Publication

Jamming transition as a paradigm to understand the loss landscape of deep neural networks