Maximum likelihood degree of surjective rational maps (Q2101709)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Maximum likelihood degree of surjective rational maps
scientific article

    Statements

    Maximum likelihood degree of surjective rational maps (English)
    0 references
    0 references
    6 December 2022
    0 references
    Let \(X=\{D_0,\ldots, D_m\}\) be a collection of effective divisors in \(\mathbb {CP}^n\). The maximum likelihood (ML) degree of \(X\) is the signed Euler characteristic of \(\mathbb {CP}^n \setminus D\), where \(D=\bigcup_{i=0}^{m}D_i\) (see [\textit{F. Catanese} et al., Am. J. Math. 128, No. 3, 671--697 (2006; Zbl 1123.13019)]). This short article computes explicitly the ML degree under the following condition: Each \(D_i\) is defined by a section \(f_i\), and \(\{f_i\}_{i=0}^{m}\) spans a linear system defining a surjective rational map \(f\colon \mathbb {CP}^n\dashrightarrow \mathbb {CP}^n\). In particular, all \(f_i\) have the same degree and \(n=m\). The ML degree will be the coefficient of a certain rational polynomial defined by \(X\). The main idea is that \(f\) can be used to define a vector bundle \(E_f\). This \textit{intermediate} vector bundle is used to construct a short exact sequence having \(F=\Omega_{\mathbb {CP}^n}^1(\log D)\) at its center. Now we use that the ML degree is the degree of the top Chern class of \(F\).
    0 references
    surjective rational map
    0 references
    vector bundle
    0 references
    Chern number
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references