From Galaxy Zoo to Euclid: classifying galaxies with citizen science and artificial intelligence

Sandor Kruk

ESTEC

Galaxies have different shapes, sizes and colours which tell us about how they formed and evolved. Citizen Science projects such as Galaxy Zoo have demonstrated that hundreds of thousands of volunteers can successfully classify galaxies in existing large surveys and make significant contributions to research. Future big data missions such as Euclid, however, will provide an unprecedented number of galaxies, which cannot be classified with visual inspection. New methods based on deep learning are increasingly powerful and can aid tackle this problem, but they require large, accurately labelled samples of training data. In this talk, I will present how we can combine human and machine intelligence for the accurate classification of galaxies on short timescales, which can be applied to future missions such as Euclid.