(Single choice question) Compare autoencoder and generative adversarial network, which of the following statements is false. ()
A、They are both depth generation models
B、Autoencoder can be used as a generative adversarial network
C、For the data with different distribution from the training sample, the output of the autoencoder will be different from the training sample. Generative adversarial networks will be judged as false samples
D、Both can be used for data enhancement