Mathematics Home / Research Seminars: Geometry & Topology
Time & Location: All talks are on Thursday in Gibson Hall 308 at 12:30 PM unless otherwise noted.
Organizer: Komendarczyk, Rafal
Title: Vietoris-Rips thickenings: Problems for birds and frogs
Henry Adams | Colorado State
Abstract: An artificial distinction is to describe some mathematicians as birds, who from their high vantage point connect disparate areas of mathematics through broad theories, and other mathematicians as frogs, who dig deep into particular problems to solve them one at a time. Neither type of mathematician thrives without the help of the other! In this talk, I will survey open problems related to Vietoris-Rips complexes that are attractive to both birds and frogs. Though Vietoris-Rips complexes are frequently used to approximate the shape of a dataset, many questions remain about their mathematical properties. Frogs may delight in open problems such as the homotopy types of Vietoris-Rips complexes of spheres, ellipsoids, tori, graphs, Cayley graphs of groups, geodesic spaces, subsets of the plane, and even the integer lattice Z^n with the taxicab metric for n >= 4. Birds may enjoy emerging connections between Vietoris-Rips complexes and a variety of areas in pure mathematics, including metric geometry (Gromov-Hausdorff distances), quantitative topology (Gromov's filling radius), measure theory (optimal transport), topological combinatorics (Borsuk-Ulam theorems), geometric group theory (finiteness properties of groups), and geometric topology (thick-thin decompositions).
Time: 10:00am CT
Title: AATRN: Vietoris-Rips Persistent Homology, Injective Metric Spaces, and The Filling Radius
Facundo Mémoli | Ohio State
Abstract: The persistent homology induced by the Vietoris-Rips simplicial filtration is a standard method for capturing topological information from metric spaces. We consider a different, more geometric way of generating persistent homology of metric spaces which arises by first embedding a given metric space X into a larger ambient metric space E and then considering neighborhoods of the original space X inside E.
We then prove that the persistent homology of the Vietoris-Rips filtration is isomorphic to our geometric persistent homology provided that the ambient metric space E satisfies a property called injectivity.
As an application of this isomorphism result we are able to precisely characterize the type of intervals that appear in the persistence barcodes of the Vietoris-Rips filtration of any compact metric space and also to give succinct proofs of the characterization of the persistent homology of products and metric gluings of metric spaces.
Our results also permit proving several bounds on the length of intervals in the Vietoris-Rips barcode by other metric invariants.
As another application, we connect this geometric persistent homology to the notion of filling radius of manifolds introduced by Gromov and show some consequences related to (1) the homotopy type of the Vietoris-Rips complexes of spheres which follow from work of M. Katz and (2) characterization (rigidity) results for spheres in terms of their Vietoris-Rips persistence barcodes which follow from work of F. Wilhelm.
Time: 10:00am CT
Title: AATRN: Graph representation learning and its applications to biomedicine
Marinka Zitnik - Harvard University
Abstract: The success of machine learning depends heavily on the choice of representations used for prediction tasks. Graph representation learning has emerged as a predominant choice for learning representations of networks. In this talk, I describe our efforts to expand the scope and ease the applicability of graph representation learning. First, I outline SubGNN, a subgraph neural network for learning disentangled subgraph embeddings. SubGNN generates embeddings that capture complex subgraph topology, including structure, neighborhood, and position of subgraphs in a graph. Second, I will discuss applications in biology and medicine. The new methods predicted disease treatments that were experimentally verified in the wet laboratory. Further, the methods helped to discover dozens of combinations of drugs safe for patients with considerably fewer unwanted side effects than today's treatments. Lastly, I describe our efforts in learning actionable representations that allow users of our models to receive predictions that can be interpreted meaningfully.
Recordings of most of the talks will be posted to the AATRN YouTube Channel
Time: 10:00am CT
Title: AATRN: How optimal transport can help us to determine the curvature of complex networks?
Marzieh Eidi | Max Planck
Abstract: Ollivier Ricci curvature is a notion that originated from Riemannian Geometry and is suitable for applying on different settings from smooth manifolds to discrete structures such as (directed) hypergraphs. In the past few years, alongside Forman Ricci curvature, this curvature as an edge-based measure has become a popular and powerful tool for network analysis. This notion is defined based on optimal transport problem (Wasserstein distance) between sets of probability measures supported on data points and can nicely detect some important features such as clustering and sparsity in their structures. After introducing this notion for (directed) hypergraphs and mentioning some of its properties, as one of the main recent applications, I will present the result of the implementation of this tool for the analysis of chemical reaction networks.
Recordings of most of the talks will be posted to the AATRN YouTube Channel.
Time: 10:00am CT
Title: AATRN: High-dimensional data, level-set geometry, and Voronoi analysis of spatial point sets
Menachem Lazar | Bar-Ilan University
Abstract: Physical systems are regularly studied as spatial point sets, and so understanding the structure of such sets is a very natural problem. However, aside from special cases, describing the manner in which a set of points is arranged in space can be quite challenging. In the first part of this talk, I will show how consideration of the configuration space of local arrangements of neighbors can shed light on essential challenges of this problem, and in the classification of high-dimensional data more generally. In the second part of the talk I will introduce some ideas from Voronoi cell topology and show how they can be used to define crystals, defects, and order more generally in a somewhat precise manner.
Recordings of most of the talks will be posted to the AATRN YouTube Channel.
Time: 10:00am CT