Maximal components of Noether-Lefschetz locus for Beilinson-Hodge cycles (Q2480877): Difference between revisions

Let \(U\) be a smooth complex quasiprojective variety, and \(\text{CH}^m(U,m)\) an instance of S. Bloch's higher Chow groups, that most resembles Milnor \(K\)-theory. Let MHS stand for the category of mixed Hodge structures. There is a cycle class map \(\text{cl} : \text{CH}^m(U,m)\otimes {\mathbb{Q}} \to\Gamma\big(H^{m}(U,{\mathbb{Q}}(m))\big) := \hom_{\text{MHS}}\big({\mathbb{Q}}(0),H^{m}(U,{\mathbb{Q}}(m))\big)\), the target space \(\Gamma\big(H^{m}(U,{\mathbb{Q}}(m))\big)\) referred to as the space of Beilinson-Hodge cycles. This cycle class map \(\text{cl}\) has been studied rather extensively by the authors, with the anticipation that it may be surjective. So far, the evidence seems to support this expectation. Now suppose that we have a family \(\{U_t\}_{t\in M}\) of smooth quasiprojective varieties with ``nontrivial'' MHS. The Noether-Lefschetz locus \(M_{\text{NL}}\subset M\) is the subset of those \(t\in M\) for which \(\dim \Gamma\big(H^{m}(U_t,{\mathbb{Q}}(m))\big)\) is bigger than that for sufficiently general \(t\). It is well known that \(M_{\text{NL}}\) is a countable union of analytic varieties. The authors arrive at analogous results of the works of M. Green and others on the dimension of the components of \(M_{\text{NL}}\), in the case where \(\{U_t\}\) is a certain family of open surfaces in \(\mathbb{P}^3\). Their method can be described as a tour de force construction, using infinitesimal Hodge theory techniques, together with a reduction to Jacobian rings.