Machine-learning-based Brokers for Real-time Classification of the LSST Alert Stream

Gautham Narayan, Tayeb Zaidi, Monika D. Soraisam, Zhe Wang, Michelle Lochner, Thomas Matheson, Abhijit Saha, Shuo Yang, Zhenge Zhao, John D Kececioglu, Carlos Eduardo Scheidegger, Richard Thomas Snodgrass, Tim Axelrod, Tim Jenness, Robert S Maier, Stephen T. Ridgway, Robert L. Seaman, Eric Michael Evans, Navdeep Singh, Clark TaylorJackson Toeniskoetter, Eric Welch, Songzhe Zhu

Research output: Contribution to journalArticle

13 Citations (Scopus)

Abstract

The unprecedented volume and rate of transient events that will be discovered by the Large Synoptic Survey Telescope (LSST) demand that the astronomical community update its follow-up paradigm. Alert-brokers- A utomated software system to sift through, characterize, annotate, and prioritize events for follow-up-will be critical tools for managing alert streams in the LSST era. The Arizona-NOAO Temporal Analysis and Response to Events System (ANTARES) is one such broker. In this work, we develop a machine learning pipeline to characterize and classify variable and transient sources only using the available multiband optical photometry. We describe three illustrative stages of the pipeline, serving the three goals of early, intermediate, and retrospective classification of alerts. The first takes the form of variable versus transient categorization, the second a multiclass typing of the combined variable and transient data set, and the third a purity-driven subtyping of a transient class. Although several similar algorithms have proven themselves in simulations, we validate their performance on real observations for the first time. We quantitatively evaluate our pipeline on sparse, unevenly sampled, heteroskedastic data from various existing observational campaigns, and demonstrate very competitive classification performance. We describe our progress toward adapting the pipeline developed in this work into a real-time broker working on live alert streams from time-domain surveys.

Original languageEnglish (US)
Article number9
JournalAstrophysical Journal, Supplement Series
Volume236
Issue number1
DOIs
StatePublished - May 1 2018

Fingerprint

machine learning
telescopes
temporal analysis
photometry
purity
computer programs
software
simulation

Keywords

  • data analysis-methods
  • general-supernovae
  • general-surveys-virtual observatory tools
  • methods
  • statistical-stars
  • variables

ASJC Scopus subject areas

  • Astronomy and Astrophysics
  • Space and Planetary Science

Cite this

Narayan, G., Zaidi, T., Soraisam, M. D., Wang, Z., Lochner, M., Matheson, T., ... Zhu, S. (2018). Machine-learning-based Brokers for Real-time Classification of the LSST Alert Stream. Astrophysical Journal, Supplement Series, 236(1), [9]. https://doi.org/10.3847/1538-4365/aab781

Machine-learning-based Brokers for Real-time Classification of the LSST Alert Stream. / Narayan, Gautham; Zaidi, Tayeb; Soraisam, Monika D.; Wang, Zhe; Lochner, Michelle; Matheson, Thomas; Saha, Abhijit; Yang, Shuo; Zhao, Zhenge; Kececioglu, John D; Scheidegger, Carlos Eduardo; Snodgrass, Richard Thomas; Axelrod, Tim; Jenness, Tim; Maier, Robert S; Ridgway, Stephen T.; Seaman, Robert L.; Evans, Eric Michael; Singh, Navdeep; Taylor, Clark; Toeniskoetter, Jackson; Welch, Eric; Zhu, Songzhe.

In: Astrophysical Journal, Supplement Series, Vol. 236, No. 1, 9, 01.05.2018.

Research output: Contribution to journalArticle

Narayan, G, Zaidi, T, Soraisam, MD, Wang, Z, Lochner, M, Matheson, T, Saha, A, Yang, S, Zhao, Z, Kececioglu, JD, Scheidegger, CE, Snodgrass, RT, Axelrod, T, Jenness, T, Maier, RS, Ridgway, ST, Seaman, RL, Evans, EM, Singh, N, Taylor, C, Toeniskoetter, J, Welch, E & Zhu, S 2018, 'Machine-learning-based Brokers for Real-time Classification of the LSST Alert Stream', Astrophysical Journal, Supplement Series, vol. 236, no. 1, 9. https://doi.org/10.3847/1538-4365/aab781
Narayan, Gautham ; Zaidi, Tayeb ; Soraisam, Monika D. ; Wang, Zhe ; Lochner, Michelle ; Matheson, Thomas ; Saha, Abhijit ; Yang, Shuo ; Zhao, Zhenge ; Kececioglu, John D ; Scheidegger, Carlos Eduardo ; Snodgrass, Richard Thomas ; Axelrod, Tim ; Jenness, Tim ; Maier, Robert S ; Ridgway, Stephen T. ; Seaman, Robert L. ; Evans, Eric Michael ; Singh, Navdeep ; Taylor, Clark ; Toeniskoetter, Jackson ; Welch, Eric ; Zhu, Songzhe. / Machine-learning-based Brokers for Real-time Classification of the LSST Alert Stream. In: Astrophysical Journal, Supplement Series. 2018 ; Vol. 236, No. 1.
@article{05ec44d00e314feb80b145cce327be99,
title = "Machine-learning-based Brokers for Real-time Classification of the LSST Alert Stream",
abstract = "The unprecedented volume and rate of transient events that will be discovered by the Large Synoptic Survey Telescope (LSST) demand that the astronomical community update its follow-up paradigm. Alert-brokers- A utomated software system to sift through, characterize, annotate, and prioritize events for follow-up-will be critical tools for managing alert streams in the LSST era. The Arizona-NOAO Temporal Analysis and Response to Events System (ANTARES) is one such broker. In this work, we develop a machine learning pipeline to characterize and classify variable and transient sources only using the available multiband optical photometry. We describe three illustrative stages of the pipeline, serving the three goals of early, intermediate, and retrospective classification of alerts. The first takes the form of variable versus transient categorization, the second a multiclass typing of the combined variable and transient data set, and the third a purity-driven subtyping of a transient class. Although several similar algorithms have proven themselves in simulations, we validate their performance on real observations for the first time. We quantitatively evaluate our pipeline on sparse, unevenly sampled, heteroskedastic data from various existing observational campaigns, and demonstrate very competitive classification performance. We describe our progress toward adapting the pipeline developed in this work into a real-time broker working on live alert streams from time-domain surveys.",
keywords = "data analysis-methods, general-supernovae, general-surveys-virtual observatory tools, methods, statistical-stars, variables",
author = "Gautham Narayan and Tayeb Zaidi and Soraisam, {Monika D.} and Zhe Wang and Michelle Lochner and Thomas Matheson and Abhijit Saha and Shuo Yang and Zhenge Zhao and Kececioglu, {John D} and Scheidegger, {Carlos Eduardo} and Snodgrass, {Richard Thomas} and Tim Axelrod and Tim Jenness and Maier, {Robert S} and Ridgway, {Stephen T.} and Seaman, {Robert L.} and Evans, {Eric Michael} and Navdeep Singh and Clark Taylor and Jackson Toeniskoetter and Eric Welch and Songzhe Zhu",
year = "2018",
month = "5",
day = "1",
doi = "10.3847/1538-4365/aab781",
language = "English (US)",
volume = "236",
journal = "Astrophysical Journal, Supplement Series",
issn = "0067-0049",
publisher = "IOP Publishing Ltd.",
number = "1",

}

TY - JOUR

T1 - Machine-learning-based Brokers for Real-time Classification of the LSST Alert Stream

AU - Narayan, Gautham

AU - Zaidi, Tayeb

AU - Soraisam, Monika D.

AU - Wang, Zhe

AU - Lochner, Michelle

AU - Matheson, Thomas

AU - Saha, Abhijit

AU - Yang, Shuo

AU - Zhao, Zhenge

AU - Kececioglu, John D

AU - Scheidegger, Carlos Eduardo

AU - Snodgrass, Richard Thomas

AU - Axelrod, Tim

AU - Jenness, Tim

AU - Maier, Robert S

AU - Ridgway, Stephen T.

AU - Seaman, Robert L.

AU - Evans, Eric Michael

AU - Singh, Navdeep

AU - Taylor, Clark

AU - Toeniskoetter, Jackson

AU - Welch, Eric

AU - Zhu, Songzhe

PY - 2018/5/1

Y1 - 2018/5/1

N2 - The unprecedented volume and rate of transient events that will be discovered by the Large Synoptic Survey Telescope (LSST) demand that the astronomical community update its follow-up paradigm. Alert-brokers- A utomated software system to sift through, characterize, annotate, and prioritize events for follow-up-will be critical tools for managing alert streams in the LSST era. The Arizona-NOAO Temporal Analysis and Response to Events System (ANTARES) is one such broker. In this work, we develop a machine learning pipeline to characterize and classify variable and transient sources only using the available multiband optical photometry. We describe three illustrative stages of the pipeline, serving the three goals of early, intermediate, and retrospective classification of alerts. The first takes the form of variable versus transient categorization, the second a multiclass typing of the combined variable and transient data set, and the third a purity-driven subtyping of a transient class. Although several similar algorithms have proven themselves in simulations, we validate their performance on real observations for the first time. We quantitatively evaluate our pipeline on sparse, unevenly sampled, heteroskedastic data from various existing observational campaigns, and demonstrate very competitive classification performance. We describe our progress toward adapting the pipeline developed in this work into a real-time broker working on live alert streams from time-domain surveys.

AB - The unprecedented volume and rate of transient events that will be discovered by the Large Synoptic Survey Telescope (LSST) demand that the astronomical community update its follow-up paradigm. Alert-brokers- A utomated software system to sift through, characterize, annotate, and prioritize events for follow-up-will be critical tools for managing alert streams in the LSST era. The Arizona-NOAO Temporal Analysis and Response to Events System (ANTARES) is one such broker. In this work, we develop a machine learning pipeline to characterize and classify variable and transient sources only using the available multiband optical photometry. We describe three illustrative stages of the pipeline, serving the three goals of early, intermediate, and retrospective classification of alerts. The first takes the form of variable versus transient categorization, the second a multiclass typing of the combined variable and transient data set, and the third a purity-driven subtyping of a transient class. Although several similar algorithms have proven themselves in simulations, we validate their performance on real observations for the first time. We quantitatively evaluate our pipeline on sparse, unevenly sampled, heteroskedastic data from various existing observational campaigns, and demonstrate very competitive classification performance. We describe our progress toward adapting the pipeline developed in this work into a real-time broker working on live alert streams from time-domain surveys.

KW - data analysis-methods

KW - general-supernovae

KW - general-surveys-virtual observatory tools

KW - methods

KW - statistical-stars

KW - variables

UR - http://www.scopus.com/inward/record.url?scp=85047260546&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047260546&partnerID=8YFLogxK

U2 - 10.3847/1538-4365/aab781

DO - 10.3847/1538-4365/aab781

M3 - Article

AN - SCOPUS:85047260546

VL - 236

JO - Astrophysical Journal, Supplement Series

JF - Astrophysical Journal, Supplement Series

SN - 0067-0049

IS - 1

M1 - 9

ER -