Self-Supervised Clustering on Image-Subtracted Data with Deep-Embedded Self-Organizing Map

Abstract

Developing an effective automatic classifier to separate genuine sources from artifacts is essential for transient follow-ups in wide-field optical surveys. The identification of transient detections from the subtraction artifacts after the image differencing process is a key step in such classifiers, known as real-bogus classification problem. We apply a self-supervised machine learning model, the deep-embedded self-organizing map (DESOM) to this "real-bogus" classification problem. DESOM combines an autoencoder and a self-organizing map to perform clustering in order to distinguish between real and bogus detections, based on their dimensionality-reduced representations. We use 32x32 normalized detection thumbnails as the input of DESOM. We demonstrate different model training approaches, and find that our best DESOM classifier shows a missed detection rate of 6.6% with a false positive rate of 1.5%. DESOM offers a more nuanced way to fine-tune the decision boundary identifying likely real detections when used in combination with other types of classifiers, for example built on neural networks or decision trees. We also discuss other potential usages of DESOM and its limitations.

References in corpus (33)

• Multi-messenger Observations of a Binary Neutron Star Merger
• The Physics of Gamma-Ray Bursts
• An Ordinary Short Gamma-Ray Burst with Extraordinary Implications: Fermi-GBM Detection of GRB 170817A
• A method for optimal image subtraction
• Swope Supernova Survey 2017a (SSS17a), the Optical Counterpart to a Gravitational Wave Source
• Short-Duration Gamma-Ray Bursts
• Physical processes shaping GRB X-ray afterglow lightcurves: theoretical implications from the Swift XRT observations
• The Electromagnetic Counterpart of the Binary Neutron Star Merger LIGO/VIRGO GW170817. II. UV, Optical, and Near-IR Light Curves and Comparison to Kilonova Models
• INTEGRAL Detection of the First Prompt Gamma-Ray Signal Coincident with the Gravitational Wave Event GW170817
• A "kilonova" associated with short-duration gamma-ray burst 130603B
• Exploring the Optical Transient Sky with the Palomar Transient Factory
• The Electromagnetic Counterpart of the Binary Neutron Star Merger LIGO/VIRGO GW170817. IV. Detection of Near-infrared Signatures of r-process Nucleosynthesis with Gemini-South
• A Radio Counterpart to a Neutron Star Merger
• An r-Process Kilonova Associated with the Short-Hard GRB 130603B
• The Electromagnetic Counterpart of the Binary Neutron Star Merger LIGO/VIRGO GW170817. V. Rising X-ray Emission from an Off-Axis Jet
• Design and operation of the ATLAS Transient Science Server
• Localization and broadband follow-up of the gravitational-wave transient GW150914
• Real-bogus classification for the Zwicky Transient Facility using deep learning
• Short GRB 160821B: a reverse shock, a refreshed shock, and a well-sampled kilonova
• Automating Discovery and Classification of Transients and Variable Stars in the Synoptic Survey Era
• A Comprehensive Study of Gamma-Ray Burst Optical Emission: I. Flares and Early Shallow Decay Component
• The Electromagnetic Counterpart of the Binary Neutron Star Merger LIGO/VIRGO GW170817. VII. Properties of the Host Galaxy and Constraints on the Merger Timescale
• Deep-HiTS: Rotation Invariant Convolutional Neural Network for Transient Detection
• Searching for Electromagnetic Counterparts to Gravitational-wave Merger Events with the Prototype Gravitational-wave Optical Transient Observer (GOTO-4)
• iPTF14yb: The First Discovery of a GRB Afterglow Independent of a High-Energy Trigger
• Infrared emission from kilonovae: the case of the nearby short hard burst GRB 160821B
• GRB 081007 and GRB 090424: the surrounding medium, outflows and supernovae
• Convolutional Neural Networks for Transient Candidate Vetting in Large-Scale Surveys
• ZTF20aajnksq (AT2020blt): A Fast Optical Transient at $z\approx 2.9$ With No Detected Gamma-Ray Burst Counterpart
• Transient-optimised real-bogus classification with Bayesian Convolutional Neural Networks -- sifting the GOTO candidate stream
• Machine Learning for Transient Recognition in Difference Imaging With Minimum Sampling Effort
• Searching For Fermi GRB Optical Counterparts With The Prototype Gravitational-Wave Optical Transient Observer (GOTO)
• An astronomical image content-based recommendation system using combined deep learning models in a fully unsupervised mode

Cited by in corpus (1)

• A review of unsupervised learning in astronomy