Estimating Vector Fields from Noisy Time Series

12/06/2020
by   Harish S. Bhat, et al.
11

While there has been a surge of recent interest in learning differential equation models from time series, methods in this area typically cannot cope with highly noisy data. We break this problem into two parts: (i) approximating the unknown vector field (or right-hand side) of the differential equation, and (ii) dealing with noise. To deal with (i), we describe a neural network architecture consisting of tensor products of one-dimensional neural shape functions. For (ii), we propose an alternating minimization scheme that switches between vector field training and filtering steps, together with multiple trajectories of training data. We find that the neural shape function architecture retains the approximation properties of dense neural networks, enables effective computation of vector field error, and allows for graphical interpretability, all for data/systems in any finite dimension d. We also study the combination of either our neural shape function method or existing differential equation learning methods with alternating minimization and multiple trajectories. We find that retrofitting any learning method in this way boosts the method's robustness to noise. While in their raw form the methods struggle with 1 accurate vector fields from data with 10

READ FULL TEXT

page 1

page 2

page 3

page 4

research
07/09/2020

A blueprint for building efficient Neural Network Differential Equation Solvers

Neural Networks are well known to have universal approximation propertie...
research
05/23/2019

Neural ODEs with stochastic vector field mixtures

It was recently shown that neural ordinary differential equation models ...
research
09/14/2021

Multiple shooting with neural differential equations

Neural differential equations have recently emerged as a flexible data-d...
research
09/23/2021

Learning Dynamics from Noisy Measurements using Deep Learning with a Runge-Kutta Constraint

Measurement noise is an integral part while collecting data of a physica...
research
05/19/2021

Efficient approximation of high-dimensional exponentials by tensornetworks

In this work a general approach to compute a compressed representation o...
research
10/05/2021

Deep Learning for the Approximation of a Shape Functional

Artificial Neuronal Networks are models widely used for many scientific ...
research
05/26/2021

SG-PALM: a Fast Physically Interpretable Tensor Graphical Model

We propose a new graphical model inference procedure, called SG-PALM, fo...

Please sign up or login with your details

Forgot password? Click here to reset