Essential dimension of cubics (Q1879663)

The concept of essential dimension tries to make precise the notion of how many algebraically independent parameters are needed to determine a given structure. This idea was introduced by \textit{J.~Buhler} and \textit{Z.~Reichstein} [Compos. Math. 106, No.~2, 159--179 (1997; Zbl 0905.12003)]. There, the essential dimension of a finite group \(G\) over a field \(k\) was defined as follows. Consider algebraic \(k\)-varieties \(Y\) with a faithful \(G\)-action for which there exists a dominant \(G\)-equivariant rational map \(V\rightarrow Y\) defined over \(k\), where \(G\rightarrow \text{GL}(V)\) is a faithful finite-dimensional representation. Then the essential dimension of \(G\) is the minimum possible dimension of such an algebraic variety \(Y\). In an earlier paper [Doc. Math., J. DMV 8, 279--330 (2003; Zbl 1101.14324)], the authors defined the essential dimension of certain functors, based on unpublished work of A.~Merkurjev. In the paper under review, the authors show that over any field of characteristic different from 2 and 3, the functor \(\text{Cub}_3\) of homogeneous cubic polynomials in three variables (up to linear changes in coordinates) has essential dimension 3. The proof basically shows that defining a nonsingular cubic up to projective equivalence amounts to specifying a configuration of nine flex points in the plane, which requires two parameters, and a value of the \(j\)-invariant. The authors use Galois cohomology functors of some suitable algebraic groups to carry out the proof. In the course of the proof, the authors show that the functor \(\text{Cub}^-_3\) of singular cubics in three variables has essential dimension 2. They conjecture that the essential dimension of singular hypersurfaces should be less than the essential dimension of nonsingular hypersurfaces for higher degrees as well.