Why does normalization work in batch?

Related questions

1. What is the difference between CNN and Ann?
2. Which is major benefit of reducing batch size?
3. What are strides in CNN?
4. What does convolution layer do?
5. What is 2d convolution layer?
6. Why do we need convolution?
7. What does batch size do?
8. How does CNN decide how many layers?
9. What is batch normalization in keras?
10. Why is Max pooling done?

Tags

1. #stable
2. #therefore
3. #would
4. #allows
5. #variance
6. #inputs
7. #learn
8. #supposedly
9. #training
10. #normalization
11. #accelerate
12. #value
13. #desires
14. #practice
15. #batch
16. #layer
17. #parameters
18. #distribution
19. #convert
20. #network