Visit CST

Mid-Atlantic Numerical Analysis Day

A conference on numerical analysis and scientific computing for graduate students and postdocs in the Mid-Atlantic region.

Friday, 15 November 2024

The Conference

This one-day meeting will start at 10am to allow same-day travel.
It will be held in Room 617 Wachman Hall, Temple University, 1805 North Broad Street, just north of Montgomery Avenue.
It is an opportunity for graduate students and postdocs to present their research, and to meet other researchers.
There will be contributed talks and a poster session.
There is no registration fee, and no support for travel. Lunch will be provided.
We ask every participant to please register in advance, even if they are not planning to give a talk.

Keynote Speaker

Mark Embree, Virginia Tech

Spectral Computations for Quasicrystals

In 2011, Dan Shechtman was awarded the Nobel Prize in Chemistry for his discovery of quasicrystals, novel materials with properties somewhere between the regularity of crystals and the disorder of random structures. In parallel with this scientific breakthrough, mathematicians have developed models of aperiodic order, such as Fibonacci substitutions and Penrose tilings. The self-adjoint linear operators based on such models often exhibit intriguing spectral structure. For example, the spectrum of the Fibonacci Hamiltonian is a zero-measure Cantor set. How can one approximate such fine structure using eigenvalue computations with finite dimensional matrices? Can one estimate quantities like fractal dimension of the spectrum and the integrated density of states? We will describe several aperiodic models, discuss their spectral properties, illustrate some numerical tools we can use to approach these problems, and show results from our calculations. This talk describes collaborations with Matt Colbrook, David Damanik, Jake Fillman, Anton Gorodetski, and May Mei.

Registration and/or Abstract Submission

If you would like to participate (in any form), please register using the online registration form.
Deadline for the submission of talks: October 18, 2024.

Conference Poster and Booklet

Conference Poster

Click to download a PDF.

Conference Booklet

Click to download the conference booklet.

Schedule

9:15-9:50	Registration and breakfast (provided)
9:50-10:00	Opening remarks
10:00-11:00	Presentations (Algorithms)
11:00-11:15	Coffee Break
11:15-12:15	Presentations (Data Science)
12:15-1:30	Posters and lunch (provided)
1:30-2:30	Keynote lecture (Mark Embree)
2:30-2:45	Coffee break
2:45-3:45	Presentations (Numerical Differential Equations)
3:45-4:00	Coffee break
4:00-5:00	Presentations (Numerical Linear Algebra)
5:00-5:10	Closing remarks
6:00-8:00	Group dinner (attendance optional)

Speakers

Algorithms
Chenyang	Cao	Purdue University	Kernel matrix approximations by sums of exponentials and stability of fast structured transforms
Chenyang	Islam	Lehigh University	Acceleration of approximate maps for matrices arising in discretized PDEs
Hai	Zhu	Flatiron Institute	Recursive reduction quadrature for the evaluation of Laplace layer potentials in three dimensions
Data Science
Mikhail	Lepilov	Emory University	Estimating kernel matrix eigenvalues
Sonia Marlena	Reilly	New York University	Bayesian inversion of PDE-based problems using integrated nested Laplace approximations
Annan	Yu	Cornell University	Training an LTI system without an objective: A numerical analyst's perspectives on state-space models
Numerical Differential Equations
Hamad	El Kahza	University of Delaware	Adaptive-rank implicit time integrator for advection-diffusion transport equations with inhomogeneous coefficients
Zachary	Miksis	Temple University	A new Fick-Jacobs derivation with applications to computational branched diffusion networks
Mansur	Shakipov	University of Maryland, College Park	Inf-sup stability of parabolic TraceFEM
Numerical Linear Algebra
Noah	Amsel	New York University	Fixed-sparsity matrix approximation from matrix-vector products
Robin John	Armstrong	Cornell University	“Collect, commit, expand”: A strategy for faster CPQR-based column selection on short, wide matrices
Tyler	Chen	New York University	Near-optimal hierarchical matrix approximation from matrix-vector products

Posters

Pierre	Amenoagbadji	APAM Columbia	Wave propagation in junctions of periodic half-spaces
Caroline	Huber	New York University	Preconditioning without a preconditioner: Faster ridge-regression and Gaussian sampling with randomized block Krylov methods
Sean T.	McQuade	Temple University	Modeling lipid metabolism for multiple classes of virtual patient
Afrina Asad	Meghla	Temple University	Modeling and simulation of calcium influx through NMDA receptors and its activation of TRPM4 channels
Madison	Shoraka	Temple University	Modeling growth of pCF10-induced complex structures in Enterococcus faecalis biofilm under erythromycin treatment
Jovan	Zigic	McMaster University	Stochastic variants of gradient descent for PDE-constrained optimization on Riemannian manifolds