Georgia Institute of Technology College of Sciences

Micro-organisms are known to respond to certain dissolved chemicalsubstances in their environment by moving preferentially away or towardtheir source in a process called chemotaxis. We study such chemotacticresponses at the population level when the micro-swimmers arehydrodynamically coupled to each-other as well as the chemicalconcentration. We include a chemotactic bias based on the known bacteriarun-and-tumble phenomenon in a kinetic model of motile suspension dynamicsdeveloped recently to study hydrodynamic interactions. The chemicalsubstance can be produced or consumed by the swimmers themselves, as wellas be advected by the fluid flows created by their movement. The linearstability analysis of the system will be discussed, as well as the entropyanalysis. Nonlinear dynamics are investigated using numerical simulationin two dimensions of the full system of equations. We show examples ofaggregation in suspensions of pullers (front-actuated swimmers) anddiscuss how chemotaxis affects the mixing flows in suspensions of pushers(rear-actuated swimmers). Last, I will discuss recent work on numericalsimulations of discrete particle/swimmer suspensions that have achemotactic bias.

A theorem of Chris Wendl allows you to completely characterize symplectic fillings of certain open book decompositions by factorizations of their monodromy into Dehn twists. Olga Plamenevskaya and I use this to generalize results of Eliashberg, McDuff and Lisca to classify the fillings of certain Lens spaces. I'll discuss this and a newer version of Wendl's theorem, joint with Wendl and Sam Lisi, this time for spinal open books, and discuss a few more applications.

The k-disjoint paths problem takes as input a graph G and k pairs of vertices (s_1, t_1),..., (s_k, t_k) and determines if there exist internally disjoint paths P_1,..., P_k such that the endpoints of P_i are s_i and t_i for all i=1,2,...,k. While the problem is NP-complete when k is allowed to be part of the input, Robertson and Seymour showed that there exists a polynomial time algorithm for fixed values of k. The existence of such an algorithm is the major algorithmic result of the Graph Minors series. The original proof of Robertson and Seymour relies on the whole theory of graph minors, and consequently is both quite technical and involved. Recent results have dramatically simplified the proof to the point where it is now feasible to present the proof in its entirety. This seminar series will do just that, with the level of detail aimed at a graduate student level.

Aim of the talk is to make a survey on some recent results concerning analysis over spaces with Ricci curvature bounded from below. I will show that the heat flow in such setting can be equivalently built either as gradient flow of the natural Dirichlet energy in L^2 or as gradient flow if the relative entropy in the Wasserstein space. I will also show how such identification can lead to interesting analytic and geometric insights on the structures of the spaces themselves. From a collaboration with L.Ambrosio and G.Savare

I will describe recent work with Marina Chugunovaand Ben Stephens on the evolution of a thin-filmequation that models a "coating flow" on a horizontalcylinder. Formally, the equation defines a gradientflow with respect to an energy that controls theH^1-norm.We show that for each given mass there exists aunique steady state, given by a droplet hanging from thebottom of the cylinder that meets the dry region withzero contact angle. The droplet minimizes the energy andattracts all strong solutions that satisfy certain energyand entropy inequalities. (Such solutions exist for arbitraryinitial values of finite energy and entropy, but it is notknown if they are unique.) The distance of any solutionfrom the steady state decays no faster than a power law.

Consider any dynamical system with the phase space (set of all states) M. One gets an open dynamical system if M has a subset H (hole) such that any orbit escapes ("disappears") after hitting H. The question in the title naturally appears in dealing with some experiments in physics, in some problems in bioinformatics, in coding theory, etc. However this question was essentially ignored in the dynamical systems theory. It occurred that it has a simple and counter intuitive answer. It also brings about a new characterization of periodic orbits in chaotic dynamical systems. Besides, a duality with Dynamical Networks allows to introduce dynamical characterization of the nodes (or edges) of Networks, which complements such static characterizations as centrality, betweenness, etc. Surprisingly this approach allows to obtain new results about such classical objects as Markov chains and introduce a hierarchy in the set of their states in regard of their ability to absorb or transmit an "information". Most of the results come from a finding that one can make finite (rather than traditional large) time predictions on behavior of dynamical systems even if they do not contain any small parameter. It looks plausible that a variety of problems in dynamical systems, probability, coding, imaging ... could be attacked now. No preliminary knowledge is required. The talk will be accessible to students.

Deducing the state or structure of a system from partial, noisy measurements is a fundamental task throughout the sciences and engineering. The resulting inverse problems are often ill-posed because there are fewer measurements available than the ambient dimension of the model to be estimated. In practice, however, many interesting signals or models contain few degrees of freedom relative to their ambient dimension: a small number of genes may constitute the signature of a disease, very few parameters may specify the correlation structure of a time series, or a sparse collection of geometric constraints may determine a molecular configuration. Discovering, leveraging, or recognizing such low-dimensional structure plays an important role in making inverse problems well-posed. In this talk, I will propose a unified approach to transform notions of simplicity and latent low-dimensionality into convex penalty functions. This approach builds on the success of generalizing compressed sensing to matrix completion, and greatly extends the catalog of objects and structures that can be recovered from partial information. I will focus on a suite of data analysis algorithms designed to decompose general signals into sums of atoms from a simple---but not necessarily discrete---set. These algorithms are derived in a convex optimization framework that encompasses previous methods based on l1-norm minimization and nuclear norm minimization for recovering sparse vectors and low-rank matrices. I will provide sharp estimates of the number of generic measurements required for exact and robust recovery of a variety of structured models. These estimates are based on computing certain Gaussian statistics related to the latent model geometry. I will detail several example applications and describe how to scale the corresponding inference algorithms to very large data sets. (Joint work with Venkat Chandrasekaran, Pablo Parrilo, and Alan Willsky)

The spectacular success of search and display advertising -- to businesses and search engine companies -- and its huge growth potential has attracted the attention of researchers from many aspects of computer science. Since a core problem in this area is that of effective ad allocation, an inherently algorithmic and game-theoretic question, numerous theoreticians have worked in this area in recent years. Ad allocation involves matching ad slots to advertisers, under demand and supply constraints. In short, the better the matching, the more efficient the market. Interestingly, the seminal work on online matching, by Karp, Vazirani and Vazirani, was done over two decades ago, well before the advent of the Internet economy! In this talk, I will give an overview of several key algorithmic papers in this area, starting with its purely academic beginnings, to papers that actually address the Adwords problem. The elegant -- and deep -- theory behind these algorithms involves new combinatorial, probabilistic and linear programming techniques. Besides the classic KVV paper (STOC 1990), this talk will refer to several papers with my co-authors: Mehta, Saberi, Vazirani, Vazirani (FOCS 05, J. ACM 07), Goel, Mehta (SODA 08), Feldman, Mehta, Mirrokni, Muthukrishnan (FOCS 09), Aggarwal, Goel, Karande, Mehta (SODA 10), Karande, Mehta, Tripathi (STOC 11).

We begin this talk by discussing four different problems that arenumber theoretic or combinatorial in nature. Two of these problems remainopen and the other two have known solutions. We then explain how these seeminglyunrelated problems are connected to each other. To disclose a little more information,one of the problems with a known solution is the following: Is it possible to find anirrational number $q$ such that the infinite geometric sequence $1, q, q^{2}, \dots$has 4 terms in arithmetic progression?

Torsion of a curve in Euclidean 3-space is a quantity which together with the curvature completely determines the curve up to a rigid motion. In this talk we use the curve shortening flow to show that the number of zero torsion points (or vertices) v a closed space curve c and the number p of the pair of parallel tangent lines of c satisfy the following sharp inequality: v + 2p > 5.

I'll present a new, simple proof that the Torelli group is generated by (infinitely many) bounding pair maps. At the end, I'll explain an application of this approach to the hyperelliptic Torelli group. The key is to take advantage of the "complex of minimizing cycles."

The Chv\'atal--Erd\H{o}s Theorem states that every graph whose connectivityis at least its independence number has a spanning cycle. In 1976, Fouquet andJolivet conjectured an extension: If $G$ is an $n$-vertex $k$-connectedgraph with independence number $a$, and $a \ge k$, then $G$ has a cycle of lengthat least $\frac{k(n+a-k)}{a}$. We prove this conjecture. This is joint work with Suil O and Douglas B. West.