Why batch normalization is used?

Related questions

1. Which of these is a benefit of batch normalization?
2. Does it make sense to initialize all weights in a deep network to 0 justify?
3. Why does Xavier initialization work?
4. What is weight Ann?
5. Why is Softmax used?
6. How do I initialize an unallocated disk?
7. What does Batch Norm do?
8. What is S shaped curve?
9. How do I find the hidden layer size?
10. What is the cost function in machine learning?

Tags

1. #weight
2. #initialisation
3. #train
4. #creating
5. #difficult
6. #using
7. #reduce
8. #weights
9. #batch
10. #networks
11. #rates
12. #allows
13. #normalisation
14. #which
15. #sensitivity
16. #starting
17. #helps
18. #initial
19. #especially
20. #initialise
21. #higher
22. #learning
23. #easier
24. #increasing
25. #deeper
26. #makes
27. #speed