Unrolled Generative Adversarial Networks [article]

Luke Metz, Ben Poole, David Pfau, Jascha Sohl-Dickstein
2017 arXiv   pre-print
We introduce a method to stabilize Generative Adversarial Networks (GANs) by defining the generator objective with respect to an unrolled optimization of the discriminator. This allows training to be adjusted between using the optimal discriminator in the generator's objective, which is ideal but infeasible in practice, and using the current value of the discriminator, which is often unstable and leads to poor solutions. We show how this technique solves the common problem of mode collapse,
more » ... ilizes training of GANs with complex recurrent generators, and increases diversity and coverage of the data distribution by the generator.
arXiv:1611.02163v4 fatcat:likkc2vogjghvpgfleg76ciafe