In this talk, a similarity search method for finding missed transients called the “SNAD Miner”, developed during the fourth workshop of the SuperNova Anomaly Detection (SNAD) team— an international network of researchers from Sternberg astronomical institute (http://www.sai.msu.su/), Laboratoire de Physique de Clermont (http://clrwww.in2p3.fr/), Space Research Institute (http://www.iki.rssi.ru/eng/), Moscow Institute of Physics and Technology (https://mipt.ru/english/), and the University of Illinois Urbana-Champaign (https://astro.illinois.edu/) joined together to solve the problem of detecting unusual objects in astronomical databases with machine learning methods ---will be discussed. The proof-of-concept mining strategy employs physically motivated features extracted from both real light curves and four simulated light curve models (SN Ia, SN II, TDE, SLSN-I). These features are input to a k-D tree algorithm, from which we calculate nearest neighbors to simulated light curves amongst our real data set, and visually inspect the light curves of unique sources. To date, the SNAD Miner is approaching 200 missed supernova discoveries within the ZTF Data Release. Our result illustrates the potential of coherently incorporating domain knowledge and automatic learning algorithms, which is one of the guiding principles directing the SNAD team. It also demonstrates that the ZTF DR is a suitable testing ground for data mining algorithms aiming to prepare for the next generation of astronomical data.
Join ZoomGov Meeting