Weakly supervised classifiers trained on background-versus-mixture samples can identify anomalous gamma-ray sources without labeled signal templates, approaching supervised performance in controlled benchmarks.
3FGLzoo. Classifying 3FGL Unassociated Fermi-LAT Gamma-ray Sources by Artificial Neural Networks
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
abstract
In its first four years of operation, the Fermi Large Area Telescope (LAT) detected 3033 $\gamma$-ray emitting sources. In the Fermi-LAT Third Source Catalogue (3FGL) about 50% of the sources have no clear association with a likely $\gamma$-ray emitter. We use an artificial neural network algorithm aimed at distinguishing BL Lacs from FSRQs to investigate the source subclass of 559 3FGL unassociated sources characterised by $\gamma$-ray properties very similar to those of Active Galactic Nuclei. Based on our method, we can classify 271 objects as BL Lac candidates, 185 as FSRQ candidates, leaving only 103 without a clear classification. we suggest a new zoo for $\gamma$-ray objects, where the percentage of sources of uncertain type drops from 52% to less than 10%. The result of this study opens up new considerations on the population of the $\gamma$-ray sky, and it will facilitate the planning of significant samples for rigorous analyses and multiwavelength observational campaigns.
fields
astro-ph.HE 2representative citing papers
Filtering unidentified Fermi sources and comparing to repopulated VL-II simulations yields upper limits of 4e-26 cm3/s (10 GeV) and 5e-25 cm3/s (100 GeV) on tau-pair annihilation.
citing papers explorer
-
Weakly supervised machine learning for model-agnostic searches of new phenomena in the $\gamma$-ray sky
Weakly supervised classifiers trained on background-versus-mixture samples can identify anomalous gamma-ray sources without labeled signal templates, approaching supervised performance in controlled benchmarks.
-
Unidentified Gamma-ray Sources as Targets for Indirect Dark Matter Detection with the Fermi-Large Area Telescope
Filtering unidentified Fermi sources and comparing to repopulated VL-II simulations yields upper limits of 4e-26 cm3/s (10 GeV) and 5e-25 cm3/s (100 GeV) on tau-pair annihilation.