Mathematical Sciences

School of Natural Sciences & Mathematics

Computational Science Seminar AY 15-16

Date Speaker Affiliation Title Abstract
Apr 21 Chris Jones UNC Where does Math come into Climate Science?

Climate science currently revolves around massive computational models that run at just a few centers around the world. These are mathematical objects in that they offer mathematical replicas of the world we inhabit. Nevertheless it is hard to see how the art of mathematics fits in with computational modeling at this scale. Why we mathematicians are badly needed will be discussed in this lecture, as well as how we can contribute to the climate science enterprise. I will illustrate this with two recent pieces of work that I have been involved in.

A buffet lunch will follow in the 2nd floor Atrium at 1pm.

Sponsored by the UTD-SMU SIAM Student Chapter

Apr 14 Prabir Daripa Mathematics, Texas A&M Modeling and simulation of multiphase multicomponent porous media flows
in the context of chemical enhanced oil recovery
March 31st Faruck Morcos Biological Sciences, Utd Inferring direct couplings to unveil coevolutionary signals in
protein 3D structure, interactions and recognition in signalling networks
Mar 10 Frederico Furtado Department of Mathematics

University of Wyoming
The displacement problem for immiscible three-phase flow in green reservoirs We discuss the displacement problem for immiscible three-phase flow described by a system of two conservation laws with fluxes originating from Corey-type permeabilities. A mixture of water, gas, and oil is injected into a porous medium initially saturated with oil, which is partially displaced. The solutions of the resulting Riemann problems generically belong to two classes. Each class of solutions occurs for injection states in one of two regions of the state space, separated by a curve of states for which the interstitial velocities of water and gas are equal. This is a separatrix curve because on one side water appears at breakthrough, while gas appears for injection states on the other side. In other words, the behavior near breakthrough is flow of oil and of the dominant phase, either water or gas; the non dominant phase is left behind. This description of the solutions is valid for any values of phase viscosities. The inevitable loss of strict hyperbolicity for such flow models seems to be the cause of this solution structure.
Mar 3 Hejun Zhu Geophysics, Utd Cycle Skipping and Uncertainty Estimation in Full Waveform Inversion Abstract
Feb 25 Kathryn Hedrick SMU Attractor Neural Networks and the Brain’s Representation of Space An attractor network is a network of nodes (e.g., neurons) whose state vector converges in time to a stable pattern. Attractor network theory has been highly influential in our understanding of how the brain remembers. It naturally accounts for many neural processes, such as full recall from partial cues and persistent neural activity given a transient stimulus. In this talk, I will briefly review how attractor network theory has been used to model the brain’s representation of space. I will then describe a new type of attractor neural network called a megamap that is capable of stably representing a large, naturalistic environment. Results from simulations and analysis of the megamap point to a general computational strategy by which the brain may enjoy the stability of attractor dynamics without sacrificing the flexibility needed to represent a complex, changing world.
Feb 18 Albert Montillo UTSW Statistical Learning for Personalized Diagnostics and Prognostics Abstract
Jan 28 Pankaj Choudhary and John Zweck SIAM Student Chapter Tips for Studying for Qualifying Exams