September 22 talk will be on the use of neural networks:

Data Driven Governing Equations Recovery with Deep Neural Networks

Dongbin Xiu
Department of Mathematics
The Ohio State University

We present effective numerical algorithms for recovering unknown governing equations from measurement data. Upon recasting the problem into a function approximation problem, we discuss the importance of using a large number of short bursts of trajectory data, rather than using data from a small number of long trajectories. Several recovery strategies using deep neural networks (DNNs) are presented. We demonstrate that residual network (ResNet) is particularly suitable for equation discovery, as it can produce exact time integrator for numerical prediction. We then present a set of applications of the DNN learning of unknown dynamical systems, which may contain random parameters or missing variables, as well as learning of unknown partial differential equations. The numerical examples demonstrate that DNNs can be a highly effective tool for data driven physics recovery.

From: Tobias von Petersdorff tvp@math.umd.edu

Subject: Sayas Numerics Seminar, ONLINE, Sep 2020

We are pleased to announce SAYAS NUMERICS SEMINAR, a new online seminar on computational mathematics jointly organized by

– University of Maryland, College Park, MD

– University of Maryland, Baltimore County, MD

– George Mason University, Fairfax, VA

– University of Delaware, Newark, DE

The seminar is named in honor of Prof. Francisco-Javier Sayas (1968-2019). In addition to his many achievements he played a key role in establishing and

running the annual meeting DelMar Numerics Day, and was dedicated to its mission of giving younger mathematicians an opportunity to share their work

with the community.

The seminar will meet via Zoom on Tuesdays at 3:30pm (Eastern Time) unless noted otherwise. Please go to https://sayasseminar.math.umd.edu to REGISTER,

SUBMIT A CONTRIBUTED TALK, and to see the INVITED TALKS.

Researchers from anywhere in the world are welcome to attend, but REGISTRATION IS REQUIRED.

The organizing committee: Harbir Antil, Andrei Draganescu, Ricardo H. Nochetto, Petr Plechac, Tobias von Petersdorff