Learning Dynamical Systems with Side Information (short version)

Amir Ali Ahmadi; Bachir El Khadir

Learning Dynamical Systems with Side Information (short version)

Research output: Contribution to journal › Conference article › peer-review

Abstract

We present a mathematical formalism and a computational framework for the problem of learning a dynamical system from noisy observations of a few trajectories and subject to side information (e.g., physical laws or contextual knowledge). We identify six classes of side information which can be imposed by semidefinite programming and that arise naturally in many applications. We demonstrate their value on two examples from epidemiology and physics. Some density results on polynomial dynamical systems that either exactly or approximately satisfy side information are also presented.

Original language	English (US)
Pages (from-to)	718-727
Number of pages	10
Journal	Proceedings of Machine Learning Research
Volume	120
State	Published - 2020
Event	2nd Annual Conference on Learning for Dynamics and Control, L4DC 2020 - Berkeley, United States Duration: Jun 10 2020 → Jun 11 2020

All Science Journal Classification (ASJC) codes

Artificial Intelligence
Software
Control and Systems Engineering
Statistics and Probability

Keywords

Dynamical Systems
Learning
Semidefinite Programming
Sum of Squares Optimization

Cite this

@article{43e576b74a464d4c8caddd340c861eb3,

title = "Learning Dynamical Systems with Side Information (short version)",

abstract = "We present a mathematical formalism and a computational framework for the problem of learning a dynamical system from noisy observations of a few trajectories and subject to side information (e.g., physical laws or contextual knowledge). We identify six classes of side information which can be imposed by semidefinite programming and that arise naturally in many applications. We demonstrate their value on two examples from epidemiology and physics. Some density results on polynomial dynamical systems that either exactly or approximately satisfy side information are also presented.",

keywords = "Dynamical Systems, Learning, Semidefinite Programming, Sum of Squares Optimization",

author = "Ahmadi, {Amir Ali} and Khadir, {Bachir El}",

note = "Publisher Copyright: {\textcopyright} 2020 A. Ahmadi & B. El Khadir.; 2nd Annual Conference on Learning for Dynamics and Control, L4DC 2020 ; Conference date: 10-06-2020 Through 11-06-2020",

year = "2020",

language = "English (US)",

volume = "120",

pages = "718--727",

journal = "Proceedings of Machine Learning Research",

issn = "2640-3498",

}

TY - JOUR

T1 - Learning Dynamical Systems with Side Information (short version)

AU - Ahmadi, Amir Ali

AU - Khadir, Bachir El

PY - 2020

Y1 - 2020

N2 - We present a mathematical formalism and a computational framework for the problem of learning a dynamical system from noisy observations of a few trajectories and subject to side information (e.g., physical laws or contextual knowledge). We identify six classes of side information which can be imposed by semidefinite programming and that arise naturally in many applications. We demonstrate their value on two examples from epidemiology and physics. Some density results on polynomial dynamical systems that either exactly or approximately satisfy side information are also presented.

AB - We present a mathematical formalism and a computational framework for the problem of learning a dynamical system from noisy observations of a few trajectories and subject to side information (e.g., physical laws or contextual knowledge). We identify six classes of side information which can be imposed by semidefinite programming and that arise naturally in many applications. We demonstrate their value on two examples from epidemiology and physics. Some density results on polynomial dynamical systems that either exactly or approximately satisfy side information are also presented.

KW - Dynamical Systems

KW - Learning

KW - Semidefinite Programming

KW - Sum of Squares Optimization

UR - http://www.scopus.com/inward/record.url?scp=85161090466&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85161090466&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:85161090466

SN - 2640-3498

VL - 120

SP - 718

EP - 727

JO - Proceedings of Machine Learning Research

JF - Proceedings of Machine Learning Research

T2 - 2nd Annual Conference on Learning for Dynamics and Control, L4DC 2020

Y2 - 10 June 2020 through 11 June 2020

ER -

Learning Dynamical Systems with Side Information (short version)

Abstract

All Science Journal Classification (ASJC) codes

Keywords

Other files and links

Fingerprint

Cite this