Timezone: »

Oral
A Gradual, Semi-Discrete Approach to Generative Network Training via Explicit Wasserstein Minimization
Yucheng Chen · Matus Telgarsky · Chao Zhang · Bolton Bailey · Daniel Hsu · Jian Peng

Wed Jun 12 11:30 AM -- 11:35 AM (PDT) @ Hall A

This paper provides a simple procedure to fit generative networks to target distributions, with the goal of a small Wasserstein distance (or other optimal transport costs). The approach is based on two principles: (a) if the source randomness of the network is a continuous distribution (the semi-discrete'' setting), then the Wasserstein distance is realized by a deterministic optimal transport mapping; (b) given an optimal transport mapping between a generator network and a target distribution, the Wasserstein distance may be reduced via a regression between the generated data and the mapped target points. The procedure here therefore alternates these two steps, forming an optimal transport and regressing against it, gradually adjusting the generator network towards the target distribution. Mathematically, this approach is shown to both minimize the Wasserstein distance to both the empirical target distribution, and its underlying population counterpart. Empirically, good performance is demonstrated on the training and test sets of MNIST, and Thin-8 datasets. The paper closes with a discussion of the unsuitability of the Wasserstein distance for certain tasks, as has been identified in prior work (Arora et al., 2017; Huang et al., 2017).