Vetting the optical transient candidates detected by the GWAC network using convolutional neural networks.

The observation of the transient sky through a multitude of astrophysical messengers hasled to several scientific breakthroughs these last two decades thanks to the fast evolution ofthe observational techniques and strategies employed by the astronomers. Now, it requiresto be able to coordinate multi-wavelength and multi-messenger follow-up campaign withinstruments both in space and on ground jointly capable of scanning a large fraction of thesky with a high imaging cadency and duty cycle. In the optical domain, the key challengeof the wide field of view telescopes covering tens to hundreds of square degrees is to dealwith the detection, the identification and the classification of hundreds to thousands of opticaltransient (OT) candidates every night in a reasonable amount of time. In the last decade, newautomated tools based on machine learning approaches have been developed to perform thosetasks with a low computing time and a high classification efficiency. In this paper, we presentan efficient classification method using Convolutional Neural Networks (CNN) to discard anybogus falsely detected in astrophysical images in the optical domain. We designed this toolto improve the performances of the OT detection pipeline of the Ground Wide field AngleCameras (GWAC) telescopes, a network of robotic telescopes aiming at monitoring the opticaltransient sky down to R=16 with a 15 seconds imaging cadency. We applied our trainedCNN classifier on a sample of 1472 GWAC OT candidates detected by the real-time detectionpipeline. It yields a good classification performance with 94% of well classified event and afalse positive rate of 4%.