Seminars and Colloquia by Series

Extreme value theory for random walks in space-time random media

Series
Job Candidate Talk
Time
Wednesday, January 29, 2025 - 11:00 for 1 hour (actually 50 minutes)
Location
Skiles 006
Speaker
Shalin ParekhUniversity of Maryland

 

The KPZ equation is a singular stochastic PDE arising as a scaling limit of various physically and probabilistically interesting models. Often, this equation describes the “crossover” between Gaussian and non-Gaussian fluctuation behavior in simple models of interacting particles, directed polymers, or interface growth. It is a difficult and elusive open problem to elucidate the nature of this crossover for general stochastic interface models. In this talk, I will discuss a series of recent works where we have made progress in understanding the KPZ crossover for models of random walks in dynamical random media. This was done through a tilting-based approach to study the extreme tails of the quenched probability distribution. This talk includes joint work with Sayan Das and Hindy Drillick.

Zoom link:

https://gatech.zoom.us/j/96535844666

Separators in sphere intersection graphs

Series
Graph Theory Seminar
Time
Tuesday, January 28, 2025 - 15:30 for 1 hour (actually 50 minutes)
Location
Skiles 255
Speaker
Rose McCartyGeorgia Tech

We discuss the sphere dimension of a graph. This is the smallest integer $d$ such that the graph can be represented as the intersection graph of a collection of spheres in $\mathbb{R}^d$. We show that graphs with small sphere dimension have small balanced separators, as long as they exclude a complete bipartite graph $K_{t,t}$. This property is connected to forbidding shallow minors and can be used to develop divide-and-conquer algorithms. This is joint work with James Davies, Agelos Georgakopoulos, and Meike Hatzel.

Recent progress on the horocycle flow on strata of translation surfaces - NEW DATE

Series
Job Candidate Talk
Time
Tuesday, January 28, 2025 - 11:00 for 1 hour (actually 50 minutes)
Location
Skiles 006
Speaker
Jon ChaikaUniversity of Utah

For about 2 decades the horocycle flow on strata of translation surfaces was studied, very successfully, in analogy with unipotent flows on homogeneous spaces, which by work of Ratner, Margulis, Dani and many others, have striking rigidity properties. In the past decade Eskin-Mirzakhani and Eskin-Mirzakhani-Mohammadi proved some analogous rigidity results for SL(2,R) and the full upper triangular subgroup on strata of translation surfaces. This talk will begin by introducing ergodic theory and translation surfaces. Then it will describe some of the previously mentioned rigidity theorems before moving on to its goal, that many such rigidity results fail for the horocycle flow on strata of translation surfaces. Time permitting we will also describe a rigidity result for special sub-objects in strata of translation surfaces. This will include joint work with Osama Khalil, John Smillie, Barak Weiss and Florent Ygouf. 

 

https://gatech.zoom.us/j/95951300274?pwd=dZE89RkP2k6Ri4xbgJP3cSucsi9xna.1

Meeting ID: 959 5130 0274
Passcode: 412458

Symmetries of Legendrian links and their exact Lagrangian fillings

Series
Geometry Topology Seminar
Time
Monday, January 27, 2025 - 14:00 for 1 hour (actually 50 minutes)
Location
Skiles 006
Speaker
James HughesDuke University

Given a Legendrian link L in the contact 3-sphere, one can hope to classify the set of exact Lagrangian fillings of L, i.e. exact Lagrangian surfaces in the symplectic 4-ball with boundary equal to L. Much of the recent progress towards this classification relies on establishing a connection between sheaf-theoretic invariants of Legendrians and cluster algebras. In this talk, I will describe this connection and how these invariants behave with respect to certain symmetries of Legendrian links and their fillings. Parts of this are joint work with Agniva Roy.

From centralized to federated learning of neural operators: Accuracy, efficiency, and reliability

Series
Applied and Computational Mathematics Seminar
Time
Monday, January 27, 2025 - 14:00 for 1 hour (actually 50 minutes)
Location
Skiles 005 and https://gatech.zoom.us/j/94954654170
Speaker
Lu LuYale University

As an emerging paradigm in scientific machine learning, deep neural operators pioneered by us can learn nonlinear operators of complex dynamic systems via neural networks. In this talk, I will present the deep operator network (DeepONet) to learn various operators that represent deterministic and stochastic differential equations. I will also present several extensions of DeepONet, such as DeepM&Mnet for multiphysics problems, DeepONet with proper orthogonal decomposition or Fourier decoder layers, MIONet for multiple-input operators, and multifidelity DeepONet. I will demonstrate the effectiveness of DeepONet and its extensions to diverse multiphysics and multiscale problems, such as bubble growth dynamics, high-speed boundary layers, electroconvection, hypersonics, geological carbon sequestration, full waveform inversion, and astrophysics. Deep learning models are usually limited to interpolation scenarios, and I will quantify the extrapolation complexity and develop a complete workflow to address the challenge of extrapolation for deep neural operators. Moreover, I will present the first operator learning method that only requires one PDE solution, i.e., one-shot learning, by introducing a new concept of local solution operator based on the principle of locality of PDEs. I will also present the first systematic study of federated scientific machine learning (FedSciML) for approximating functions and solving PDEs with data heterogeneity.

Tensor decompositions with applications to LU and SLOCC equivalence of multipartite pure states

Series
Algebra Seminar
Time
Monday, January 27, 2025 - 13:00 for 1 hour (actually 50 minutes)
Location
Skiles 005
Speaker
Ian TanAuburn University

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

We introduce a broad lemma, one consequence of which is the higher order singular value decomposition (HOSVD) of tensors defined by DeLathauwer, DeMoor and Vandewalle (2000). By an analogous application of the lemma, we find a complex orthogonal version of the HOSVD. Kraus's (2010) algorithm used the HOSVD to compute normal forms of almost all n-qubit pure states under the action of the local unitary group. Taking advantage of the double cover SL2(C)×SL2(C)→SO4(C) , we produce similar algorithms (distinguished by the parity of n) that compute normal forms for almost all n-qubit pure states under the action of the SLOCC group.

Convexity in Whitney Problems

Series
Colloquia
Time
Monday, January 27, 2025 - 11:00 for 1 hour (actually 50 minutes)
Location
Skiles 006
Speaker
Marjorie DrakeMIT

Please Note: This is a job talk, it will be also broadcast by Zoom, in addition to in-person: https://gatech.zoom.us/j/91499035568

Let $E \subset \mathbb{R}^n$ be a compact set, and $f: E \to \mathbb{R}$. How can we tell if there exists a smooth convex extension $F \in C^{1,1}(\mathbb{R}^n)$ of $f$, i.e. satisfying $F|_E = f|_E$? Assuming such an extension exists, how small can one take the Lipschitz constant $\text{Lip}(\nabla F): = \sup_{x,y \in \mathbb{R}^n, x \neq y} \frac{|\nabla F(x) - \nabla F(y)|}{|x-y|}$? I will provide an answer to these questions for the non-linear space of strongly convex functions by presenting recent work of mine proving there is a Finiteness Principle for strongly convex functions in $C^{1,1}(\mathbb{R}^n)$. This work is the first attempt to understand the constrained interpolation problem for *convex* functions in $C^{1,1}(\mathbb{R}^n)$, building on techniques developed by P. Shvartsman, C. Fefferman, A. Israel, and K. Luli to understand whether a function has a smooth extension despite obstacles to their direct application. We will finish with a discussion of challenges in adapting my proof of a Finiteness Principle for the space of convex functions in $C^{1,1}(\mathbb{R})$ ($n=1$) to higher dimensions.

On the Three-Dimensional, Quadratic Diffeomorphism: Anti-integrability, Attractors, and Chaos

Series
CDSNS Colloquium
Time
Friday, January 24, 2025 - 15:30 for 1 hour (actually 50 minutes)
Location
Skiles 314
Speaker
Amanda HamptonGeorgia Tech

Please Note: Zoom link (if needed): https://gatech.zoom.us/j/91390791493?pwd=QnpaWHNEOHZTVXlZSXFkYTJ0b0Q0UT09

 

We give a comprehensive parameter study of the three-dimensional quadratic diffeomorphism to understand its attracting and chaotic dynamics. For large parameter values, we use a concept introduced 30 years ago for the Frenkel--Kontorova model of condensed matter physics: the anti-integrable (AI) limit. At the traditional AI limit, orbits of a map degenerate to sequences of symbols and the dynamics is reduced to the shift operator, a pure form of chaos. For the 3D quadratic map, the AI limit that we study becomes a pair of one-dimensional maps, introducing symbolic dynamics on two symbols. Using contraction arguments, we find parameter domains such that each symbol sequence corresponds to a unique AI state. In some of these domains, sufficient conditions are then found for each such AI state to continue away from the limit becoming an orbit of the original 3D map. Numerical continuation methods extend these results, allowing computation of bifurcations, and allowing us to obtain orbits with horseshoe-like structures and intriguing self-similarity.

For small parameter values, we focus on the dissipative, orientation preserving case to study the codimension-one and two bifurcations. Periodic orbits, born at resonant, Neimark-Sacker bifurcations, give rise to Arnold tongues in parameter space. Aperiodic attractors include invariant circles and chaotic orbits; these are distinguished by rotation number and Lyapunov exponents. Chaotic orbits include Hénon-like and Lorenz-like attractors, which can arise from period-doubling cascades, and those born from the destruction of invariant circles. The latter lie on paraboloids near the local unstable manifold of a fixed point.

Lastly, we present a generalized proof for the existence of AI states using similar contraction arguments to find larger parameter domains for the one-to-one correspondence of symbol sequences and AI states. We apply numerical continuation to these results to determine the persistence of low-period and heteroclinic AI states to the full, deterministic 3D map for a volume-contracting case. We find the corresponding AI state of a chaotic attractor and continue this state towards the full map. The numerical results show that the AI states continue to resonant and chaotic attractors along a 3D folded horseshoe that is similar to the classical 2D Hénon attractor.

On the Meissner state for type-II inhomogeneous superconductors

Series
Math Physics Seminar
Time
Friday, January 24, 2025 - 11:00 for 1 hour (actually 50 minutes)
Location
Skiles 005
Speaker
Carlos RománPontificia Universidad Católica de Chile

The Ginzburg-Landau model is a phenomenological description of superconductivity. A key feature of type-II superconductors is the emergence of singularities, known as vortices, which occur when the external magnetic field exceeds the first critical field. Determining the location and number of these vortices is crucial. Furthermore, the presence of impurities in the material can influence the configuration of these singularities.

In this talk, I will present an estimation of the first critical field for inhomogeneous type-II superconductors and show that the model admits stable local minimizers without vortices, corresponding to Meissner type solutions, even when the external magnetic field intensity significantly exceeds the first critical field, approaching the so-called superheating field. This work is in collaboration with Matías Díaz-Vera.

Expansion and torsion homology of 3-manifolds

Series
Job Candidate Talk
Time
Thursday, January 23, 2025 - 11:00 for 1 hour (actually 50 minutes)
Location
Skiles 006
Speaker
Jonathan ZungMIT

We say that a Riemannian manifold has good higher expansion if every rationally null-homologous i-cycle bounds an i+1 chain of comparatively small volume. The interactions between expansion, spectral geometry, and topology have long been studied in the settings of graphs and surfaces. In this talk, I will explain how to construct rational homology 3-spheres which are good higher expanders. On the other hand, I will show that such higher expanders must be rather topologically complicated; in particular, we will demonstrate a super-polynomial-in-volume lower bound on their torsion homology.

Pages