First PSU workshop on learning nonparametric differential equations from data
When: Tuesday, September 13th, 10 am to 16:00 pm.
Where: Portland State University, FMH-4-462 MTH LargeConferenceRoom
Background: Consider data arising from a differential equation, ordinary, partial, or stochastic. This data is univariate, multivariate, or functional and observed with noise at a small number of time points. How to learn these differential equations? Are they even learnable? If yes, what guarantee can be provided? We are interested in the nonparametric case when there is no known analytical form for the differential equations.
Sponsors: Google Research Award and the Research Training Group in Computation and Data-Enabled Science https://sites.google.com/pdx.edu/rtg-in-cades
Program:
10-11 Bruno Jedynak: On learning nonparametric differential equations from data and application to Alzheimer's disease data
11-12 Kamel Lahouel: Chaining, Dudley’s theorem and application to the consistency of estimated ODE trajectories
12-12:30 lunch (pizzas)
12:30-13 Michael Wells: Learning Ordinary Differential Equations from Data using the penalty method of optimization
13-13:30 Victor Rielly: Learning Ordinary Differential Equations, Some Established Techniques
13:30-14 Ethan Lew: Dynamical System Reachability, Conformance, and Control via Koopman Linearization
14:14:30 Vicky Haney: Speaker Conversion: An approach using a Partial Differential Equation
14:30-15:30 Round table discussion with the attendees, including Dr. Ehsan Variani, from Google Inc.