Powered by Indico
v3.2
Lecturer
Giacomo Gradenigo giacomo.gradenigo@gssi.it
Timetable and workload
Lectures: 12 hours
Course description and outcomes
The course is about one of the most challenging and exciting topics in statistical mechanics nowadays: ergodicity breaking transitions in disordered systems and their relationship with phase transitions in algorithms.
The purpose of this short course is to provide the basic ingredients to understand the statistical mechanics of disordered systems and to convince you, by going through some actual calculations, that phase transitions in such systems are closely related to, and provide the fundamental background to understand, efficiency bottlenecks and phase transitions in inference and learning algorithms.
The course, which is closely related to the ongoing series of seminars "Artificial Intelligence 2021", is conceived for 1-st year students but anyone who is interested in the topic is warmly encouraged and welcomed to follow.
References
The material covered in this course can be found partly in Chapter 7 of this notes, and partly in this manuscript, which also represents notes of lectures on this topic (delivered at Padua University).
Giacomo Gradenigo