Batch normalization primarily helps by:
Answer options
A
Removing the need for activation functions
B
Stabilizing and speeding up training
C
Making models deterministic
D
Creating new data
Correct answer: Stabilizing and speeding up training
Explanation
Quick AnswerThe correct answer is Stabilizing and speeding up training because it directly addresses the core logic of Generative AI.
Batch normalization normalizes layer inputs during training, reducing internal covariate shift, which stabilizes and accelerates training.