What is the "contradictory loss" in the "Old Photo Restoration via Deep Latent Space Translation" paper?

Question

In page 4 of the paper Old Photo Restoration via Deep Latent Space Translation, it says the encoder $E_{R,X}$ of $VAE_1$ tries to fool the discriminator with a contradictory loss to ensure that $R$ and $X$ are mapped to the same space. What do they mean by "contradictory loss"?

Answer 1

The contradictory loss is the same loss function that the discriminator would normally use, except with deliberately incorrect labels. That is, when you train the generator, the output of the generator is fed to the discriminator, but instead of the correct label (typically $0$ for a false image), the opposite label is applied (e.g. $1$ for a real image).

This label is condractictory in that it is incorrect and the opposite of the ground truth. However, this is useful, because reducing the loss of this signal represents the goal of the generator. It is important to note that during this phase of the training, you should not update any parameters of the discriminator. The process is followed in order to find gradients for improving the generator only.