Memorization in Generative Networks and Mixtures of GANs

Simple item page
dc.contributor.advisorBaraniuk, Richard
dc.creatorLuzi, Lorenzo
dc.date.accessioned2020-04-23T16:22:41Z
dc.date.available2021-05-01T05:01:11Z
dc.date.created2020-05
dc.date.issued2020-04-22
dc.date.submittedMay 2020
dc.date.updated2020-04-23T16:22:41Z
dc.description.abstractWe demonstrate that memorization (perfectly fitting the training data) is necessary to avoid mode collapse in generative networks. Using a straightforward measure of the distance between the training data points and the closest point in the range of the generator, we study how well current generative models memorize in terms of the training dataset size, data distribution, and generator architecture. An important hallmark of our GOoF measure is that it does not require a second, trained model as with Frechet Inception Distance or Inception Score. The GOoF measure quantifies that the successful, popular generative models DCGAN, WGAN, and BigGAN fall far short of memorization. Our analysis inspires a new method to circumvent mode collapse by subsampling the training data (either randomly or with $k$-means clustering); we discuss the links to overparameterization. Mixtures of generative adversarial networks (GANs) are closely related to subsampling methods. We study these mixtures in the context of memorization and density estimation to show that mixtures of GANs are superior to training a single GAN under certain assumptions. Furthermore, we construct a theoretic framework that explains how single GANs, mixtures of GANs, conditional GANs, and Gaussian mixture GANs are all related to each other by modifying the typical GAN optimization problem. Finally, we show empirically that our modified optimization problem has a memorization sweet spot which can be found with hyperparameter tuning.
dc.embargo.terms2021-05-01
dc.format.mimetypeapplication/pdf
dc.identifier.citationLuzi, Lorenzo. "Memorization in Generative Networks and Mixtures of GANs." (2020) Master’s Thesis, Rice University. <a href="https://hdl.handle.net/1911/108344">https://hdl.handle.net/1911/108344</a>.
dc.identifier.urihttps://hdl.handle.net/1911/108344
dc.language.isoeng
dc.rightsCopyright is held by the author, unless otherwise indicated. Permission to reuse, publish, or reproduce the work beyond the bounds of fair use or other exemptions to copyright law must be obtained from the copyright holder.
dc.subjectGANs
dc.subjectmemorization
dc.subjectmixtures
dc.titleMemorization in Generative Networks and Mixtures of GANs
dc.typeThesis
dc.type.materialText
thesis.degree.departmentElectrical and Computer Engineering
thesis.degree.disciplineEngineering
thesis.degree.grantorRice University
thesis.degree.levelMasters
thesis.degree.nameMaster of Science
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
LUZI-DOCUMENT-2020.pdf
Size:
4.14 MB
Format:
Adobe Portable Document Format
Download
License bundle
Now showing 1 - 2 of 2
No Thumbnail Available
Name:
PROQUEST_LICENSE.txt
Size:
5.84 KB
Format:
Plain Text
Description:
Download
No Thumbnail Available
Name:
LICENSE.txt
Size:
2.61 KB
Format:
Plain Text
Description:
Download
Collections
Rice University Electronic Theses and Dissertations