## Algebra, Combinatorics, Measure

Seminar is on **Mondays**. This year we will have a mix of a reading seminar and a limited number of external speakers.

Topics covered might include, but are by no means limited to, group rings of infinite groups, ${l}^{2}$-invariants, appoximations of infinite groups (for example sofic groups), asymptotic and measured group theory, Borel and measurable combinatorics, graph limits, interactions between algebra/combinatorics and operator algebras, probability, etc.

We have a mailing list: . In order to participate either to sign you up, or send a *plain text* email to , containing only the two words `subscribe acmseminar`

in the body (you can leave the subject empty).

### 2016-17

### Abstracts

This talk is devoted to the problem of characterizing the family of factorial classes of graphs, i.e., hereditary classes in which the number of n-vertex graphs grows as a factorial type function. First, we will discuss background and motivation of the problem. Then we will consider a special case of characterizing the family of factorial classes of bipartite graphs. I will present a recent result showing that the class of P_7-free bipartite graphs is factorial. The result completes the description of monogenic factorial classes of bipartite graphs. The proof is based on a new decomposition scheme of bipartite graphs, which is of independent interest. Finally, we will discuss some open problems.

Based on joint work with Vadim Lozin.

The curve enclosing the greatest area for its length is a circle; this is the first example of an isoperimetric theorem. We can look for similar results whenever we have a notion of size and boundary: for example, in graphs. The general problem is NP-hard, but we can sometimes obtain good results in graphs with a lot of structure. I'll describe a very general result that solves the isoperimetric problem for all lattice like graphs. This is joint work with Joshua Erde.

I will describe my joint work with Sageev on properties of groups acting on CAT(0) cube complexes, like property P_naive and uniform exponential growth. I will not assume any specialist knowledge and explain how one exploits the geometry of CAT(0) cube complexes to extract group theoretic properties.

I will talk about L1 full groups of ergodic measure-preserving transformation, which are measurable analogues of topological full groups of minimal homeomorphisms of the Cantor space. As in the topological world, these groups completely remember the transformation up to flip conjugacy. I will describe basic properties of these Polish groups an then present some results on their topological rank.