Emanuele Troiani

Emanuele is a postdoctoral fellow supported by the Swiss National Science Foundation Postdoc Mobility fellowship.

His research aims to bridge the gap between theoretical understanding and practical utility in modern AI architectures. In particular, he is interested in identifying which limitations of machine learning are fundamental, and which instead arise from the particular solutions and architectures currently used in practice.

Previously, he was an EDIC PhD Fellow at EPFL under the supervision of Lenka Zdeborová, where he developed exact methods to characterize feature learning in shallow neural networks. Before that, he was an International Selection Scholar in Theoretical Physics at École Normale Supérieure.