Soit $F(X,Y)$ une forme binaire à coefficients entiers, de discriminant non nul, de degré $\geq 3$.
A quelle condition, nécessaire et suffisante, existe-t-il une forme $G (X,Y)$, non $GL(2, Z)$-équivalente à $F(X,Y)$, telle qu’on ait l’égalité des images $F(Z^2) = G(Z^2)$ ?
La condition trouvée repose sur l’existence d’un élément d’ordre $3$, d’un certain type, dans le groupe d’automorphismes de $F$.
Travail en commun avec Peter Koymans.

Let G be a semisimple real Lie group with Lie algebra g. We will show how the universal enveloping algebra U(g) naturally fits into a one-parameter family of algebras U_q(g) with interesting structure. Any of these algebras U_q(g) moreover allows for an associated C*-algebra, whose representation category closely resembles that of G. We mainly explain these ideas and results in the concrete case of SL(2,R). This is based on joint work with Joel Right Dzokou Talla.

Si l’on se fixe une variété abélienne définie sur un corps de nombre $K$, alors sa classe d’isogénie (l’ensemble des variétés abéliennes qui lui sont isogènes sur $K$) est un ensemble fini: c’est l’un des théorèmes fondamentaux de géométrie arithmétique dus à Faltings. Dans le cas particulier des courbes elliptiques définies sur $K = \mathbb{Q}$, on sait exactement à quoi ressemblent ces classes d’isogénies, mais une telle classification est hors de portée en dimensions supérieures. Dans cet exposé, je parlerai d’un algorithme efficace de calcul de classes d’isogénie dans le cas « le plus simple » des surfaces abéliennes sur $\mathbb{Q}$, fondé sur l’utilisation des fonctions thêta de Riemann. Cet algorithme a permis pour la première fois de calculer de nombreux exemples de classes d’isogénies. Il s’agit d’un travail en commun avec Raymond van Bommel, Shiva Chidambaram et Edgar Costa.

I will talk about a refinement of the Connes-Kasparov isomorphism, which is proved by understanding the structure of the Casselman algebra of rapidly decreasing functions on a real reductive group. I show that this Casselman algebra, which encodes nonunitary representation theory, and the reduced group C^*-algebra, which encodes tempered unitary representation theory, are built in very similar ways from similar elementary components. The structure of the Casselman algebra is understood using techniques from Delorme’s proof of the Paley-Wiener theorem for real reductive groups, which describes the Fourier transform of compactly supported smooth functions. Thanks to the similar structures of the two algebras, it becomes straightforward to prove that the two algebras, once cut down to certain finite sets of K-types, have isomorphic K-theory, which is the refinement of Connes-Kasparov. This work is essentially my thesis at Penn State.

In this talk, we will discuss the Gaussian behaviour of small quadratic non-residues for almost all primes in short intervals. We will begin with some background on quadratic non-residues and then briefly outline the proof. The proof uses the method of moments in conjunction with sieve methods and algebraic inputs from counting solutions of polynomial equations. This is joint work with Debmalya Basak and Alexandru Zaharescu.

Let $M$ be a manifold with a geometric structure and sufficiently nice $G$ a symmetry group, often the geometric structure can be transferred to $M/G$. In (multi-)symplectic geometry, reduction procedures permit to transfer the differential form to an even smaller space. However, all approaches working directly on the space have very strong regularity requirements.
We present an approach to reducing the algebra of (multi-)symplectic observables for general (covariant) moment maps, without any regularity assumptions of the level sets (and the symmetries).
Based on joint work with Casey Blacker and Antonio Miti.

We consider the problem of quantizing the positive nilradical of a complex semisimple Lie algebra of finite rank, together with a certain fixed direct sum decomposition. The decompositions we consider are in one-to-one correspondence with total orders on the simple roots, and exhibit the nilradical as a direct sum of graded modules for appropriate Levi factors. We show that this situation can be quantized equivariantly as a finite-dimensional subspace within the positive part of the corresponding quantized enveloping algebra. Furthermore, we show that such subspaces give rise to left coideals, with the possible exception of components corresponding to some exceptional Lie algebras, and this property singles them out uniquely. Finally, we discuss how to use these quantizations to construct covariant first-order differential calculi on quantum flag manifolds, which coincide with those introduced by Heckenberger-Kolb in the irreducible case.

À chaque polynôme unitaire sans facteur carré $D$ d’un anneau de polynômes $\mathbb{F}_q[t]$ nous pouvons associer un caractère réel quadratique $\chi_D$ et puis une fonction L de Dirichlet $L(s,\chi_D)$. Inspirés par l’article de Y. Lamzouri sur les constantes d’Euler-Kronecker d’extensions quadratiques de corps de nombres, nous étudions la famille des valeurs $-L'(1,\chi_D)/L(1,\chi_D)$ lorsque $D$ parcourt l’ensemble des polynômes unitaires sans facteur carré de $\mathbb{F}_q[t]$. Tout d’abord, nous calculons uniformément leurs moments entiers sur un intervalle particulier. Puis, en utilisant un modèle aléatoire, nous montrons que les valeurs $-L'(1,\chi_D)/L(1,\chi_D)$ possèdent une distribution limite lorsque le degré de $D$ tend vers l’infini, où la fonction de distribution admet une fonction de densité lisse. Nous prouvons également un théorème de discrépance pour la convergence des fréquences des valeurs $-L'(1,\chi_D)/L(1,\chi_D)$ vers cette fonction de distribution. Notre théorème de discrépance fournit de l’information non négligeable à propos des petites valeurs de $-L'(1,\chi_D)/L(1,\chi_D)$. Nous déduisons aussi des résultats analogues pour les constantes d’Euler-Kronecker d’extensions quadratiques. Il s’agit d’un travail en collaboration avec Amir Akbary (University of Lethbridge).

(Joint work with G. Liu and S. Mehdi)

For the last decades, representation theory of Lie groups and algebras has been a very active research topic with a multitude of ramifications and applications. Since the work, in the 1970’s, of Parthasarathy and Atiyah-Schmid, Dirac operators have become efficient tools to describe and classify the unitary dual of a real Lie a group $G$. On the one hand, any irreducible unitary representation occurring in the regular representation $L^2(G)$ can be realized as the Hilbert space of $L^2$-sections, of some twist of the spin bundle over the Riemannian symmetric space $G/K$, which belong to the kernel of the associated Dirac operator. Here $K$ is a maximal compact subgroup of $G$. On the other hand, Dirac cohomology, introduced by Vogan in the late 1990’s, defines an invariant which can be used to detect the infinitesimal character of representations (theorem of Huang and Pandzic). Therefore it is important to study the behavior of the Dirac cohomology under functors involved in representation theory.

A useful functor in representation theory of reductive groups is the so-called $\Theta$-correspondence (or the Howe duality). Howe duality relates representations and characters of two Lie groups $G_1$ and $G_2$, viewed as closed subgroups of the metaplectic group $M$ such that $Z_M(G_1) = G_2$ and $Z_M(G_2) = G_1$.

In this talk, we will study the behavior of the Dirac cohomology under the $\Theta$-correspondence in the case of complex
pairs $(G_1, G_2)$ viewed as real Lie groups.