Seminars and Colloquia by Series

A modern take on Huber regression

Series
School of Mathematics Colloquium
Time
Thursday, April 22, 2021 - 12:00 for 1 hour (actually 50 minutes)
Location
https://us02web.zoom.us/j/87011170680?pwd=ektPOWtkN1U0TW5ETFcrVDNTL1V1QT09
Speaker
Po-Ling LohUniversity of Cambridge

Please Note: Note the unusual time: 12:00pm.

In the first part of the talk, we discuss the use of a penalized Huber M-estimator for high-dimensional linear regression. We explain how a fairly straightforward analysis yields high-probability error bounds that hold even when the additive errors are heavy-tailed. However, the parameter governing the shape of the Huber loss must be chosen in relation to the scale of the error distribution. We discuss how to use an adaptive technique, based on Lepski's method, to overcome the difficulties traditionally faced by applying Huber M-estimation in a context where both location and scale are unknown.

In the second part of the talk, we turn to a more complicated setting where both the covariates and responses may be heavy-tailed and/or adversarially contaminated. We show how to modify the Huber regression estimator by first applying an appropriate "filtering" procedure to the data based on the covariates. We prove that in low-dimensional settings, this filtered Huber regression estimator achieves near-optimal error rates. We further show that the commonly used least trimmed squares and least absolute deviation estimators may similarly be made robust to contaminated covariates via the same covariate filtering step. This is based on joint work with Ankit Pensia and Varun Jog.

An Alexander method for infinite-type surfaces

Series
Geometry Topology Student Seminar
Time
Wednesday, April 21, 2021 - 14:00 for 1 hour (actually 50 minutes)
Location
ONLINE
Speaker
Roberta Shapiro

Given a surface S, the Alexander method is a combinatorial tool used to determine whether two homeomorphisms are isotopic. This statement was formalized in A Primer on Mapping Class Groups in the case that S is of finite type. We extend the Alexander method to include infinite-type surfaces, which are surfaces with infinitely generated fundamental groups.

In this talk, we will introduce a technique useful in proofs dealing with infinite-type surfaces. Then, we provide a "proof by example" of an infinite-type analogue of the Alexander method.

An analytical study of intermittency through Riemann’s non-differentiable functions

Series
Analysis Seminar
Time
Wednesday, April 21, 2021 - 14:00 for 1 hour (actually 50 minutes)
Location
ONLINE — see abstract for the Zoom link
Speaker
Victor Vilaça Da RochaGeorgia Tech

Intermittency is a property observed in the study of turbulence. Two of the most popular ways to measure it are based on the concept of flatness, one with structure functions in the physical space and the other one with high-pass filters in the frequency space. Experimental and numerical simulations suggest that the two approaches do not always give the same results. In this talk, we prove they are not analytically equivalent. For that, we first adapt them to a rigorous mathematical language, and we test them with generalizations of Riemann’s non-differentiable function. This work is motivated by the discovery of Riemann’s non-differentiable function as a trajectory of polygonal vortex filaments.

The seminar will be held on Zoom.  Here is the link

https://us02web.zoom.us/j/71579248210?pwd=d2VPck1CbjltZStURWRWUUgwTFVLZz09

Universal graph product structures

Series
Graph Theory Seminar
Time
Tuesday, April 20, 2021 - 17:45 for 1 hour (actually 50 minutes)
Location
https://us04web.zoom.us/j/77238664391. For password, please email Anton Bernshteyn (bahtoh ~at~ gatech.edu)
Speaker
David WoodMonash University

Please Note: Note the unusual time: 5:45pm.

This talk will survey recent results that describe graphs as subgraphs of products of simpler graphs. The starting point is the following theorem: every planar graph is a subgraph of the strong product of some treewidth 7 graph and a path. This result has been the key to solving several open problems, for example, regarding the queue-number and nonrepetitive chromatic number of planar graphs. The result generalises to provide a universal graph for planar graphs. In particular, if $T^7$ is the universal treewidth 7 graph (which is explicitly defined), then every countable planar graph is a subgraph of the strong product of $T^7$ and the infinite 1-way path. This result generalises in various ways for many sparse graph classes: graphs embeddable on a fixed surface, graphs excluding a fixed minor, map graphs, etc. For example, we prove that for every fixed graph $X$, there is an explicitly defined countable graph $G$ that contains every countable $X$-minor-free graph as a subgraph, and $G$ has several desirable properties such as every $n$-vertex subgraph of $G$ has a $O(\sqrt{n})$-separator. On the other hand, as a lower bound we strengthen a theorem of Pach (1981) by showing that if a countable graph $G$ contains every countable planar graph, then $G$ must contain an infinite complete graph as a minor. 

3-manifolds that bound no definite 4-manifold

Series
Geometry Topology Seminar
Time
Monday, April 19, 2021 - 14:00 for 1 hour (actually 50 minutes)
Location
ONLINE
Speaker
Marco GollaUniversité de Nantes

All 3-manifolds bound 4-manifolds, and many construction of 3-manifolds automatically come with a 4-manifold bounding it. Often times these 4-manifolds have definite intersection form. Using Heegaard Floer correction terms and an analysis of short characteristic covectors in bimodular lattices, we give an obstruction for a 3-manifold to bound a definite 4-manifold, and produce some concrete examples. This is joint work with Kyle Larson.

Symmetrically processed splitting integrators for enhanced Hamiltonian Monte Carlo sampling

Series
Applied and Computational Mathematics Seminar
Time
Monday, April 19, 2021 - 14:00 for 1 hour (actually 50 minutes)
Location
ONLINE https://bluejeans.com/884917410
Speaker
Prof. Sergio BlanesUniversidad Politécnica de Valencia

We construct integrators to be used in Hamiltonian (or Hybrid) Monte Carlo sampling. The new integrators are easily implementable and, for a given computational budget, may deliver five times as many accepted proposals as standard leapfrog/Verlet without impairing in any way the quality of the samples. They are based on a suitable modification of the   processing technique first introduced by J.C. Butcher. The idea of modified processing may also be useful for other purposes, like the construction of high-order splitting integrators with positive coefficients.

Joint work with Mari Paz Calvo, Fernando Casas, and Jesús M. Sanz-Serna

Approximate Schauder Frames for Banach Sequence Spaces

Series
Dissertation Defense
Time
Friday, April 16, 2021 - 16:00 for 1.5 hours (actually 80 minutes)
Location
ONLINE
Speaker
Yam-Sung ChengGeorgia Institute of Technology

The main topics of this thesis concern two types of approximate Schauder frames for the Banach sequence space $\ell_1$. The first main topic pertains to finite-unit norm tight frames (FUNTFs) for the finite-dimensional real sequence space $\ell_1^n$. We prove that for any $N \geq n$, FUNTFs of length $N$ exist for real $\ell_1^n$. To show the existence of FUNTFs, specific examples are constructed for various lengths. These constructions involve repetitions of frame elements. However, a different method of frame constructions allows us to prove the existence of FUNTFs for real $\ell_1^n$ of lengths $2n-1$ and $2n-2$ that do not have repeated elements.

The second main topic of this thesis pertains to normalized unconditional Schauder frames for the sequence space $\ell_1$. A Schauder frame provides a reconstruction formula for elements in the space, but need not be associated with a frame inequality. Our main theorem on this topic establishes a set of conditions under which an $\ell_1$-type of frame inequality is applicable towards unconditional Schauder frames. A primary motivation for choosing this set of hypotheses involves appropriate modifications of the Rademacher system, a version of which we prove to be an unconditional Schauder frame that does not satisfy an $\ell_1$-type of frame inequality.

Bluejeans link to meeting: https://bluejeans.com/544995272
 

Mutation probabilities and moments of step functions

Series
Mathematical Biology Seminar
Time
Friday, April 16, 2021 - 15:00 for 1 hour (actually 50 minutes)
Location
ONLINE
Speaker
Zvi RosenFlorida Atlantic University

Suppose that n sample genomes are collected from the same population. The expected sample frequency spectrum (SFS) is the vector of probabilities that a mutation chosen at random will appear in exactly k out of the n individuals. This vector is known to be highly dependent on the population size history (demography); for this reason, geneticists have used it for demographic inference. What does the set of all possible vectors generated by demographies look like? What if we specify that the demography has to be piecewise-constant with a fixed number of pieces? We will draw on tools from convex and algebraic geometry to answer these and related questions.

Meeting Link: https://gatech.bluejeans.com/348270750

Symbolic dynamics and oscillatory motions in the 3 Body Problem

Series
CDSNS Colloquium
Time
Friday, April 16, 2021 - 13:00 for 1 hour (actually 50 minutes)
Location
Zoom (see additional notes for link)
Speaker
Pau MartinUPC

Please Note: Zoom link: https://zoom.us/j/97732215148?pwd=Z0FBNXNFSy9mRUx3UVk4alE4MlRHdz09

Consider the three body problem with masses $m_0,m_1,m_2>0$. Take units such that $m_0+m_1+m_2 = 1$. In 1922 Chazy classified the possible final motions of the three bodies, that is the behaviors the bodies may have when time tends to infinity. One of them are what is known as oscillatory motions, that is, solutions of the three body problem such that the positions of the bodies $q_0, q_1, q_2$ satisfy
\[
\liminf_{t\to\pm\infty}\sup_{i,j=0,1,2, i\neq j}\|q_i-q_j\|<+\infty \quad \text{ and }\quad 
\limsup_{t\to\pm\infty}\sup_{i,j=0,1,2, i\neq j}\|q_i-q_j\|=+\infty.
\] At the time of Chazy, all types of final motions were known, except the oscillatory ones. We prove that, if all three masses $m_0,m_1,m_2>0$ are not equal to $1/3$, such motions exist. In fact, we prove more, since our result is based on the construction of a hyperbolic invariant set whose dynamics is conjugated to the Bernouilli shift of infinite symbols, we prove (if all masses are not all three equals to $1/3$) 1) the existence of chaotic motions and positive topological entropy for the three body problem, 2) the existence of periodic orbits of arbitrarily large period in the 3BP. Reversing time, Chazy's classification describes ``starting'' motions and then, the question if starting and final motions need to coincide or may be different arises.  We also prove that one can construct solutions of the three body problem whose starting and final motions are of different type.

Two approximate versions of Jackson’s conjecture [Special time/day!]

Series
Combinatorics Seminar
Time
Thursday, April 15, 2021 - 18:00 for 1 hour (actually 50 minutes)
Location
https://bluejeans.com/751242993/PASSWORD (To receive the password, please email Lutz Warnke)
Speaker
Anita LiebenauUNSW Sydney

A diregular bipartite tournament is a balanced complete bipartite graph whose edges are oriented so that every vertex has the same in- and outdegree. 
In 1981, Jackson showed that a diregular bipartite tournament contains a Hamilton cycle, and conjectured that in fact the edge set of it can be partitioned into Hamilton cycles. 
We prove an approximate version of this conjecture: for every $\epsilon>0$ there exists $n_0$ such that every diregular bipartite tournament on $2n>n_0$  vertices contains a collection of $(1/2-\epsilon)n$ cycles of length at least $(2-\epsilon)n$. 
Increasing the degree by a small proportion allows us to prove the existence of many Hamilton cycles: for every $c>1/2$ and $\epsilon>0$ there exists $n_0$ such that every $cn$-regular bipartite digraph on $2n>n_0$ vertices contains $(1-\epsilon)cn$ edge-disjoint Hamilton cycles.

Base on joint work with Yanitsa Pehova, see https://arxiv.org/abs/1907.08479

Please note the special time/day: Thursday 6pm

Pages