Machine learning (ML) encompasses a large range of algorithms used for a broad range of techniques, from data processing to interpretation and emulators at different scales. These techniques have entered most scientific disciplines in recent years and astronomy is particularly well suited to these now mature methods. The size of current and future astronomical datasets, acquired by telescopes or produced by numerical simulations, is a BigData challenge for the community, which can be tackled by fast and well-designed ML methods. ML can be used for i) image processing, ii) object classification, iii) cosmological inference, or iv) emulate expensive simulations. For several years, our PNCG community has developed supervised (classification, regression) or unsupervised (clustering) methods, applied to cosmology, galaxy populations (morphologies, redshifts…) and several other purposes. The goal of this workshop is to gather people interested in Machine Learning to show recent results demonstrating the power of these methods, share skills and experience with future users thanks to more dedicated talks and to discuss future prospects (Euclid, SKA, Model inversion..).
