The aim of the ''Mining the variable sky'' project is to establish an independent research group that will study the history of the Galaxy and the physics driving the variability of astrophysical sources through the application of machine learning techniques.

The transient and variable sky are still under-explored. Combining the results from the surveys across the electromagnetic spectrum allows us to discover and study populations of heretofore unknown transients, for example the mysterious fast radio bursts or rare transients such as the tidal disruptions of stars by supermassive black holes. Furthermore, time-domain surveys can help in pinpointing the sources of gravitational waves by identifying their electromagnetic counterparts.

In the case of more common transients like the type Ia supernovae, larger samples are required to study the expansion history of the Universe in more detail. It is therefore necessary to develop rapid and reliable classification methods that will optimize the use of the scarce follow-up resources. Complementary studies of stellar variability, in conjunction with the parallaxes provided by the Gaia mission, allow us to precisely and accurately anchor and apply luminosity-based distance estimation methods on galactic to cosmological scales, study stellar evolution and the Milky Way's history of formation and accretion.