Seminars and Colloquia by Series

New perspectives on learning networks from dynamics

Series
Stochastics Seminar
Time
Tuesday, November 4, 2025 - 15:30 for 1 hour (actually 50 minutes)
Location
Skiles 006
Speaker
Ani SridharNew Jersey Institute of Technology

Suppose that a continuous-time, stochastic diffusion (i.e., the Susceptible-Infected process) spreads on an unknown graph. We only observe the time at which the diffusion reaches each vertex, i.e., the set of infection times. What can be learned about the unknown graph from the infection times? While there is far too little information to learn individual edges in the graph, we show that certain high-level properties -- such as the number of vertices of sufficiently high degree, or super-spreaders -- can surprisingly be determined with certainty. To achieve this goal, we develop a suite of algorithms that can efficiently detect vertices of degree asymptotically greater than sqrt(n) from infection times, for a natural and general class of graphs with n vertices. To complement these results, we show that our algorithms are information-theoretically optimal: there exist graphs for which it is impossible to tell whether vertices of degree larger than n^{1/2 - \epsilon} exist from vertices' infection times, for any \epsilon > 0. Finally, we discuss the broader implications of our ideas for change-point detection in non-stationary point processes. This talk is based on joint work with Anna Brandenberger (MIT) and Elchanan Mossel (MIT).

Lipschitz functions on weak expanders

Series
Graph Theory Seminar
Time
Tuesday, November 4, 2025 - 15:30 for 1 hour (actually 50 minutes)
Location
Skiles 005
Speaker
Lina LiUniversity of Mississippi

Given a connected finite graph $G$, an integer-valued function $f$ on $V(G)$ is called $M$-Lipschitz if the value of $f$ changes by at most $M$ along the edges of $G$. In 2013, Peled, Samotij, and Yehudayoff showed that random $M$-Lipschitz functions on graphs with sufficiently good expansion typically exhibit small fluctuations, giving sharp bounds on the typical range of such functions, assuming $M$ is not too large. We prove that the same conclusion holds under a relaxed expansion condition and for larger $M$, (partially) answering questions of Peled et al. Our approach combines Sapozhenko’s graph container method with entropy techniques from information theory.

 

This is joint work with Krueger and Park.

New perspectives on Heegaard Floer satellite operators

Series
Geometry Topology Seminar
Time
Monday, November 3, 2025 - 14:00 for 1 hour (actually 50 minutes)
Location
Skiles 006
Speaker
Ian ZemkeUniversity of Oregon

Satellite operations are one of the most basic operations in knot theory. Many researchers have studied the behavior of knot Floer homology under satellite operations. Most of these results use Lipshitz, Ozsvath and Thurston's bordered Heegaard Floer theory. In this talk, we discuss a new technique for studying these operators, and we apply this technique to a family of operators called L-space operators. Using this theory, we are able in many cases to give a simple formula for the behavior of the concordance invariant tau under such operators. This formula generalizes a large number of existing formulas for the behavior of tau under satellite operations (such as cabling, 1-bridge braids and generalized Mazur patterns), and also has a number of topological applications. This is joint work with Daren Chen and Hugo Zhou.

Multiscale Representation and Learning of Molecules

Series
Applied and Computational Mathematics Seminar
Time
Monday, November 3, 2025 - 14:00 for 1 hour (actually 50 minutes)
Location
Skiles 005 and https://gatech.zoom.us/j/94954654170
Speaker
Bao WangUniversity of Utah

Artificial intelligence (AI) has become a transformative force in scientific discovery---known as AI for Science---with profound impact on computational molecular design, as highlighted by the 2024 Nobel Prize in Chemistry. Due to their remarkable capability in analyzing complex structures, message-passing neural networks and diffusion- and flow-based generative models stand out as effective tools for molecular property prediction and structure generation. However, message-passing neural networks struggle to efficiently integrate multiscale molecular features and complex 3D geometry for accurate property prediction, and (2) the generative processes of generative models are often computationally intensive and error-prone. 

In this talk, I will present our recent advances toward overcoming these limitations: (1) multiscale graph representations and message-passing architectures for efficient and accurate molecular learning, and (2) one-step flow-based generative models that enable high-fidelity molecule generation with dramatically reduced computational cost.

Diophantine Approximation and Solving Sparse Polynomial Equations

Series
Algebra Seminar
Time
Monday, November 3, 2025 - 13:00 for 1 hour (actually 50 minutes)
Location
Skiles 005
Speaker
Maurice RojasTexas A&M University

Please Note: There will be a pre-seminar 10:55-11:15 in Skiles 005.

Suppose $f$ is a univariate polynomial with integer coefficients of absolute value at most $H$, exactly $t$ monomial terms, and degree $d$. Suppose also that $p,q$ are integers of absolute value at most $H$. Then one can determine the sign of $f(p/q)$ in time $(d^3 \log H)^{2+\epsilon}$, by combining work of Liouville from about 170 years ago, work of Mahler from about 60 years ago, work of Neff, Reif, and Pan from about 30 years ago, and more recent refinements.

However, when $t$ is small, one can do much better: When $t=2$, one can determine the sign of $f(p/q)$ in time $\log^5(dH)$, via work of Baker from about 55 years ago. Koiran asked, around 2016, about the $t=3$ case, after proving new bounds on the minimal separation of complex roots of univariate trinomials.

We make progress on Koiran's question by giving a new, dramatically faster algorithm for solving univariate trinomials over the real numbers. The key innovation is a new family of non-hypergeometric series for the roots of $f$ when $f$ is close to having a degenerate root. This is joint work with Emma Boniface and Weixun Deng.

The Brownian and Poisson transport maps

Series
Stochastics Seminar
Time
Friday, October 31, 2025 - 13:00 for 1 hour (actually 50 minutes)
Location
Skiles 006
Speaker
Yair ShenfeldBrown University

Please Note: Please note the nonstandard date (Friday) and time (1pm).

Transport maps serve as a powerful tool to transfer information from source to target measures. However, this transfer of information is possible only if the transport map is sufficiently regular, which is often difficult to show. I will explain how taking the source measure to be an infinite-dimensional measure, and building transport maps based on stochastic processes, solves some of these challenges both in the continuous and discrete settings. 

First critical field in the 3D Ginzburg-Landau model for inhomogeneous type-II superconductors

Series
Math Physics Seminar
Time
Friday, October 31, 2025 - 11:00 for 1 hour (actually 50 minutes)
Location
SKiles 006
Speaker
Matias Diaz VeraGeorgia Tech

We characterize global minimizers below the so-called first critical field of the inhomogeneous version of the Ginzburg-Landau energy functional in a three-dimensional setting. Minimizers of this functional describe the behavior of type-II superconductors exposed to an external magnetic field, which is characterized by the presence of codimension 2 singularities called vortices where superconductivity is locally suppressed. We will talk about how to adapt the results from the standard Ginzburg-Landau theory into an inhomogeneous framework and present results from a recent work in collaboration with Carlos Roman (Pontificia Universidad Catolica de Chile).

The Convexity Conjecture, the Kahn-Kalai Conjecture, and introduction to k-thresholds

Series
Atlanta Combinatorics Colloquium
Time
Thursday, October 30, 2025 - 04:44 for 1 hour (actually 50 minutes)
Location
Skiles 006
Speaker
Jinyoung Park Courant Institute of Mathematical Sciences NYU

Please Note: Light refreshments will be offered before the talk at 4pm in the atrium.

The "Convexity Conjecture" by Talagrand asks (very roughly) whether one can "create convexity" in constant steps regardless of the dimension of the ambient space. Talagrand also suggested a discrete version of the Convexity Conjecture and called it "my lifetime favorite problem," offering $1,000 prize for its solution. We introduce a reformulation of the discrete Convexity Conjecture using the new notion of "k-thresholds," which is an extension of the traditional notion of thresholds, introduced by Talagrand. Some ongoing work on understanding k-thresholds, along with a (vague) connection between the Kahn-Kalai Conjecture and the discrete Convexity Conjecture, will also be discussed. Joint work with Michel Talagrand.

A High-Frequency Uncertainty Principle for the Fourier-Bessel Transform

Series
Analysis Seminar
Time
Wednesday, October 29, 2025 - 14:00 for 1 hour (actually 50 minutes)
Location
Skiles 005
Speaker
Rahul SethiGeorgia Institute of Technology

Motivated by problems in control theory concerning decay rates for the damped wave equation $$w_{tt}(x,t) + \gamma(x) w_t(x,t) + (-\Delta + 1)^{s/2} w(x,t) = 0,$$ we consider an analogue of the classical Paneah-Logvinenko-Sereda theorem for the Fourier Bessel transform. In particular, if $E \subset \mathbb{R}^+$ is $\mu_\alpha$-relatively dense (where $d\mu_\alpha(x) \approx x^{2\alpha+1}\, dx$) for $\alpha > -1/2$, and $\operatorname{supp} \mathcal{F}_\alpha(f) \subset [R,R+1]$, then we show $$\|f\|_{L^2_\alpha(\mathbb{R}^+)} \lesssim \|f\|_{L^2_\alpha(E)},$$ for all $f\in L^2_\alpha(\mathbb{R}^+)$, where the constants in $\lesssim$ do not depend on $R > 0$. Previous results on PLS theorems for the Fourier-Bessel transform by Ghobber and Jaming (2012) provide bounds that depend on $R$. In contrast, our techniques yield bounds that are independent of $R$, offering a new perspective on such results. This result is applied to derive decay rates of radial solutions of the damped wave equation. This is joint work with Ben Jaye.   

Reverse-engineering exotic 4-manifolds

Series
Geometry Topology Student Seminar
Time
Wednesday, October 29, 2025 - 14:00 for 1 hour (actually 50 minutes)
Location
Skiles 006
Speaker
Cooper KofronGeorgia Tech

4-manifold topology is characterized by unexpected differences between the smooth and topological categories. For instance, it is the only dimension where there can exist infinitely many manifolds $Y_i$ which are homeomorphic to but not diffeomorphic to $X$. A natural question: how does one construct examples of this phenomenon? In this talk, we focus on the method of reverse engineering, which allows for the construction of “small” exotic 4-manifolds. Surprisingly, symplectic geometry is the main ingredient that makes this approach work! We survey the known results related to reverse engineering, and try to pinpoint an error in a paper of Akhmedov-Park, which claimed the existence of an exotic $S^2 \times S^2$.

Pages