When applying deep learning to real-world data, how should Batch normalization be understood? Consider practical model design.

Correct Answer: C

Batch normalization normalizes intermediate activations to stabilize and speed up training

The correct answer is batch normalization normalizes intermediate activations to stabilize and speed up training. This matches the Deep Learning course topic: Batch normalization.