We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: Variance Regularizing Adversarial Learning

Abstract: We introduce a novel approach for training adversarial models by replacing the discriminator score with a bi-modal Gaussian distribution over the real/fake indicator variables. In order to do this, we train the Gaussian classifier to match the target bi-modal distribution implicitly through meta-adversarial training. We hypothesize that this approach ensures a non-zero gradient to the generator, even in the limit of a perfect classifier. We test our method against standard benchmark image datasets as well as show the classifier output distribution is smooth and has overlap between the real and fake modes.
Comments: Method is out of date and some results are incorrect
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Cite as: arXiv:1707.00309 [stat.ML]
  (or arXiv:1707.00309v2 [stat.ML] for this version)

Submission history

From: R Devon Hjelm [view email]
[v1] Sun, 2 Jul 2017 15:36:07 GMT (3619kb,D)
[v2] Sun, 19 Aug 2018 20:21:33 GMT (0kb,I)

Link back to: arXiv, form interface, contact.