The paper investigates a special class of complex geometric spaces called hyperfujiki manifolds, which are the natural generalizations of hyperkähler manifolds when you allow the geometry to be somewhat less rigid. Hyperkähler manifolds are among the most beautiful objects in modern geometry, partly because they carry a special algebraic tool called the Beauville-Bogomolov-Fujiki (BBF) form, a kind of generalized inner product defined on cohomology classes. The central question the paper addresses is: when does the positivity of this BBF form force the underlying space to be "Moishezon," meaning it has enough algebraic structure to be approximated by projective varieties?
The authors prove two main results connecting BBF positivity to Moishezonness. First, they show that any Moishezon hyperfujiki manifold automatically carries a geometric object called a big line bundle whose associated cohomology class scores positively under the BBF form, establishing one direction of the connection. Second, and more substantially, they prove the converse under certain conditions: if a cohomology class of the right type scores positively under the BBF form, then the space must be Moishezon. This converse is established in dimension four outright, and in higher even dimensions under a mild additional assumption borrowed from the minimal model program. The proofs draw on several deep recent theories, including work on singular symplectic varieties, techniques for building simplified geometric models of spaces with trivial canonical bundle, and methods for pulling back differential forms across maps between singular spaces.
As a concrete application, the authors use their criterion to describe, in terms of pure Hodge theory, exactly which members of certain smooth families of these manifolds are Moishezon. Hodge theory is a classical toolkit that encodes geometric information in linear algebraic data attached to a manifold, so this gives a tractable, computable way to detect the algebraic-geometric complexity of these spaces within a family. The overall thrust of the paper is to extend the well-developed dictionary between positivity and algebraic geometry, familiar from the projective setting, into the broader and less understood world of compact complex manifolds that are not necessarily projective.