Classification of Fermi-LAT blazars with Bayesian neural networks

Abstract

The use of Bayesian neural networks is a novel approach for the classification of gamma-ray sources. We focus on the classification of Fermi-LAT blazar candidates, which can be divided into BL Lacertae objects and Flat Spectrum Radio Quasars. In contrast to conventional dense networks, Bayesian neural networks provide a reliable estimate of the uncertainty of the network predictions. We explore the correspondence between conventional and Bayesian neural networks and the effect of data augmentation. We find that Bayesian neural networks provide a robust classifier with reliable uncertainty estimates and are particularly well suited for classification problems that are based on comparatively small and imbalanced data sets. The results of our blazar candidate classification are valuable input for population studies aimed at constraining the blazar luminosity function and to guide future observational campaigns.

arXiv comment22 pages, 10 figures. Minor changes, matches version accepted by JCAP

References in corpus (36)

• SMOTE: Synthetic Minority Over-sampling Technique
• Fermi Large Area Telescope Fourth Source Catalog
• Fermi Large Area Telescope First Source Catalog
• Hands-on Bayesian Neural Networks -- a Tutorial for Deep Learning Users
• Active Galactic Nuclei: what's in a name?
• The Fourth Catalog of Active Galactic Nuclei Detected by the Fermi Large Area Telescope
• The Fermi blazar sequence
• The Cosmic Evolution of Fermi BL Lacertae Objects
• The Luminosity Function of Fermi-detected Flat-Spectrum Radio Quasars
• The Origin of the Extragalactic Gamma-Ray Background and Implications for Dark-Matter Annihilation
• SuperNNova: an open-source framework for Bayesian, Neural Network based supernova classification
• The nature of the Diffuse Gamma-Ray Background
• Classification and Ranking of Fermi LAT Gamma-ray Sources from the 3FGL Catalog using Machine Learning Techniques
• Dissecting the regions around IceCube high-energy neutrinos: growing evidence for the blazar connection
• Fermi's Sibyl: Mining the gamma-ray sky for dark matter subhaloes
• The Extragalactic Gamma-ray Sky in the Fermi era
• 3FGL Demographics Outside the Galactic Plane using Supervised Machine Learning: Pulsar and Dark Matter Subhalo Interpretations
• Deciphering Contributions to the Extragalactic Gamma-Ray Background from 2 GeV to 2 TeV
• The diffuse $\gamma$-ray background is dominated by star-forming galaxies
• Interpretation of the diffuse astrophysical neutrino flux in terms of the blazar sequence
• Blazar Flaring Patterns (B-FlaP): Classifying Blazar Candidates of Uncertain type in the third Fermi-LAT catalog by Artificial Neural Networks
• Deep-Learning Jets with Uncertainties and More
• The GCE in a New Light: Disentangling the $\gamma$-ray Sky with Bayesian Graph Convolutional Neural Networks
• Search for gamma-ray-emitting active galactic nuclei in the Fermi-LAT unassociated sample using machine learning
• Statistical Measurement of the Gamma-ray Source-count Distribution as a Function of Energy
• Per-Object Systematics using Deep-Learned Calibration
• 3FGLzoo. Classifying 3FGL Unassociated Fermi-LAT Gamma-ray Sources by Artificial Neural Networks
• Deriving the contribution of blazars to the Fermi-LAT Extragalactic $\gamma$-ray background at $E>10$ GeV with efficiency corrections and photon statistics
• An investigation on the factors affecting machine learning classifications in $\gamma$-ray astronomy
• Testing gamma-ray models of blazars in the extragalactic sky
• Classification of Blazar Candidates of Uncertain Type from the Fermi LAT 8-Year Source Catalog with an Artificial Neural Network
• Optical spectroscopic observations of gamma-ray blazar candidates. X. Results from the 2018--2019 SOAR and OAN-SPM observations of blazar candidates of uncertain type
• Multiwavelength Spectral Analysis and Neural Network Classification of Counterparts to 4FGL Unassociated Sources
• Classification of Fermi-LAT sources with deep learning using energy and time spectra
• Identification of point sources in gamma rays using U-shaped convolutional neural networks and a data challenge
• Identifying the 3FHL catalog: V. Results of the CTIO-COSMOS optical spectroscopy campaign 2019

Cited by in corpus (2)

• The Dark Energy Survey 5-year photometrically identified Type Ia Supernovae
• Gradient boosting decision trees classification of blazars of uncertain type in the fourth Fermi-LAT catalog