Stat 369 – Methods From Statistical Physics

Andrea Montanari, Stanford University, Autumn 2021

Mathematical techniques from statistical physics have been applied with increasing success on problems form computer science, statistics, machine learning. These methods are non-rigorous, but in several cases they were proved to yield correct predictions. This course provides a working knowledge of these methods for non-physicists. No background in physics is required.

Topics might include:

  • The p-spin model and tensor PCA

  • {mathbf Z}_2 synchronization and the Sherrington-Kirkpatrick model

  • Sharp analysis of high-dimensional regression

  • The Hopfield model of a neural network

  • Models on sparse random graphs

Class Times and Locations

  • Building 320, Room 109

  • Friday, 9:45AM-12:45PM

  • First lecture on Friday, September 24