@article{taylorizedtrainingtowardsbetter,
  title = {Taylorized Training: Towards Better Approximation of Neural Network Training at Finite Width},
  author = {Yu Bai and Ben Krause and Huan Wang and Caiming Xiong and Richard Socher},
  year = {2020},
  eprint = {2002.04010},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2002.04010v2},
}