Files in this item



application/pdfCHEN-THESIS-2019.pdf (2MB)
(no description provided)PDF


Title:Deep generative models via explicit Wasserstein minimization
Author(s):Chen, Yucheng
Advisor(s):Peng, Jian
Department / Program:Computer Science
Discipline:Computer Science
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Deep Generative Models
Generative Adversarial Networks
Optimal Transport
Abstract:This thesis provides a 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 decreased 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 minimize the Wasserstein distance to both the empirical target distribution, and also its underlying population counterpart. Empirically, good performance is demonstrated on the training and testing sets of the MNIST and Thin-8 data. As a side product, the thesis proposes several effective metrics of measure performance of deep generative models. The thesis closes with a discussion of the unsuitability of the Wasserstein distance for certain tasks, as has been identified in prior work.
Issue Date:2019-04-25
Rights Information:Copyright 2019 Yucheng Chen
Date Available in IDEALS:2019-08-23
Date Deposited:2019-05

This item appears in the following Collection(s)

Item Statistics