Machine learning for transient discovery in Pan-STARRS1 difference imaging

Abstract

Efficient identification and follow-up of astronomical transients is hindered by the need for humans to manually select promising candidates from data streams that contain many false positives. These artefacts arise in the difference images that are produced by most major ground-based time domain surveys with large format CCD cameras. This dependence on humans to reject bogus detections is unsustainable for next generation all-sky surveys and significant effort is now being invested to solve the problem computationally. In this paper we explore a simple machine learning approach to real-bogus classification by constructing a training set from the image data of ~32000 real astrophysical transients and bogus detections from the Pan-STARRS1 Medium Deep Survey. We derive our feature representation from the pixel intensity values of a 20x20 pixel stamp around the centre of the candidates. This differs from previous work in that it works directly on the pixels rather than catalogued domain knowledge for feature design or selection. Three machine learning algorithms are trained (artificial neural networks, support vector machines and random forests) and their performances are tested on a held-out subset of 25% of the training data. We find the best results from the random forest classifier and demonstrate that by accepting a false positive rate of 1%, the classifier initially suggests a missed detection rate of around 10%. However we also find that a combination of bright star variability, nuclear transients and uncertainty in human labelling means that our best estimate of the missed detection rate is approximately 6%.

17 pages, 16 figures, accepted for publication in MNRAS

References in corpus (23)

• The Sloan Digital Sky Survey: Technical Summary
• Proceedings of the 29th International Conference on Machine Learning (ICML-12)
• First Results from the Catalina Real-time Transient Survey
• The Pan-STARRS1 Photometric System
• Exploring the Optical Transient Sky with the Palomar Transient Factory
• SkyMapper and the Southern Sky Survey
• An ultraviolet-optical flare from the tidal disruption of a helium-rich stellar core
• Photometric Calibration of the First 1.5 Years of the Pan-STARRS1 Survey
• A Wolf-Rayet-like progenitor of supernova SN 2013cu from spectral observations of a wind
• Cosmological Constraints from Measurements of Type Ia Supernovae discovered during the first 1.5 years of the Pan-STARRS1 Survey
• The Pan-STARRS 1 Photometric Reference Ladder, Release 12.0
• Discovery, Progenitor & Early Evolution of a Stripped Envelope Supernova iPTF13bvn
• Testing LMC Microlensing Scenarios: The Discrimination Power of the SuperMACHO Microlensing Survey
• Pan-STARRS1 Discovery of Two Ultra-Luminous Supernovae at z ~ 0.9
• Automating Discovery and Classification of Transients and Variable Stars in the Synoptic Survey Era
• Using Machine Learning for Discovery in Synoptic Survey Imaging
• PS1-10bzj: A Fast, Hydrogen-Poor Superluminous Supernova in a Metal Poor Host Galaxy
• PS1-10afx at z=1.388: Pan-STARRS1 Discovery of a New Type of Superluminous Supernova
• The superluminous supernova PS1-11ap: bridging the gap between low and high redshift
• How to Find More Supernovae with Less Work: Object Classification Techniques for Difference Imaging
• First Results from Pan-STARRS1: Faint, High Proper Motion White Dwarfs in the Medium-Deep Fields
• Ultra-Luminous Supernovae as a New Probe of the Interstellar Medium in Distant Galaxies
• GALEX and Pan-STARRS1 Discovery of SN IIP 2010aq: The First Few Days After Shock Breakout in a Red Supergiant Star

Cited by in corpus (53)

• Design and operation of the ATLAS Transient Science Server
• Proper image subtraction - optimal transient detection, photometry and hypothesis testing
• Photometric Supernova Classification With Machine Learning
• Real-bogus classification for the Zwicky Transient Facility using deep learning
• Automated Transient Identification in the Dark Energy Survey
• Machine Learning for the Zwicky Transient Facility
• The weirdest SDSS galaxies: results from an outlier detection algorithm
• Pan-STARRS Pixel Analysis : Source Detection and Characterization
• PS18kh: A New Tidal Disruption Event with a Non-Axisymmetric Accretion Disk
• A Morphological Classification Model to Identify Unresolved PanSTARRS1 Sources: Application in the ZTF Real-Time Pipeline
• The IPAC Image Subtraction and Discovery Pipeline for the intermediate Palomar Transient Factory
• Machine Learning-based Brokers for Real-time Classification of the LSST Alert Stream
• An automatic taxonomy of galaxy morphology using unsupervised machine learning
• Radio Galaxy Zoo: Compact and extended radio source classification with deep learning
• Pan-STARRS and PESSTO search for an optical counterpart to the LIGO gravitational wave source GW150914
• A DECam Search for an Optical Counterpart to the LIGO Gravitational Wave Event GW151226
• Observations of the GRB afterglow ATLAS17aeu and its possible association with GW170104
• A transient search using combined human and machine classifications
• An analysis of feature relevance in the classification of astronomical transients with machine learning methods
• A search for an optical counterpart to the gravitational wave event GW151226
• Convolutional Neural Networks for Transient Candidate Vetting in Large-Scale Surveys
• Anomaly Detection in the Open Supernova Catalog
• Transient-optimised real-bogus classification with Bayesian Convolutional Neural Networks -- sifting the GOTO candidate stream
• Searches after Gravitational Waves Using ARizona Observatories (SAGUARO): Observations and Analysis from Advanced LIGO/Virgo's Third Observing Run
• Machine Learning for Observational Cosmology
• Optical follow-up of gravitational wave events during the second advanced LIGO/VIRGO observing run with the DLT40 Survey
• Euclid: Superluminous supernovae in the Deep Survey
• Machine-learning Selection of Optical Transients in Subaru/Hyper Suprime-Cam Survey
• Light curve classification with recurrent neural networks for GOTO: dealing with imbalanced data
• Discovery and Rapid Follow-up Observations of the Unusual Type II SN 2018ivc in NGC 1068
• Machine Learning Based Real Bogus System for HSC-SSP Moving Object Detecting Pipeline
• Search for transient optical counterparts to high-energy IceCube neutrinos with Pan-STARRS1
• Machine-z: Rapid Machine Learned Redshift Indicator for Swift Gamma-ray Bursts
• Machine Learning for Transient Recognition in Difference Imaging With Minimum Sampling Effort
• Accurately constraining velocity information from spectral imaging observations using machine learning techniques
• deepSIP: Linking Type Ia Supernova Spectra to Photometric Quantities with Deep Learning
• Panning for gold, but finding helium: discovery of the ultra-stripped supernova SN2019wxt from gravitational-wave follow-up observations
• Quantifying correlations between galaxy emission lines and stellar continua
• GPU-Based High-Performance Imaging for Mingantu Spectral RadioHeliograph
• Ask The Machine: Systematic detection of wind-type outflows in low-mass X-ray binaries
• Results from the Pan-STARRS Search for Kilonovae: Contamination by Massive Stellar Outbursts
• Merged or monolithic? Using machine-learning to reconstruct the dynamical history of simulated star clusters
• Identifying Transients in the Dark Energy Survey using Convolutional Neural Networks
• Noise reduction on single-shot images using an autoencoder
• GWOPS: A VO-technology Driven Tool to Search for the Electromagnetic Counterpart of Gravitational Wave Event
• Deep-learning Real/Bogus classification for the Tomo-e Gozen transient survey
• Automated Transient Detection with Shapelet Analysis in Image-subtracted Data
• The classification of real and bogus transients using active learning and semi-supervised learning
• Real-bogus scores for active anomaly detection
• The Gravitational-wave Optical Transient Observer (GOTO): prototype performance and prospects for transient science
• Image plane detection of FRB121102 with the MeerKAT radio telescope
• The automation of optical transient discovery and classification in Rubin-era time-domain astronomy
• Phenomenological classification of the Zwicky Transient Facility astronomical event alerts