Mathematical and Computational Aspects of Machine Learning

The past few years have witnessed an impressive growth and development of machine learning methods. Enhanced by large collected databases and improved computational power, these techniques have made spectacular progress in fields like image recognition and have even reached the ability to surpass humans in certain tasks and in games like Go. Still, the conceptual mechanisms on which such forms of learning work are largely not understood. Moreover, there is a complete lack of prediction ability, not only in the efficiency of machine learning, but also on its ability to work at all on a new problem. All this calls for a strong commitment on the part of the mathematical community. The present school aims at connecting international experts at the forefront of research on the mathematical and computational aspects of the problem with the interested scholars, especially the young generations. Courses The programme includes 5 minicourses of 6 hours each. Lectures will be delivered by experts of international standing. Deadline for registration: 31st August 2019. Admission to the courses will be notified by 15th September after the Scientific Committee's decision. Jean Barbier Philipp Grohs Gabriel Peyré Lars Ruthotto Stefano Soatto