🏢 University of Chile

Overparametrized neural networks’ learning dynamics are analyzed under data symmetries using mean-field theory, revealing that data augmentation, feature averaging, and equivariant architectures asymp…