Difference between revisions of "SIAM Student Chapter Seminar"
(No difference)

Latest revision as of 12:47, 20 February 2020
 When: Every other Friday at 1:30 pm
 Where: B333 Van Vleck Hall
 Organizers: Xiao Shen
 Faculty advisers: JeanLuc Thiffeault, Steve Wright
 To join the SIAM Chapter mailing list: email [joinsiamchapter@lists.wisc.edu].
Spring 2020
date  speaker  title 

Jan 31  Lorenzo Najt (Math)  Ensemble methods for measuring gerrymandering: Algorithmic problems and inferential challenges 
Feb 14  Polly Yu (Math)  Algebra, Dynamics, and Chemistry with Delay Differential Equations 
Feb 21  Gage Bonner (Physics)  Growth of historydependent random sequences 
Abstracts
Jan 31, Lorenzo Najt (Math)
Ensemble methods for measuring gerrymandering: Algorithmic problems and inferential challenges
We will review some recent work regarding measuring gerrymandering by sampling from the space of maps, including two methods used in a recent amicus brief to the supreme court. This discussion will highlight some of the computational challenges of this approach, including some complexitytheory lower bounds and bottlenecks in Markov chains. We will examine the robustness of these statistical methods through their connection to phase transitions in the selfavoiding walk model, as well as their dependence on artifacts of discretization. This talk is largely based on https://arxiv.org/abs/1908.08881
Feb 14, Polly Yu (Math)
Algebra, Dynamics, and Chemistry with Delay Differential Equations
Delay differential equations (DDEs) can exhibit more complicated behavior than their ODE counterparts. What is stable in the ODE setting could exhibit oscillation in DDE. Where do delay equations show up anyway? In this talk, we’ll introduce DDEs, and how (sortof)linear algebra gives information about the stability of DDEs.
Feb 21, Gage Bonner (Physics)
Growth of historydependent random sequences
Unlike discrete Markov chains, historydependent random sequences are sequences of random variables whose "next" term depends on all others seen previously. For this reason, they can be difficult to analyze. I will discuss some simple and fun cases where the longterm behavior of the sequence can be computed explicitly in expectation.