@misc{indiciaec58c158047cc,
  title = {Universal scaling laws in the gradient descent training of neural networks},
  author = {Maksim Velikanov and Dmitry Yarotsky},
  year = {2021},
  url = {https://arxiv.org/abs/2105.00507},
  note = {Source identifier: 2105.00507}
}