Sebastian Goldt

Role

Lecturer

Affiliation

SISSA

S. Goldt leads the Theory of neural networks group at SISSA. The goal of his group is to develop theories of learning artificial and biological neural networks. His current focus is on understanding how the interplay of data structure, learning rule and architecture shapes learning.

He studied physics at Cambridge University and did his PhD with Udo Seifert at the University of Stuttgart, where he worked on the stochastic thermodynamics of learning. He then spent three years as a post-doc with Florent Krzakala and Lenka Zdeborovà in Paris working at the interface between machine learning and statistical physics. He joined SISSA in September 2020.