solved by Gammel and Kossovskiy, while the case \(n=1\) holds as observed by Kruglikov.