HOME    »    PROGRAMS/ACTIVITIES    »    
Annual Program Seminars

During each Annual Thematic Program, several seminars are offered. Talks will include graduate-level lectures as well as seminars covering various topics related to the theme.

  • Representing Large and Complex Data

    Gunnar Carlsson, Stanford University
    September 18, 2013 10:30 AM - 12:00 PM
    Lind Hall 305 [Map]

    Abstract forthcoming.

  • The Structure of Persistence Modules

    Vin de Silva, Pomona College
    September 18, 2013 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract forthcoming.

  • Persistent Homology and Applications

    Gunnar Carlsson, Stanford University
    September 24, 2013 3:30 PM - 5:00 PM
    Keller Hall 3-180 [Map]

    Abstract forthcoming.

  • Topological Methods in the Analysis of Biological Data

    F. Javier Arsuaga, San Francisco State University
    September 25, 2013 1:30 PM - 2:30 PM
    Keller Hall 3-180 [Map]

    Abstract

    In this introductory talk I will explain some of the topological tools that we are using to analyze data obtained in molecular biology and genomics experiments. The first part of my talk will be devoted to present well established applications of knot theory to the analysis of 3D DNA structure. In the second half of my talk, I will introduce our homology methods for the analysis of copy number and gene expression data in cancer.
  • Persistent Holes in the Universe

    Pratyush Pranav, Rijksuniversiteit te Groningen
    October 14, 2013 11:00 AM - 12:00 PM
    Lind Hall 305 [Map]

    Abstract

    I will present a new topological formalism that, for the first time, describes topology as a multi-scale concept. This has a direct and strong relevance to the topological analysis of structure formation process in the cosmos, given that this proceeds in a hierarchical fashion. Rooted in algebraic topology, the concepts I will describe stem from (persistence) homology and Morse theory. Although the mathematical theory behind the concepts have been known for over a century, only recently have they become of practical relevance, due to breakthroughs in computational topology over the last decade. Taking gaussian random fields and the web-like spatial distribution of matter in the Universe as running examples, I will demonstrate that the formalism allows us to describe topology at a level of detail that far supersedes those provided by the standard topological descriptors like Euler characteristic and genus. In this context, I will introduce persistence and ratio landscapes as an empirical statistical description of persistence homology. They prove to be powerful tools to discriminate between the spatial structure emanating in different cosmologies. Most promising is our recent realization that it provides for a novel way of hunting for primordial non-gaussianities. In a related aspect, I will present a software for interactive visualization of the hierarchical topological structures. It exploits the geometric aspects of Morse theory, to detect, describe and quantify the filamentary patterns of the Cosmic web. The software is a promising tool applicable to the analysis of structural patterns in a wide range of astronomical areas of interest -- the detection and characterization of stellar streams at galactic scales being a promising candidate.
  • Beyond the Borsuk-Ulam/Dold theorem (in Combinatorial Geometry)

    Pavle Blagojević, Freie Universität Berlin
    October 16, 2013 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    In this talk we present an evolution of equivariant topology methods in Combinatorial Geometry. We start with (a) the Topological Radon's theorem, an application of the Borsuk-Ulam theorem, and proceed, via non-planarity of K_{3,3}, to (b) the Topological Tverberg and the Weak Colored Tverberg theorem for primes, which are applications of Dold's theorem, to continue with (c) the Topological Tverberg for prime powers, an application beyond Dold's theorem based on the connectivity and localization theorem for elementary abelian groups, to finally ask: "What needs to be done in the case of Barany-Larman conjecture and Nandakumar & Ramana-Rao problem when all the previously known methods fail?"
  • Climate Seminar: Budyko's Model as a Dynamical System

    Richard McGehee, University of Minnesota, Twin Cities
    October 22, 2013 11:15 AM - 12:15 PM
    Lind Hall 305 [Map]

    Abstract forthcoming.

    Read more...

  • Annual Program Series

    Frederick Cohen, University of Rochester
    October 23, 2013 10:30 AM - 12:00 PM
    Lind Hall 305 [Map]

    Abstract forthcoming.

  • Towards a Topos Bais for Persistent Homology

    Mikael Vejdemo-Johansson, Royal Institute of Technology (KTH)
    October 23, 2013 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    We describe a topos of sheaves with the property that classical persistent homology of a filtered complex (should) be the internal simplicial homology functor of the logic specified by the base space of the sheaves. All relevant background to understand definitions and their ramifications will be provided in the talk.
  • Annual Program Series

    Frederick Cohen, University of Rochester
    October 25, 2013 10:30 AM - 12:00 PM
    Lind Hall 305 [Map]

    Abstract forthcoming.

  • Celestial Influences on Glacial Cycles

    Richard McGehee, University of Minnesota, Twin Cities
    October 29, 2013 11:15 AM - 12:15 PM
    Lind Hall 305 [Map]

    Abstract forthcoming.

    Read more...

  • Some Properties of Clustering Methods

    Facundo Mémoli, The Ohio State University
    November 6, 2013 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    I'll describe joint work with Gunnar Carlsson about clustering methods. I will concentrate on the characterization, stability, and convergence of hierarchical clustering methods operating on metric spaces.Some properties of clustering methods
  • Andrew Blumberg, University of Texas, Austin
    November 13, 2013 10:30 AM - 12:00 PM
    Lind Hall 305 [Map]

    Abstract forthcoming.

  • Asymptotic topology and group actions

    Greg Bell, University of North Carolina, Greensboro
    November 13, 2013 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    The large-scale approach to metric spaces was popularized by Gromov who applied these ideas to finitely generated groups. Although this "asymptotic topology" is interesting in its own right, a great deal of interest in large-scale dimension followed from Guoliang Yu’s result that groups with finite asymptotic dimension satisfy the Novikov higher signature conjecture. In this talk we’ll define some dimension-theoretic notions in asymptotic topology and discuss the role geometric group theory plays in proving large-scale analogs of classical results.
  • Andrew Blumberg, University of Texas, Austin
    November 18, 2013 10:30 AM - 12:00 PM
    Keller Hall 3-180 [Map]

    Abstract forthcoming.

  • Limit Theory for Random Cech Complexes

    Yogeshwaran Dhandapani, Technion-Israel Institute of Technology
    November 20, 2013 1:30 PM - 2:30 PM
    Keller Hall 3-180 [Map]

    Abstract

    In this talk, I shall try to describe some asymptotic results for random Cech complexes. These shall concern thresholds for vanishing and formation of Homology groups as well as identify the limiting distribution of the Betti numbers under different regimes. I shall try to emphasize the usefulness of various stochastic geometric tools such as martingale techniques, stabilization theory, Palm formula et al in the context of random topology. Though most of the talk shall be restricted to the case of Poisson or i.i.d. point process, we shall briefly discuss extensions to general point processes too.
  • Persistence Modules and Category Theory

    Vin de Silva, Pomona College
    December 4, 2013 10:30 AM - 12:00 PM
    Lind Hall 305 [Map]

    Abstract forthcoming.

  • Scheduling spaces in concurrency via directed topology

    Martin Raussen, Aalborg University
    December 4, 2013 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    Concurrency theory in Computer Science studies the effects that arise when several processors run simultaneously sharing common resources. It attempts to advise methods to deal with the "state space explosion problem", sometimes using models with a combinatorial/topological flavor. It is a common feature of these models that a schedule corresponds to a directed path (d-path), and that d-homotopies (preserving the directions) result in computations with the same result. I shall discuss particular classical examples of directed spaces, a class of Higher Dimensional Automata (HDA). For such a state space, I shall describe a (nerve lemma) method that determines the homotopy type of the space of traces (schedules) as a prodsimplicial complex - with products of simplices as building blocks. A description of that complex opens up for (machine) calculations of homology groups and other topological invariants of the trace space. The determination of the path components of trace space is particularly important for applications. Unfortunately, the resulting prodsimplicial complexes grow still quickly in both dimension and the number of cells. I shall sketch ongoing work with K. Ziemiański (Warsaw) that tries to find smaller homotopy equivalent simplicial complexes via suitable homotopy decompositions of trace spaces.
  • Persistence Modules and Category Theory

    Vin de Silva, Pomona College
    December 6, 2013 10:30 AM - 12:00 PM
    Lind Hall 305 [Map]

    Abstract forthcoming.

  • The Variational Bicomplex

    Irina Kogan, North Carolina State University
    January 29, 2014 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    Introduction of the variational bicomplex can be motivated by drawing an analogy with vector calculus. It is well known that one can reformulate vector calculus in terms of differential forms, and use the exterior derivative to represent gradient, divergence, and curl operators. Thus vector calculus can be efficiently expressed by the de Rham complex. Similarly to vector calculus, many important aspects of variational calculus, such as Euler-Lagrange operator, Helmholtz operator, or Noether correspondence, can be formulated in terms of complexes of differential forms. This leads to notion of the variational bicomplex (and related notions of the variational complex, the variational spectral sequence). These constructions originated with the work of Dedecker (1957) and a large body of literature has appeared afterwards. The purpose of my talk is to define the variational bicomplex, explain how it encodes various aspects of variational calculus and what role is played by its cohomology.
  • Towards Utilizing Topological Data Analysis for Studying Machining Models

    Firas Khasawneh, State University of New York Institute of Technology
    February 19, 2014 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    Delay differential equations (DDEs) appear in many models in science and engineering either as an intrinsic component or as a modeling decision. Machining dynamics are an important example of processes that include delays as an intrinsic part of the system. The infinite dimensionality of DDEs significantly complicates the resulting analysis from both an analytical and numerical perspective. Recent developments in the theory of DDEs have helped with obtaining more accurate mathematical models for capturing the process dynamics. However, even though machining processes are known to be stochastic, the majority of existing machining models are deterministic. Further, data analysis methods for DDEs are either few or non-existant largely due to the non-Markovian nature of these systems. This talk briefly discusses the challenges associated with DDEs and how they are incorporated into machining dynamics models. We also discuss some new results which suggest that topological data analysis, specifically persistent homology, can be a very useful tool for analyzing numerical and experimental time series corresponding to time delay systems.
  • Lipschitz Extensions and Higher-Order Metric Certificates

    Vin de Silva, Pomona College
    February 26, 2014 1:30 PM - 2:30 PM
    Lind Hall 409 [Map]

    Abstract

    The traditional TDA (topological data analysis) pipeline converts data to a filtered simplicial complex, and the filtered simplicial complexes to a persistence diagram (PD). The PD can then be studied for information about the input data. This process can be recursed: given lots of data sets, each can be turned into a PD, and this collection of PDs can be viewed as a data set in diagram space. One can then build Cech or Vietoris-Rips complexes from this data.
  • Homology-vanishing Theorems for Random Simplicial Complexes

    Matthew Kahle, The Ohio State University
    March 12, 2014 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    Linial-Meshulam random 2-complexes are analogues of Erdős-Rényi random graphs, and their topological properties have been the subject of extensive study. We now know a few methods for proving homology-vanishing theorems in settings like this: (1) cocycle counting, i.e. combinatorial methods, (2) spectral methods, using concentration results for spectral gaps of certain random matrices, and (3) what might be called random linear algebra. Each of these has its advantages, and I will briefly overview each. The main new result I'll discuss by the end of the talk is using the third method to describe the threshold for vanishing of homology with integer coefficients, a problem which was very resistant to earlier methods. This is joint work with Christopher Hoffman and Elliot Paquette.
  • Multiparameter Persistent Homology for Shape Comparison: Continuous versus Discrete

    Tomasz Kaczynski, University of Sherbrooke
    March 19, 2014 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    The theory of multiparameter persistent homology was initially developed in the discrete setting of filtered simplicial complexes. Stability of persistence was proved for topological spaces filtered by continuous vector-valued functions. Our aim is to provide a formal bridge between the continuous setting, where stability properties hold, and the discrete setting, where actual computations are carried out. The existence of this bridge is not obvious for two reasons. One is related to the Sard’s lemma, namely, we do not have the isolation of critical values for generic vector functions. The other one is due to the phenomenon of structural gap between the two settings which appears in the multi-parameter case when using the standard piecewise linear interpolation of the discrete model.
  • Necktie knots, formal languages and network security

    Mikael Vejdemo-Johansson, Royal Institute of Technology (KTH)
    March 26, 2014 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    A chore for some, space for personal expression for others, the necktie knot used to have very few speci?c knots in widespread use. In their 1999 paper, Fink & Mao (Designing tie knots by random walks." Nature 398, no. 6722 (1999): 31-32) list all possible ways to tie a necktie. They limit their enumeration task by focusing on knots that present a flat front just like all the classical tie knots. This way they established a list of 85 possible tie knots. Tie knots with intricate patterns of the necktie winding into symmetric but no longer at front displays have emerged in the past decade, introduced by the movie Matrix Reloaded and recreated hobbyists. These tie knots are tied with the narrow end of the tie, wrapping it to create patterns on the surface of the tie knot. As such these knots are not covered by the listing proposed by Fink & Mao. With a team of collaborators I have extended the listing by Fink and Mao to cover these new tie knots. While doing this we have been able to determine the computational complexity classes of the grammar that describes tie knots. The formal language techniques that help us analyze the tie knot grammars are used in contemporary security research: a large class of security problems online emerge from different implementations of a communications protocol disagreeing on the actual grammar used. We will talk about the connections between the language techniques for tie knots and those that help us analyze security. http://www.newscientist.com/article/dn25019-matrix-villain-spawns-177000-ways-to-knot-a-tie.html http://phys.org/news/2014-02-mathematicians-ways.html
  • Random Linear Algebra and the Vanishing Threshold for Integer Homology

    Matthew Kahle, The Ohio State University
    April 2, 2014 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    There are a few different approaches now to homology-vanishing theorems for random simplicial complexes: (1) cocycle counting, (2) spectral gap methods, and a new approach, (3) random linear algebra. In this talk I will talk about the third approach, and new joint work with Hoffman and Paquette, giving the vanishing threshold for homology with integer coefficients. This was a problem that had resisted attempts by the other methods for several years, and the approach is also elementary. (This is a follow up to an earlier talk, but this talk will also aim to be self contained.)
  • A-infinity Persistence

    Aniceto Murillo, University of Málaga
    April 9, 2014 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    We will introduce A-infinity persistence of a given homology filtration of topological spaces. This is a family homological invariants which provide information not readily available by the (persistent) Betti numbers of the given filtration. This may help to detect noise, not just in the simplicial structure of the filtration but in further geometrical properties in which the higher codiagonals of the A-infinity structure are translated.
  • Topology and coding in neural networks

    Chad Giusti, University of Nebraska
    April 16, 2014 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract

    Natural and artificial neural networks are traditionally studied using classical methods from dynamical systems and linear algebra. However, there are many problems of interest for which topological tools can offer new perspectives. Here, we survey our recent work on the relationship between the structure of a network and that of the data it encodes: first, using classical tools from combinatorial topology, we show that a simple one-layer feed-forward network can encode any prescribed simplicial complex, but that a large class of non-convex codes cannot be encoded; second, using recent notions and computer tools for computing persistent homology, we discover signatures of "geometric structure" in the correlations of activity among neurons in the hippocampal network in rats. In the latter project, we frame persistent homology of clique complexes as a tool for the study of equivalence classes of real symmetric matrices under the action of the group of monotone increasing functions, building on work of M. Kahle describing "Betti curves" of such classes to interpret the results. Time permitting, we will also touch on work in progress regarding detecting convexity of codes. This is in various parts joint work with C. Curto, V. Itskov and W. Kronholm. No background in biology or networks is assumed!
  • Radmila Sazdanović, North Carolina State University
    April 23, 2014 1:30 PM - 2:30 PM
    Lind Hall 305 [Map]

    Abstract forthcoming.

  • Vin de Silva, Pomona College
    May 7, 2014 1:00 PM - 2:00 PM
    Lind Hall 305 [Map]

    Abstract forthcoming.

Previous Annual Seminars

Go