Why did the authors of D2GAN propose two discriminators using KL and reverse KL divergence losses instead of one discriminator using JS divergence

Asked Sep 17 '22 at 14:39

Active Sep 17 '22 at 16:27

Viewed 32 times

I have stumbled upon the D2GAN paper as part of my research and I am finding myself extremely confused by the fact that instead of using the JS divergence to capture both KL and its reverse, they opted for two discriminators with opposite objective functions. Why is this (if at all) preferable to using just one discriminator that tries to minimize the JS divergence?

edited Sep 17 '22 at 16:27

asked Sep 17 '22 at 14:39

Minuano

Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. – Community Sep 17 '22 at 16:19

Why did the authors of D2GAN propose two discriminators using KL and reverse KL divergence losses instead of one discriminator using JS divergence

0 Answers0