Large scale cosmology surveys are posing a number of new challenges related to the size of the datasets, the richness of the observations, and the need for a new generation of analytics tools that can function in a regime dominated by systematic errors, rather than statistical errors. In order to deal with this class of 'big data' problems, statistics amd machine learning methods must be pushed in new directions. The purpose of this workshop is to bring together a small number of domain scientists and statisticians and machine learning experts to discuss and work on problems in cosmological statistics.