@misc{indiciae05cda9cc1d60,
  title = {Learning deep linear neural networks: Riemannian gradient flows and convergence to global minimizers},
  author = {Bubacarr Bah and Holger Rauhut and Ulrich Terstiege and Michael Westdickenberg},
  year = {2020},
  url = {https://arxiv.org/abs/1910.05505},
  note = {Source identifier: 1910.05505}
}