Birational cobordism invariance of uniruled symplectic manifolds (Q2482717)

Many years ago Y. Ruan pointed out that it should be possible to translate theorems from birational algebraic geometry into the symplectic setting. In order to do this one would have to have a good understanding of the computation of Gromov-Witten invariants. In the interim there have been remarkable advances in computation in Gromov-Witten theory, and it is now possible to address questions from birational geometry in a symplectic setting. This paper is a step in this direction. It has the potential to translate parts of the Mori program into symplectic geometry and to give a better understanding of the role of rational curves in Mori theory. Here the authors prove that being uniruled is invariant under symplectic blow up and blow down along submanifolds. The first obstacle is to formulate a suitable definition of uniruled in the symplectic category. Recall that an algebraic surface is uniruled if it is covered by rational curves. An analogous definition in the symplectic category would admit too many manifolds. Instead, one uses the result that a projective manifold is uniruled if and only if there is a non-zero genus zero Gromov-Witten invariant involving a point constraint [\textit{J. Kollár}, in: European congress of mathematics (ECM), Budapest, Hungary, July 22-26, 1996. Volume I. Basel: Birkhäuser. Prog. Math. 168, 255--288 (1998; Zbl 0970.14001); \textit{Y. Ruan}, Turk. J. Math. 23, No. 1, 161--231 (1999; Zbl 0967.53055)]. This then becomes the definition of uniruled in the symplectic category. The main part of the paper uses relative invariants and the degeneration formula to relate the appropriate Gromov-Witten invariants of a manifold and its blow-ups. The paper begins with a nice resume of symplectic cuts and the inverse operation of normal connected sum. It then reviews relative Gromov-Witten invariants and the degeneration formula. The discussion here is rather nice because it covers the key ideas without getting lost in a maze of technical detail. They then review the Gromov-Witten characterization of uniruled projective manifolds and move to the main technical arguments of the paper.