Supernova search with active learning in ZTF DR3

Context. We provide the first results from complete SNAD adaptive learning pipeline in context of a broad scope data large-scale astronomical surveys. Aims. The main goal this work is to explore potential techniques application big sets. Methods. Our team used Active Anomaly Discovery (AAD) as tool search for new supernova (SN) candidates photometric 9.4 months Zwicky Transient Facility (ZTF) survey, namely, between March 17 and December 31, 2018 (58 194 ≤ MJD 58 483). analysed 70 ZTF fields at high galactic latitude visually inspected 2100 outliers. Results. This resulted 104 SN-like objects being found, 57 which were reported Name Server time with 47 having previously been mentioned other catalogues, either SNe known types or SN candidates. multi-colour light curves non-catalogued transients performed fittings different models assign it probable class: Ia, Ib/c, IIP, IIL, IIn. Moreover, we also identified unreported slow-evolving that are good superluminous candidates, along few objects, such red dwarf flares active nuclei. Conclusions. Beyond confirming effectiveness human-machine integration underlying AAD strategy, our shed on leaks currently available pipelines. These findings can help avoid similar losses future Furthermore, algorithm enables direct searches any type based definition an anomaly set by expert.