Question 11
Domain 2 — Data, Machine Learning, and Model DevelopmentWhat is the main purpose of using activation functions in neural networks?
Correct answer: B
Explanation
Activation functions let a neural network model complex patterns by making each neuron’s output non-linear. Without them, stacked layers would still behave like a single linear transformation, so they would not capture relationships beyond straight-line boundaries.
Why each option is right or wrong
A. To normalize input data
B. To introduce non-linearity
Activation functions are used to make each neuron’s output a non-linear transformation of its input, which is what allows a multilayer network to represent complex, non-linear relationships. Without an activation function, composing multiple layers of weighted sums collapses to a single linear map, so the network cannot learn anything beyond linear decision boundaries.
C. To reduce dimensionality
D. To cluster data