Probing a new regime of dark matter with intelligent systems.

Extensions to the cold dark matter paradigm with predictable observables present pathways to understanding what dark matter is. Upcoming telescopes, like Euclid, we will have the data to test these extensions to an unprecedented level. Yet, with precision must come accuracy. We must be able to not only make robust predictions on how these extensions manifest themselves, but also analyse the data in a systematic free way. In this talk I will show how we are using intelligent systems to probe one particular extension - Self Interacting Dark Matter. I will show how we are using super-sampled simulations in order to probe new regimes of parameter space that have not be explored, and also how deep learning, when used carefully, can aid us analyse these simulations and make quick and robust inferences on the nature of dark matter.