> I would argue that GAN's by definition aren't unsupervised

You can define the terms how you want - but in terms of how they're understood in both industry and academia, you are incorrect.