Critical Points and Convergence Analysis of Generative Deep Linear Networks Trained with Bures-Wasserstein Loss.

Pierre Brchet, Katerina Papagiannouli, Jing An, Guido Montfar

Browse the full ICML paper archive.