Sinkhorn Divergence of Topological Signature Estimates for Time Series Classification

Colin Stephen

doi:10.1109/ICMLA.2018.00113

Sinkhorn Divergence of Topological Signature Estimates for Time Series Classification

Colin Stephen

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding › peer-review

72 Downloads (Pure)

Abstract

Distinguishing between classes of time series sampled from dynamic systems is a common challenge in systems and control engineering, for example in the context of health monitoring, fault detection, and quality control. The challenge is increased when no underlying model of a system is known, measurement noise is present, and long signals need to be interpreted. In this paper we address these issues with a new non parametric classifier based on topological signatures. Our model learns classes as weighted kernel density estimates (KDEs) over persistent homology diagrams and predicts new trajectory labels using Sinkhorn divergences on the space of diagram KDEs to quantify proximity. We show that this approach accurately discriminates between states of chaotic systems that are close in parameter space, and its performance is robust to noise.

Original language	English
Title of host publication	Proceedings - 17th IEEE International Conference on Machine Learning and Applications, ICMLA 2018
Editors	M. Arif Wani, Moamar Sayed-Mouchaweh, Edwin Lughofer, Joao Gama, Mehmed Kantardzic
Publisher	IEEE
Pages	714-721
Number of pages	8
ISBN (Electronic)	978-1-5386-6805-4
ISBN (Print)	978-1-5386-6806-1
DOIs	https://doi.org/10.1109/ICMLA.2018.00113
Publication status	Published - 17 Jan 2019
Event	17th IEEE International Conference on Machine Learning and Applications, ICMLA 2018 - Orlando, United States Duration: 17 Dec 2018 → 20 Dec 2018 Conference number: 17 https://www.icmla-conference.org/icmla18/

Conference

Conference	17th IEEE International Conference on Machine Learning and Applications, ICMLA 2018
Abbreviated title	ICMLA 2018
Country/Territory	United States
City	Orlando
Period	17/12/18 → 20/12/18
Internet address	https://www.icmla-conference.org/icmla18/

Bibliographical note

© 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

Copyright © and Moral Rights are retained by the author(s) and/ or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This item cannot be reproduced or quoted extensively from without first obtaining permission in writing from the copyright holder(s). The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the copyright holders.

Keywords

Time series analysis
Measurement
Entropy
Pipelines
Histograms
Machine learning
Chaos
Time series
Dynamical system
Sinkhorn
Persistence diagram
TDA
Topological data analysis

ASJC Scopus subject areas

Artificial Intelligence
Decision Sciences (miscellaneous)
Safety, Risk, Reliability and Quality
Signal Processing
Computer Vision and Pattern Recognition
Computer Networks and Communications
Computer Science Applications

Access to Document

10.1109/ICMLA.2018.00113

Post-PrintAccepted author manuscript, 1.4 MB

Cite this

Sinkhorn Divergence of Topological Signature Estimates for Time Series Classification. / Stephen, Colin.
Proceedings - 17th IEEE International Conference on Machine Learning and Applications, ICMLA 2018. ed. / M. Arif Wani; Moamar Sayed-Mouchaweh; Edwin Lughofer; Joao Gama; Mehmed Kantardzic. IEEE, 2019. p. 714-721 8614138.

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding › peer-review

Stephen, C 2019, Sinkhorn Divergence of Topological Signature Estimates for Time Series Classification. in MA Wani, M Sayed-Mouchaweh, E Lughofer, J Gama & M Kantardzic (eds), Proceedings - 17th IEEE International Conference on Machine Learning and Applications, ICMLA 2018., 8614138, IEEE, pp. 714-721, 17th IEEE International Conference on Machine Learning and Applications, ICMLA 2018, Orlando, United States, 17/12/18. https://doi.org/10.1109/ICMLA.2018.00113

@inproceedings{5729f9402b8243d5a23138d03d430137,

title = "Sinkhorn Divergence of Topological Signature Estimates for Time Series Classification",

abstract = "Distinguishing between classes of time series sampled from dynamic systems is a common challenge in systems and control engineering, for example in the context of health monitoring, fault detection, and quality control. The challenge is increased when no underlying model of a system is known, measurement noise is present, and long signals need to be interpreted. In this paper we address these issues with a new non parametric classifier based on topological signatures. Our model learns classes as weighted kernel density estimates (KDEs) over persistent homology diagrams and predicts new trajectory labels using Sinkhorn divergences on the space of diagram KDEs to quantify proximity. We show that this approach accurately discriminates between states of chaotic systems that are close in parameter space, and its performance is robust to noise.",

keywords = "Time series analysis, Measurement, Entropy, Pipelines, Histograms, Machine learning, Chaos, Time series, Dynamical system, Sinkhorn, Persistence diagram, TDA, Topological data analysis",

author = "Colin Stephen",

note = "{\textcopyright} 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. Copyright {\textcopyright} and Moral Rights are retained by the author(s) and/ or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This item cannot be reproduced or quoted extensively from without first obtaining permission in writing from the copyright holder(s). The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the copyright holders. ; 17th IEEE International Conference on Machine Learning and Applications, ICMLA 2018, ICMLA 2018 ; Conference date: 17-12-2018 Through 20-12-2018",

year = "2019",

month = jan,

day = "17",

doi = "10.1109/ICMLA.2018.00113",

language = "English",

isbn = "978-1-5386-6806-1",

pages = "714--721",

editor = "Wani, {M. Arif} and Moamar Sayed-Mouchaweh and Edwin Lughofer and Joao Gama and Mehmed Kantardzic",

booktitle = "Proceedings - 17th IEEE International Conference on Machine Learning and Applications, ICMLA 2018",

publisher = "IEEE",

address = "United States",

url = "https://www.icmla-conference.org/icmla18/",

}

TY - GEN

T1 - Sinkhorn Divergence of Topological Signature Estimates for Time Series Classification

AU - Stephen, Colin

N1 - Conference code: 17

PY - 2019/1/17

Y1 - 2019/1/17

N2 - Distinguishing between classes of time series sampled from dynamic systems is a common challenge in systems and control engineering, for example in the context of health monitoring, fault detection, and quality control. The challenge is increased when no underlying model of a system is known, measurement noise is present, and long signals need to be interpreted. In this paper we address these issues with a new non parametric classifier based on topological signatures. Our model learns classes as weighted kernel density estimates (KDEs) over persistent homology diagrams and predicts new trajectory labels using Sinkhorn divergences on the space of diagram KDEs to quantify proximity. We show that this approach accurately discriminates between states of chaotic systems that are close in parameter space, and its performance is robust to noise.

AB - Distinguishing between classes of time series sampled from dynamic systems is a common challenge in systems and control engineering, for example in the context of health monitoring, fault detection, and quality control. The challenge is increased when no underlying model of a system is known, measurement noise is present, and long signals need to be interpreted. In this paper we address these issues with a new non parametric classifier based on topological signatures. Our model learns classes as weighted kernel density estimates (KDEs) over persistent homology diagrams and predicts new trajectory labels using Sinkhorn divergences on the space of diagram KDEs to quantify proximity. We show that this approach accurately discriminates between states of chaotic systems that are close in parameter space, and its performance is robust to noise.

KW - Time series analysis

KW - Measurement

KW - Entropy

KW - Pipelines

KW - Histograms

KW - Machine learning

KW - Chaos

KW - Time series

KW - Dynamical system

KW - Sinkhorn

KW - Persistence diagram

KW - TDA

KW - Topological data analysis

UR - http://www.scopus.com/inward/record.url?scp=85062243499&partnerID=8YFLogxK

U2 - 10.1109/ICMLA.2018.00113

DO - 10.1109/ICMLA.2018.00113

M3 - Conference proceeding

SN - 978-1-5386-6806-1

SP - 714

EP - 721

BT - Proceedings - 17th IEEE International Conference on Machine Learning and Applications, ICMLA 2018

A2 - Wani, M. Arif

A2 - Sayed-Mouchaweh, Moamar

A2 - Lughofer, Edwin

A2 - Gama, Joao

A2 - Kantardzic, Mehmed

PB - IEEE

T2 - 17th IEEE International Conference on Machine Learning and Applications, ICMLA 2018

Y2 - 17 December 2018 through 20 December 2018

ER -

Sinkhorn Divergence of Topological Signature Estimates for Time Series Classification

Abstract

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this