NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:1034
Title:Multi-marginal Wasserstein GAN

The paper introduces a new, theoretically motivated approach to multi-margin matching. The dual formulation in terms of a shared potential is quite elegant, and experiments are promising. Please take reviewer comments into account when preparing the final version of the paper.