25 September 2022 - 30 September 2022
David Belius (University of Basel), Erwin Bolthausen (University of Zurich), Giuseppe Genovese (University of Basel) Event page & registration
Few other branches of statistical mechanics are pushing the frontiers with other sciences as the theory of spin glasses. In this workshop the focus will be on the newest trends in the mathematical theory of spin glasses and its application to high dimensional statistics and machine learning. Topics of central interest include: pure states in spin glasses and their relation to energy landscapes, Thouless-Andersson-Palmer equations, approximate message passing methods, phase transitions in high dimensional statistical inference and optimization algorithms, non-convex optimisation of high dimensional landscapes, approximate message passing methods, neural network models and perceptrons.
SwissMAP Research Station, Les Diablerets, Switzerland