Why are autoencoders considered generative models?
Answer options
A
They are used for supervised learning
B
They are only used for image data
C
They can reconstruct and generate data similar to the
D
input
E
They always reduce data dimensionality
F
They are a type of neural network
Correct answer: They can reconstruct and generate data similar to the, input
Explanation
Autoencoders are considered generative models because they learn a compressed latent representation from which they can reconstruct (and generate) new data similar to the training input. Variational Autoencoders extend this by learning a probability distribution over the latent space. Options [2] and [3] together form the complete answer: 'They can reconstruct and generate data similar to the input.'