Homepage
Time & Location*: |
Wednesdays from 11am – 12pm in Hill 005 |
|
(*Some talks may be scheduled for different times or locations. Such details will be provided additionally.) |
Organizers: |
Narek Hovsepyannh507@math.rutgers.edu |
Gokul Nairgokul.nair@rutgers.edu |
|
(For inquiries, e.g. to be added to the mailing list, please contact either one of the organizers.) |
![]() |
*27 January (Monday) at 4 – 5 pm in Hill 525 |
Gabriel Rioux, Cornell University“Gromov-Wasserstein Alignment: Statistics and Computation.” |
AbstractOver the past decade, the statistical and computational properties of optimal transport (OT) have been systematically studied, driven, in part, to its broad applicability to data science. This program has culminated in an in depth understanding of the curse of dimensionality that OT distances suffer and spurred the development of computationally and statistically efficient proxies thereof via regularization. While OT distances enable a natural comparison between distributions on the same space, comparing datasets of different types (e.g., text and images) requires defining an ad hoc cost function which may not capture a meaningful correspondence between data points.
In this talk, I will survey the current statistical and computational landscape for Gromov-Wasserstein (GW) distances, a framework which enables comparing abstract metric measure spaces based on their intrinsic metric structure and, as such, have seen widespread use in applications including comparing datasets of different types. I will present the first limit laws obtained for empirical GW distances, both with and without regularization, and describe consistent resampling schemes. Additionally, I will introduce the first algorithms for computing regularized GW distance subject to formal convergence guarantees. I will conclude by highlighting a number of open questions and future directions in the study of GW distances.
Joint work with Ziv Goldfeld and Kengo Kato
|
![]() |
*30 January (Thursday) at 10:45 – 11:45 am in Hill 705 |
Joint with Math and Data SeminarSangmin Park, Carnegie Mellon University“PDEs, data science, and optimal transport.” |
AbstractIn this talk I will present two works lying at the interface between PDEs, data science, and optimal transport. The first part concerns the dissipative Hamiltonian structure of the Vlasov-Fokker-Planck equation (VFP). VFP describes the evolution of the probability density of the position and velocity of particles under the influence of external confinement, interaction, friction, and stochastic force. It is well-known that this equation can be formally seen as a dissipative Hamiltonian system in the Wasserstein space of probability measures. Moreover, this geometric structure has possible connections to the conjectured optimal convergence rates of underdamped Langevin Monte Carlo (ULMC), a sampling algorithm known to empirically outperform the (standard) Langevin Monte Carlo. This talk will focus on a time-discrete variational scheme for VFP which we introduce to more rigorously understand the geometric structure. The second part, based on a joint work with Dejan Slepcev, concerns the sliced Wasserstein distance. The sliced Wasserstein metric compares probability measures by taking averages of the Wasserstein distances between projections of the measures to lines. The distance has found a range of applications in statistics and machine learning, as it is easier to approximate and compute in high dimensions than the (classical) Wasserstein distance. We will focus on our characterization of the disparate behavior of the sliced Wasserstein distance near absolutely continuous and discrete measures. We will explain the connection between this characterization and the maximum mean discrepancies (MMDs) which helps understand the behavior of the distance in high dimensions. If time permits, we will also discuss a refined result on the sample complexity of the metric and the instability of sliced Wasserstein gradient flows. |
![]() |
*6 February (Thursday) at 2 – 3 pm in Hill 260 |
Tom Hagstrom, Southern Methodist University“Local Algorithms for Nonlocal Problems in Wave Theory.” |
AbstractAlthough many mathematical models in wave theory lead to hyperbolic initial-boundary value problems which are inherently local due to the finite domain-of-dependence of the solution at any point on past values, there are also examples where nonlocal operators, in particular space-time integral operators, arise. A primary example is the radiation boundary condition needed to truncate an unbounded domain to a finite one to enable numerical solutions, as well as closely-related operators for unidirectional propagation. In this work we show how to leverage results from rational function approximation theory to construct spectrally-convergent local algorithms for evaluating such operators. For an important class of problems – systems equivalent to the scalar wave equation, such as acoustics or Maxwell’s equations in homogeneous, isotropic media, we will explain the construction, analysis, and implementation of our complete radiation boundary conditions, which are in a certain sense optimal. We will also discuss other applications of these approximation methods, as well as the fundamental barriers to extending the successful methods to other systems. |
![]() |
19 February |
Pierre Amenoagbadji, Columbia University“Wave propagation in junctions of 2D periodic half-spaces.” |
AbstractThis talk focuses on the analysis and numerical solution of PDE models for time-harmonic wave propagation in junctions of 2D periodic half-spaces. Most existing works and numerical methods rely on the critical assumption that the medium is translation-invariant along the interface. Here, we investigate situations where this assumption no longer holds. Our analysis relies on the key observation that the medium has a quasiperiodic structure along the interface, meaning it can be viewed as a slice of a 3D translation-invariant medium. This enables us to seek solutions to our PDE as restrictions of solutions to an augmented non-elliptic PDE set in 3D, where periodicity along the interface is recovered.
In the first part of this talk, based on joint work with Sonia Fliss and Patrick Joly, I will show how this lifting approach can be used to numerically solve the time-harmonic wave equation in junctions of 2D periodic half-spaces. I will then discuss ongoing work with Michael Weinstein, where the lifting approach is applied to the study of edge states in honeycomb structures perturbed along irrational edges.
|
![]() |
12 March |
Mahadevan Ganesh, Colorado School of Mines“Heterogeneous Media and Unbounded Region Models: Wave Simulations in Frequency- and Time-domain.“ |
AbstractEfficient simulation of waves induced by heterogeneous media for measurable quantities of interest (QOI) in unbounded free space is essential for many applications. A robust mathematical model of the underlying time-dependent physical processes is crucial. This includes computationally appropriate reformulations aimed at designing high-order computational methods for QOI simulations. The development of related algorithms and analyses depends on established continuous mathematical equations, which can be framed in either the time-domain (TD) or frequency-domain (FD). Some literature suggests that the continuous FD model for certain configurations is merely a mathematical construct, while the TD model represents more tangible physical phenomena. In this talk, we will discuss our recent contributions to developing high-order computational models for wave propagation in both the TD and FD. We will also share our progress on developing computational models that incorporate all the key elements mentioned in the title of the talk.
|
![]() |
*14 March (Friday) at 3 – 4 pm in Hill 705 |
Joint with Math and Data SeminarAmit Patel, Colorado State University“Inferring Dynamics with Generalized Persistence Diagrams.” |
AbstractSuppose we are given a sample of a discrete dynamical system φ: M → M. Can we infer φ from this sample? Classical persistent homology can be used to analyze the sample and infer the homology of M. In this talk, we employ the machinery of Persistent Local Systems (Patel and MacPherson) to not only recover the persistent homology of the sample but also extract persistent information about the dynamics, all summarized in what we call the Generalized Persistence Diagram (Patel). Moreover, generalized persistence diagrams satisfy Bottleneck Stability, just like classical persistence. This implies that, under suitable assumptions on M and φ, the persistent dynamical information is stable with respect to perturbations of the sample.
|
![]() |
26 March |
Marian Gidea, Yeshiva University“Energy growth in Hamiltonian systems with small dissipation: with applications to energy harvesting.“ |
Abstract
|
![]() |
2 April |
Huan Liu, California Institute of Technology“Curved tile origami: design and application.” |
AbstractOrigami, the art of folding paper into intricate shapes, has growing practical applications across fields such as architectural design, therapeutics, deployable space structures, antenna design, and soft robotics. One promising yet largely unexplored area is curved tile origami, which can store elastic energy, offering opportunities to develop next-generation functional materials, structures, and actuators. In this talk I will present a general theory of curved origami and systematic design methods for constructing large-scale, complex curved origami structures. Additionally, I will present methods to accurately calculate the stored elastic energy and the folding motions of curved origami, and I will illustrate my theoretical results by presenting some complex structures I have folded. This theory has inspired applications of curved origami in fields ranging from medical devices to a vertical-axis wind turbine, to architected materials. |
![]() |
9 April |
Amir Sagiv, New Jersey Institute of Technology“TBA.” |
AbstractTBA. |
![]() |
16 April |
Benedict Borer, Rutgers University (Marine and Coastal Sciences)“Is intermittent fasting a hallmark of the ocean’s microbiome?” |
AbstractMicroorganisms in the ocean are exposed to a patchy resource landscape, often provided in the form of nutrient-rich sinking particles. Yet, the most abundant microorganisms in the ocean are non-motile bacteria that rely on a diffusive supply of nutrients and cannot forage for particles via swimming. How they sustain their nutrient demand remains a conundrum. We developed a mathematical framework based on a spatial Poisson point process and analytical description of particle plumes to predict the frequency with which bacteria encounter nutrient-rich plumes from sinking particles across the global oceans. We find that due to the abundance of marine particles, non-motile bacteria encounter plumes from sinking particles frequently (~15 hr) that may readily satisfy their nutrient demand. Encounter rates are sensitive to regional differences in primary productivity and particle remineralization dynamics. Our mechanistic framework challenges the common view that non-motile bacteria in the ocean rely on vanishingly low background concentrations of dissolved nutrients and instead suggests that they can thrive on intermittent nutrient bursts from passing particles and their plumes. |
![]() |
23 April |
Brittany Gelb, Rutgers University“TBA.” |
AbstractTBA. |