Shouldn't the discriminator and generator work fine even if they don't process data symmetrically? I mean, they don't only receive the final layer results of each other, they don't use data that from hidden layers.

  • 1
    $\begingroup$ Youre completely right, they dont need to be. Intuitively though, it makes sense for them to be (to me atleast), since the descriminator decomposes the images into features and makes a final classification based on this, and you want the generator to learn how to replicate these features properly, so providing the network the same power to generate as compared to discriminate may provide to be useful in the back and fourth game. $\endgroup$ – mshlis Jun 10 '19 at 20:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.