Thomas Venville (Swinburne University of Technology) -- Investigating Transient Sources in Color Space with Self Organizing Maps

The advent of several large photometric and imaging surveys, such as the Large Synoptic Survey Telescope Survey and the Euclid and WFIRST missions, in addition do existing high-volume surveys, is producing astronomical data at an unprecedented rate. The efficient and insightful processing of this data is paramount to maximize the scientific value of these programs, however the data rate is such that human classification is infeasible. Furthermore, for next generation surveys producing robust training sets for supervised machine learning will be difficult. In this talk I present methods utilizing self organizing maps to detect astrophysical transients in single epoch photometric data, without temporal information.