skip to main content
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

This content will become publicly available on December 4, 2020

Title: Avocado: Photometric Classification of Astronomical Transients with Gaussian Process Augmentation

Abstract

Upcoming astronomical surveys such as the Large Synoptic Survey Telescope (LSST) will rely on photometric classification to identify the majority of the transients and variables that they discover. We present a set of techniques for photometric classification that can be applied even when the training set of spectroscopically confirmed objects is heavily biased toward bright, low-redshift objects. Using Gaussian process regression to model arbitrary light curves in all bands simultaneously, we "augment" the training set by generating new versions of the original light curves covering a range of redshifts and observing conditions. We train a boosted decision tree classifier on features extracted from the augmented light curves, and we show how such a classifier can be designed to produce classifications that are independent of the redshift distributions of objects in the training sample. Our classification algorithm was the best-performing among the 1094 models considered in the blinded phase of the Photometric LSST Astronomical Time-Series Classification Challenge, scoring 0.468 on the organizers' logarithmic-loss metric with flat weights for all object classes in the training set, and achieving an AUC of 0.957 for classification of SNe Ia. Furthermore, our results suggest that spectroscopic campaigns used for training photometric classifiers should focus onmore » typing large numbers of well-observed, intermediate-redshift transients, instead of attempting to type a sample of transients that is directly representative of the full data set being classified.« less

Authors:
ORCiD logo [1]
  1. Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States); Univ. of California, Berkeley, CA (United States); Univ. of Washington, Seattle, WA (United States)
Publication Date:
Research Org.:
Univ. of Chicago, IL (United States)
Sponsoring Org.:
USDOE Office of Science (SC), High Energy Physics (HEP) (SC-25)
OSTI Identifier:
1593890
Grant/Contract Number:  
SC0009924
Resource Type:
Accepted Manuscript
Journal Name:
Astronomical Journal (Online)
Additional Journal Information:
Journal Name: Astronomical Journal (Online); Journal Volume: 158; Journal Issue: 6; Journal ID: ISSN 1538-3881
Publisher:
IOP Publishing - AAAS
Country of Publication:
United States
Language:
English
Subject:
79 ASTRONOMY AND ASTROPHYSICS; photometric classification; LSST; transients; supernovae

Citation Formats

Boone, Kyle. Avocado: Photometric Classification of Astronomical Transients with Gaussian Process Augmentation. United States: N. p., 2019. Web. doi:10.3847/1538-3881/ab5182.
Boone, Kyle. Avocado: Photometric Classification of Astronomical Transients with Gaussian Process Augmentation. United States. doi:10.3847/1538-3881/ab5182.
Boone, Kyle. Wed . "Avocado: Photometric Classification of Astronomical Transients with Gaussian Process Augmentation". United States. doi:10.3847/1538-3881/ab5182.
@article{osti_1593890,
title = {Avocado: Photometric Classification of Astronomical Transients with Gaussian Process Augmentation},
author = {Boone, Kyle},
abstractNote = {Upcoming astronomical surveys such as the Large Synoptic Survey Telescope (LSST) will rely on photometric classification to identify the majority of the transients and variables that they discover. We present a set of techniques for photometric classification that can be applied even when the training set of spectroscopically confirmed objects is heavily biased toward bright, low-redshift objects. Using Gaussian process regression to model arbitrary light curves in all bands simultaneously, we "augment" the training set by generating new versions of the original light curves covering a range of redshifts and observing conditions. We train a boosted decision tree classifier on features extracted from the augmented light curves, and we show how such a classifier can be designed to produce classifications that are independent of the redshift distributions of objects in the training sample. Our classification algorithm was the best-performing among the 1094 models considered in the blinded phase of the Photometric LSST Astronomical Time-Series Classification Challenge, scoring 0.468 on the organizers' logarithmic-loss metric with flat weights for all object classes in the training set, and achieving an AUC of 0.957 for classification of SNe Ia. Furthermore, our results suggest that spectroscopic campaigns used for training photometric classifiers should focus on typing large numbers of well-observed, intermediate-redshift transients, instead of attempting to type a sample of transients that is directly representative of the full data set being classified.},
doi = {10.3847/1538-3881/ab5182},
journal = {Astronomical Journal (Online)},
number = 6,
volume = 158,
place = {United States},
year = {2019},
month = {12}
}

Journal Article:
Free Publicly Available Full Text
This content will become publicly available on December 4, 2020
Publisher's Version of Record

Save / Share:

Works referenced in this record:

How to Find More Supernovae with Less Work: Object Classification Techniques for Difference Imaging
journal, August 2007

  • Bailey, S.; Aragon, C.; Romano, R.
  • The Astrophysical Journal, Vol. 665, Issue 2
  • DOI: 10.1086/519832

The Astropy Project: Building an Open-science Project and Status of the v2.0 Core Package
journal, August 2018

  • Price-Whelan, A. M.; Sipőcz, B. M.; Günther, H. M.
  • The Astronomical Journal, Vol. 156, Issue 3
  • DOI: 10.3847/1538-3881/aabc4f

Measures of location and scale for velocities in clusters of galaxies - A robust approach
journal, July 1990

  • Beers, Timothy C.; Flynn, Kevin; Gebhardt, Karl
  • The Astronomical Journal, Vol. 100
  • DOI: 10.1086/115487

HUBBLE RESIDUALS OF NEARBY TYPE Ia SUPERNOVAE ARE CORRELATED WITH HOST GALAXY MASSES
journal, May 2010


Results from the Supernova Photometric Classification Challenge
journal, December 2010

  • Kessler, Richard; Bassett, Bruce; Belov, Pavel
  • Publications of the Astronomical Society of the Pacific, Vol. 122, Issue 898
  • DOI: 10.1086/657607

The Calculation of Posterior Distributions by Data Augmentation
journal, June 1987


SNEMO: Improved Empirical Models for Type Ia Supernovae
journal, December 2018


Bayesian Single-Epoch Photometric Classification of Supernovae
journal, July 2007

  • Poznanski, Dovi; Maoz, Dan; Gal-Yam, Avishay
  • The Astronomical Journal, Vol. 134, Issue 3
  • DOI: 10.1086/520956

SNANA: A Public Software Package for Supernova Analysis
journal, September 2009

  • Kessler, Richard; Bernstein, Joseph P.; Cinabro, David
  • Publications of the Astronomical Society of the Pacific, Vol. 121, Issue 883
  • DOI: 10.1086/605984

Improved Cosmological Constraints from New, Old, and Combined Supernova Data Sets
journal, October 2008

  • Kowalski, M.; Rubin, D.; Aldering, G.
  • The Astrophysical Journal, Vol. 686, Issue 2
  • DOI: 10.1086/589937

The Pan-STARRS wide-field optical/NIR imaging survey
conference, July 2010

  • Kaiser, Nick; Burgett, William; Chambers, Ken
  • SPIE Astronomical Telescopes + Instrumentation, SPIE Proceedings
  • DOI: 10.1117/12.859188

Observational Evidence from Supernovae for an Accelerating Universe and a Cosmological Constant
journal, September 1998

  • Riess, Adam G.; Filippenko, Alexei V.; Challis, Peter
  • The Astronomical Journal, Vol. 116, Issue 3
  • DOI: 10.1086/300499

Models and Simulations for the Photometric LSST Astronomical Time Series Classification Challenge (PLAsTiCC)
journal, July 2019

  • Kessler, R.; Narayan, G.; Avelino, A.
  • Publications of the Astronomical Society of the Pacific, Vol. 131, Issue 1003
  • DOI: 10.1088/1538-3873/ab26f1

Energy Distributions, K Corrections, and the Stebbins-Whitford Effect for Giant Elliptical Galaxies
journal, October 1968

  • Oke, J. B.; Sandage, Allan
  • The Astrophysical Journal, Vol. 154
  • DOI: 10.1086/149737

Understanding the Lomb–Scargle Periodogram
journal, May 2018


Photometric Supernova Cosmology with Beams and Sdss-Ii
journal, May 2012


STANDARDIZING TYPE Ia SUPERNOVA ABSOLUTE MAGNITUDES USING GAUSSIAN PROCESS DATA REGRESSION
journal, March 2013


Improving Cosmological Distance Measurements Using twin type ia Supernovae
journal, December 2015


The Type Ia supernovae rate with Subaru/XMM-Newton Deep Survey
journal, April 2014

  • Okumura, Jun E.; Ihara, Yutaka; Doi, Mamoru
  • Publications of the Astronomical Society of Japan, Vol. 66, Issue 2
  • DOI: 10.1093/pasj/psu024

The Art of Data Augmentation
journal, March 2001

  • van Dyk, David A.; Meng, Xiao-Li
  • Journal of Computational and Graphical Statistics, Vol. 10, Issue 1
  • DOI: 10.1198/10618600152418584

Astropy: A community Python package for astronomy
journal, September 2013


TYPE Ia SUPERNOVA RATE MEASUREMENTS TO REDSHIFT 2.5 FROM CANDELS: SEARCHING FOR PROMPT EXPLOSIONS IN THE EARLY UNIVERSE
journal, June 2014

  • Rodney, Steven A.; Riess, Adam G.; Strolger, Louis-Gregory
  • The Astronomical Journal, Vol. 148, Issue 1
  • DOI: 10.1088/0004-6256/148/1/13

The Supernova Legacy Survey: measurement of $\Omega_{\mathsf{M}}$, $\Omega_\mathsf{\Lambda}$ and w from the first year data set
journal, January 2006


Evidence of environmental dependencies of Type Ia supernovae from the Nearby Supernova Factory indicated by local H α
journal, December 2013


Photometric Supernova Classification with Machine Learning
journal, August 2016

  • Lochner, Michelle; McEwen, Jason D.; Peiris, Hiranya V.
  • The Astrophysical Journal Supplement Series, Vol. 225, Issue 2
  • DOI: 10.3847/0067-0049/225/2/31

PHOTOMETRIC TYPE Ia SUPERNOVA CANDIDATES FROM THE THREE-YEAR SDSS-II SN SURVEY DATA
journal, August 2011


The Difference Imaging Pipeline for the Transient Search in the dark Energy Survey
journal, November 2015


Type Ia Supernova Discoveries at z > 1 from the Hubble Space Telescope : Evidence for Past Deceleration and Constraints on Dark Energy Evolution
journal, June 2004

  • Riess, Adam G.; Strolger, Louis‐Gregory; Tonry, John
  • The Astrophysical Journal, Vol. 607, Issue 2
  • DOI: 10.1086/383612

The rate of core Collapse Supernovae to Redshift 2.5 from the Candels and Clash Supernova Surveys
journal, November 2015

  • Strolger, Louis-Gregory; Dahlen, Tomas; Rodney, Steven A.
  • The Astrophysical Journal, Vol. 813, Issue 2
  • DOI: 10.1088/0004-637X/813/2/93

PELICAN: deeP architecturE for the LIght Curve ANalysis
journal, June 2019


Unity: Confronting Supernova Cosmology’S Statistical and Systematic Uncertainties in a Unified Bayesian Framework
journal, November 2015


Bayesian estimation applied to multiple species
journal, May 2007


Kernel PCA for Type Ia supernovae photometric classification
journal, January 2013

  • Ishida, E. E. O.; de Souza, R. S.
  • Monthly Notices of the Royal Astronomical Society, Vol. 430, Issue 1
  • DOI: 10.1093/mnras/sts650

Fast Direct Methods for Gaussian Processes
journal, February 2016

  • Ambikasaran, Sivaram; Foreman-Mackey, Daniel; Greengard, Leslie
  • IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 38, Issue 2
  • DOI: 10.1109/TPAMI.2015.2448083

A simple and robust method for automated photometric classification of supernovae using neural networks
journal, December 2012

  • Karpenka, N. V.; Feroz, F.; Hobson, M. P.
  • Monthly Notices of the Royal Astronomical Society, Vol. 429, Issue 2
  • DOI: 10.1093/mnras/sts412

SALT2: using distant supernovae to improve the use of type Ia supernovae as distance indicators
journal, February 2007


STACCATO: a novel solution to supernova photometric classification with biased training sets
journal, October 2017

  • Revsbech, E. A.; Trotta, R.; van Dyk, D. A.
  • Monthly Notices of the Royal Astronomical Society, Vol. 473, Issue 3
  • DOI: 10.1093/mnras/stx2570

Supernova Simulations and Strategies for the dark Energy Survey
journal, June 2012


Matplotlib: A 2D Graphics Environment
journal, January 2007


New Constraints on Ω M , Ω Λ , and w from an Independent Set of 11 High‐Redshift Supernovae Observed with the Hubble Space Telescope
journal, November 2003

  • Knop, R. A.; Aldering, G.; Amanullah, R.
  • The Astrophysical Journal, Vol. 598, Issue 1
  • DOI: 10.1086/378560

Optimizing spectroscopic follow-up strategies for supernova photometric classification with active learning
journal, November 2018

  • Ishida, E. E. O.; Beck, R.; González-Gaitán, S.
  • Monthly Notices of the Royal Astronomical Society, Vol. 483, Issue 1
  • DOI: 10.1093/mnras/sty3015

Semi-supervised learning for photometric supernova classification★: Semi-supervised SN classification
journal, October 2011

  • Richards, Joseph W.; Homrighausen, Darren; Freeman, Peter E.
  • Monthly Notices of the Royal Astronomical Society, Vol. 419, Issue 2
  • DOI: 10.1111/j.1365-2966.2011.19768.x

The LSST operations simulator
conference, August 2014

  • Delgado, Francisco; Saha, Abhijit; Chandrasekharan, Srinivasan
  • SPIE Astronomical Telescopes + Instrumentation, SPIE Proceedings
  • DOI: 10.1117/12.2056898

Measurements of Ω and Λ from 42 High‐Redshift Supernovae
journal, June 1999

  • Perlmutter, S.; Aldering, G.; Goldhaber, G.
  • The Astrophysical Journal, Vol. 517, Issue 2
  • DOI: 10.1086/307221

The NumPy Array: A Structure for Efficient Numerical Computation
journal, March 2011

  • van der Walt, Stéfan; Colbert, S. Chris; Varoquaux, Gaël
  • Computing in Science & Engineering, Vol. 13, Issue 2
  • DOI: 10.1109/MCSE.2011.37

Deep Recurrent Neural Networks for Supernovae Classification
journal, March 2017


Improved cosmological constraints from a joint analysis of the SDSS-II and SNLS supernova samples
journal, August 2014


A survey on Image Data Augmentation for Deep Learning
journal, July 2019