Machine Enabled Discovery
Exponential technological growth is leading to the generation of ever larger and more complexe samples of astronomical objects. It is no longer feasible to eyeball every source, or to try for every potential correlation. I will show how we use unsupervised machine learning as an intermediary that allows us to find: objects that are worth further scrutiny, sources of interest, and new trends. I will discuss this via examples over a wide range of domains, from galaxies and quasars to gravitational wave sources.