Which of the following best describes the difference between generative and discriminative models?
Answer options
A
Discriminative models can't generate data
B
Generative models are always better
C
Generative models are used for classification only
D
Generative models learn the data distribution, while
E
discriminative models learn the decision boundary
F
Generative models are older in concept
Correct answer: Generative models learn the data distribution, while, discriminative models learn the decision boundary
Explanation
Generative models learn the underlying data distribution P(X) or P(X,Y), enabling them to generate new samples. Discriminative models focus on learning the decision boundary (P(Y|X)) between classes. Options [3] and [4] together form the complete answer: 'Generative models learn the data distribution, while discriminative models learn the decision boundary.'