Séminaire de Géométrie complexe

In this talk I will present a joint work with C. Casagrande, in which we study smooth complex Fano 4-folds with a rational fibration onto a 3-fold. After an introduction on the setting and motivation, I will discuss our main result: if X is Fano 4-fold with a rational fibration onto a 3-fold and it is not a product of surfaces, then the Picard number of X is at most 9, and the bound is sharp. Moreover, I will present a classification result in a special case within the setting above, and show new examples of Fano 4-folds with large Picard number.

A fundamental result in Complex Geometry is the Kobayashi-Hitchin correspondence, stating that a holomorphic vector bundle on a Kähler manifold is poly-stable (as defined by Mumford, Takemoto) if and only if it admits a Hermitian metric solving the Hermite-Einstein equation. It has now become clear that there exist many possible different stability notions for vector bundles, that are of great interest in Algebraic Geometry and String Theory. It is natural to wonder if these stabilities are also tied to the existence of Hermitian metrics with special curvature properties. In this talk, based on joint work with Julien Keller (arXiv:2405.03312[math.DG]), we will consider a class of « polynomial » equations for the curvature of rank 2 holomorphic vector bundles on compact projective surfaces, and a corresponding class of polynomial stability conditions for these bundles. We will then explain how each of these stability conditions is related to the existence of a Hermitian metric satisfying the corresponding equation. This refines and partially confirms a conjectural correspondence between Bridgeland stability conditions and PDEs on holomorphic vector bundles, formulated by Dervan, McCarthy, and Sektnan.

We introduce the canonical model singularities (CMS) criterion for bigness of the cotangent bundle for surfaces. The CMS-criterion for bigness involves invariants for canonical singularities that we describe and we give formulas for A_n singularities. The CMS-criterion leads to conjectures and some answers about the geography and the possible ratios $c_1^2/c_2$ of surfaces with big cotangent bundle. Two cases are naturally separated: regular and irregular surfaces. For regular surfaces we apply the CMS-criterion to show the existence of deformations of hypersurfaces in $\mathbb{P}^3$ with big cotangent bundle for degree $d\ge 8$ and give an example of the regular surface with big  cotangent bundle with ratio close to 1/5. For irregular surfaces we show that there are examples with ratio as close to 1/5 as possible. If time permits, we talk about ratios  below 1/5.

Formalisme thermodynamique à basse température, dynamique symbolique et quasi-cristaux

L’étude de modèles simples de physique statistique sur le réseau $\mathbb{Z}^d$, visant à comprendre la transition du désordre vers un ordre périodique ou quasi-périodique quand la température est suffisamment basse, nécessite une interconnexion entre le formalisme des mesures de Gibbs et des états d’équilibre, la dynamique symbolique multidimensionnelle, les pavages et l’informatique théorique. En particulier, des espaces associés aux marginales finies-dimensionnelles des mesures invariantes par décalage apparaissent et possèdent une étonnante richesse. Cet exposé se propose de présenter un panorama introductif de ce domaine de recherche.

The BNS set of a finitely generated group $\Gamma$ is a certain canonical subset of the space of real additive characters on $\Gamma$. It is a subtle invariant of the group that naturally comes up in different questions of geometric and homological group theory. In the case when $\Gamma$ is the fundamental group of a compact Kähler manifold $X$, Thomas Delzant found a beautiful description of its BNS set in terms of holomorphic fibrations of $X$ over hyperbolic orbifold curves. Using it, he showed that if the fundamental group of a compact Kähler manifold is virtually solvable, it is in fact virtually nilpotent. I will explain the main ideas behind Delzant’s proof and how to generalise his theorems to the case when $X$ is a smooth complex quasi-projective variety. Time permitting, I will also discuss some applications and the case of quasi-Kähler manifolds.

Title: Affine buildings from real algebraic geometry

Abstract: The theory of symmetric spaces and the theory of buildings have a rich history of parallels and interactions. We describe symmetric spaces in terms of real algebraic geometry and then replace the real numbers by valued real closed fields to construct an affine Λ-building. A key tool is a transfer principle from model theory.

Abstract: Galois groups have a long history in enumerative geometry, encoding the intrinsic symmetries of enumerative problems. In this talk, after revisiting the core properties of enumerative Galois groups and their connections with monodromy, we focus on the Fano variety F of lines on a cubic fourfold Y, a hyperkähler fourfold, and investigate the monodromy of the Voisin map, a degree 16 self-rational map of F. We show that its Galois group is « as large as possible », and, in doing so, delve into the geometry of the LLSvS variety—a hyperkähler manifold parameterizing twisted cubics on Y. This is based on joint work with L.Giovenzana.

Quand X est une variété complexe projective lisse, de dimension d, l’i-ème itéré du cup-produit avec une section hyperplane induit un isomorphisme entre les espaces de cohomologie singulière H^(d-i)(X) et H^(d+i)(X). La conjecture standard de type Lefschetz pour X, formulée par Grothendieck dans les années 60 et encore largement ouverte, prédit que les inverses de ces isomorphismes devraient être induits par des cycles algébriques sur X \times X. Dans cet exposé, après une introduction à ces idées, je parlerai de travaux récents avec Ancona, Laterveer et Saccà, dans lesquels nous démontrons la conjecture pour certaines variétés hyperkähleriennes munies d’une fibration lagrangienne. De nouvelles idées nous permettent en fait de traiter certaines fibrations où les fibres ne sont pas toutes irréductibles, ainsi éliminant l’une des hypothèses les plus restrictives faites dans notre premier article.