Login / Signup

Batch Normalization Is Blind to the First and Second Derivatives of the Loss.

Zhanpeng ZhouWen ShenHuixin ChenLing TangYuefeng ChenQuanshi Zhang
Published in: AAAI (2024)
Keyphrases
  • higher order
  • batch mode
  • quality prediction
  • artificial neural networks
  • normalization method
  • neural network
  • data sets
  • data mining
  • finite difference