Schedule, Notes/Handouts & Videos

We will discuss a quantitative equidistribution statement for adelic homogeneous subsets whose stabilizer is maximal and semisimple.  An application to certain equidistribution theorems will also be given. This is a joint work with Einsiedler, Margulis and Venkatesh. 

We study eigenfunctions of the discrete laplacian on large regular graphs, and prove a ``quantum ergodicity'' result for these eigenfunctions : for most eigenfunctions $\psi$, the probability measure $|\psi(x)|^2$, defined on the set of vertices, is close to the uniform measure.

Although our proof is specific to regular graphs, we'll discuss possibilities of adaptation to more general models, like the Anderson model on regular graphs.

The topic of my talk is what the possibilities are for the closure of a totally geodesic immersion of a hyperbolic plane into a complete hyperbolic 3 manifold.  For finite volume hyperbolic 3 manifolds, work of Shah and Ratner shows that strong rigidity properties hold: any such immersed plane is either closed or dense.  I will describe recent joint work with Curt McMullen and Amir Mohammadi  which shows some of this rigidity persists in some cases of infinite volume hyperbolic 3-manifolds

Singular systems of linear forms were introduced by Khintchine in the 1920s, and it was shown by Dani in the 1980s that they are in one-to-one correspondence with certain divergent orbits of one-parameter diagonal groups on the space of lattices. We give a (conjecturally sharp) upper bound on the Hausdorff dimension of singular systems of linear forms (equivalently the set of points with divergent trajectories). This extends work by Cheung, as well as by Chevallier and Cheung, on the vector case. For a diagonal action on the space of lattices, we also consider the relation of the entropy of an invariant measure to its mass in a fixed compact set. Our technique is based on the method of integral inequalities developed by Eskin, Margulis, and Mozes. This is a joint work with D. Kleinbock, E. Lindenstrauss, and G. A. Margulis

(Joint with R. Boutonnet, A. Ioana) The notion of local spectral gap for general measure preserving actions will be defined. We prove that the left translation action of a dense subgroup of a simple Lie group has local spectral gap if the subgroup has algebraic entries. This extends to the non-compact setting recent works of Bourgain-Gamburd and Benoist-de Saxce. We also extend Bourgain-Yehudayoff’s result. The application of this result to Banach-Ruziewicz problem, delayed random-walk, and monotone expanders will be explained. 

A natural generalization of classical diophantine approximation consists in taking products of elements in a Lie group and ask how fast they can approach the identity. For example, classically, a number x is Diophantine if and only if sums of N terms each equal to 1, -1, x or -x, cannot get closer to zero than an inverse power of N. Replacing x and 1 by two or more vectors in a vector space is the subject of simultaneous approximation, and diophantine approximation in matrices. In this talk I will consider the case when the elements are chosen from a more general Lie group and focus on nilpotent groups. It turns out that the computation of optimal exponents for nilpotent Lie groups can be reduced to that of certain algebraic subvarieties of matrices. We give a formula for the exponent of arbitrary submanifolds of matrices, showing it depends only on their algebraic closure, answering a question of Kleinbock and Margulis, and use it to express the exponent of nilpotent Lie groups in terms of various representation theoretic data. This is joint work with M. Aka, L. Rosenzweig and N. de Saxce

Let A denote the diagonal group in SL(n,R) acting on the space of unimodular lattices in R^n. In this talk I will explain a construction of a sequence of A-invariant ergodic probability measures (supported on compact A-orbits) which converge to the zero measure. In fact the "geometry" of these orbits will be described in some detail. 

The Oppenheim conjecture proved in mideighties states that the set of values at integral points of an irrational indefinite quadratic form in n>2 variables is dense in R. I will talk about the history and proof of this conjecture and will also talk about various problems related to the conjecure

I will present recent results on the asymptotic behaviour of random walks on semisimple groups and ask open questions. This is joint work with Yves Benoist

We will discuss a recent work with Aaron Brown and Federico Rodriguez Hertz on smooth classification of Anosov actions by higher rank lattices on nilmanifolds. In particular, we will explain how the existence of an Anosov diffeomorphism from the group action leads to Anosov property of generic elements in the acting group, allowing to make use of large abelian subgroups. 

The moduli space of holomorphic one-forms on a closed surface of genus $g$ has a natural piecewise linear structure and carries an action of GL(2,R). One can investigate the geometry and dynamics of an individual flat surface by studying its orbit under this linear group action. I will discuss several open problems regarding the properties of these orbit closures. This talk is based on joint work with Alex Eskin, Amir Mohammadi and Alex Wright.

We study the dynamics of a point particle in an array of spherical scatterers centered at the points of a quasicrystal (of cut-and-project type). It turns out that, as in the case of a periodic scatterer configuration, the stochastic process that governs the time evolution for random initial data in the limit of low scatterer density (Boltzmann-Grad limit) is Markovian after an appropriate extension of the phase space. The definition and properties of the transition kernel for this limiting Markov process are more complicated in the quasicrystal case, and we will discuss this both in general and in specific examples. Homogeneous dynamics enters as a key tool in the proofs. This is joint work with Jens Marklof

Let G be a semisimple Lie group, and \Gamma a lattice. in G. Let \mu be a measure on G. We show that under a certain condition on \mu, any \mu-stationary measure on G/\Gamma is in fact invariant under the group generated by the support of \mu. We give an alternative argument (which bypasses the Local Limit Theorem) for some of the breakthrough results of Benoist and Quint in this area. This is joint work with Elon Lindenstrauss