Speaker
Description
It has been said that there is nothing as useless as a radio
source. Certainly, the science is greatly magnified if the radio sources can be cross-identified with other multi-messenger catalogues such as CTA, and redshifts measured or estimated. But it is non-trivial to do so for the catalogues of tens of millions of sources that will be generated by
next-generation radio continuum surveys, especially since many of those sources are extended or contain multiple components, none of which may correspond to the optical host galaxy. In this talk I review recent developments, mainly using machine learning techniques, for radio source
cross-identification, classification, and redshift estimation, and outline the science, including the unexpected, that we may expect to result from them.
