Applied Dynamics Seminar: Brian Hunt and Zhixin Liu| University of Maryland

Thursday, March 9, 2017
12:30 p.m.
ERF 1207
Taylor Prendergast
301 405 4951
tprender@umd.edu

Speakers: Brian Hunt, University of Maryland| Department of Mathematics & 

Zhixin LuUniversity of Maryland| Institute for Reasearch in Electonics and Applied Physics

Title: Toward a theory of reservoir computing prediction

Abstract: We consider the problem of predicting a chaotic time series from a
system whose equations of motion are unknown.  We use a
machine-learning technique called reservoir computing, which we find
is often able to learn the dynamics of the system that generated the
time series, in the following sense.  In addition to making accurate
short-term predictions, the reservoir predictor can generate a
long-term "climate" forecast that stays close to the attractor of the
actual system.  We give examples, and we discuss a preliminary theory
relating reservoir predictor performance to Lyapunov exponents and
generalized synchronization in an associated dynamical system.

Audience: Graduate  Undergraduate  Faculty 

remind we with google calendar

 

March 2024

SU MO TU WE TH FR SA
25 26 27 28 29 1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30
31 1 2 3 4 5 6
Submit an Event