Adversarial Feature Matching for Text Generation

TL;DR

This paper introduces a novel adversarial training framework for text generation that matches high-dimensional feature distributions to produce realistic sentences, addressing GAN training challenges with discrete data.

Contribution

It proposes a feature matching approach using LSTM generator and CNN discriminator, improving stability and realism in text generation compared to standard GANs.

Findings

Achieves superior quantitative performance in text realism

Effectively alleviates mode collapse in adversarial training

Generates realistic-looking sentences

Abstract

The Generative Adversarial Network (GAN) has achieved great success in generating realistic (real-valued) synthetic data. However, convergence issues and difficulties dealing with discrete data hinder the applicability of GAN to text. We propose a framework for generating realistic text via adversarial training. We employ a long short-term memory network as generator, and a convolutional network as discriminator. Instead of using the standard objective of GAN, we propose matching the high-dimensional latent feature distributions of real and synthetic sentences, via a kernelized discrepancy metric. This eases adversarial training by alleviating the mode-collapsing problem. Our experiments show superior performance in quantitative evaluation, and demonstrate that our model can generate realistic-looking sentences.