A geometric realization of an integrable system is an evolution of curves on a manifold M, invariant under the action of a group G, and such that it becomes the integrable system when the action of G is mod out. The best known geometric realization is the Vortex filament flow (VF), a flow of curves in Euclidean space which is invariant under the Euclidean group. The VF equation becomes the nonlinear Shrodinger equation when written in terms of the natural curvatures of the flow - Hasimoto proved this way the integrability of VF -. In this talk I will review what is known about the classical geometry of curves in homogeneous spaces and its relation to different types of integrable systems. In particular we will talk about how one can reduce different Hamiltonian structures to the space of curvatures (Euclidean, projective, conformal, etc) and how those reductions indicates the existence of biHamiltonian (integrable) systems. We will also describe how curvatures of Schwarzian type for curves in homogeneous spaces usually describe evolutions of KdV type. I will present background material so the talk should be accessible, at least in part, to different audiences.

I will discuss the metric behavior of the Kahler-Ricci flow on Hirzebruch surfaces assuming that the initial metric is invariant under a maximal compact subgroup of the automorphism group. I will describe how, in the sense of Gromov-Hausdorff, the flow either shrinks to a point, collapses to P^1 or contracts an exceptional divisor. This confirms a conjecture of Feldman-Ilmanen-Knopf. This is a joint work with Jian Song.

We will discuss the existence and uniqueness of the foliations by stable spheres with constant mean curvature for asymptotically flat manifolds satisfying the parity condition at infinity. The concept of center of mass in general relativity will also be discussed. This work generalizes the earlier results of Huisken-Yau, R. Ye, and Metzger.

We discuss 3-Sasakian 7-manifolds carrying metrics of positive curvature and an example of this kind on a new diffeomorphism type.