Modeling documents with Generative Adversarial Networks
arXiv:1612.09122
Abstract
This paper describes a method for using Generative Adversarial Networks to learn distributed representations of natural language documents. We propose a model that is based on the recently proposed Energy-Based GAN, but instead uses a Denoising Autoencoder as the discriminator network. Document representations are extracted from the hidden layer of the discriminator and evaluated both quantitatively and qualitatively.
References in corpus (7)
- Auto-Encoding Variational Bayes
- Stochastic Backpropagation and Approximate Inference in Deep Generative Models
- DRAW: A Recurrent Neural Network For Image Generation
- Energy-based Generative Adversarial Network
- Adversarial Feature Learning
- Adversarially Learned Inference
- Autoencoding beyond pixels using a learned similarity metric
Cited by in corpus (7)
- Learning Graph Embedding with Adversarial Training Methods
- Generating Multi-label Discrete Patient Records using Generative Adversarial Networks
- Adversarial Network Embedding
- eCommerceGAN : A Generative Adversarial Network for E-commerce
- Adversarially Regularized Graph Attention Networks for Inductive Learning on Partially Labeled Graphs
- Explaining Away Syntactic Structure in Semantic Document Representations
- Generative Adversarial Nets for Multiple Text Corpora