John Glover

Modeling documents with Generative Adversarial Networks

28 June, 2017

I wrote a brief overview of my work on Modeling documents with Generative Adversarial Networks, which I presented at the NIPS 2016 Workshop on Adversarial Training. It shows how to use a GAN to learn representations of news articles (using the 20 Newsgroups dataset) in an unsupervised manner, and describes the TensorFlow code that can be used to replicate the work.

For more details read the full post or have a look at the code repository on Github.