DP-GAN: Diversity-Promoting Generative Adversarial Network for Generating Informative and Diversified Text

Jingjing Xu, Xuancheng Ren, Junyang Lin, Xu Sun
2/5/2018
cs.CL

Official code: lancopku/DPGAN (147 stars, last commit 2018-09-15).

Abstract

Existing text generation methods tend to produce repeated and "boring" expressions. To tackle this problem, we propose a new text generation model, called Diversity-Promoting Generative Adversarial Network (DP-GAN). The proposed model assigns low reward for repeatedly generated text and high reward for "novel" and fluent text, encouraging the generator to produce diverse and informative text. Moreover, we propose a novel language-model based discriminator, which can better distinguish novel text from repeated text without the saturation problem compared with existing classifier-based discriminators. The experimental results on review generation and dialogue generation tasks demonstrate that our model can generate substantially more diverse and informative text than existing baselines. The code is available at https://github.com/lancopku/DPGAN

View on arXivView PDF

Code Implementations(1)

lancopku/DPGANOfficial100%

Diversity-Promoting Generative Adversarial Network for Generating Informative and Diversified Text (EMNLP2018)

14736Shell, PythonFeb 5, 20188 years ago

Similar Papers

Reward-Augmented Decoding: Efficient Controlled Text Generation With a Unidirectional Reward Model

85% match1 repo452023

Long Text Generation via Adversarial Training with Leaked Information

83% match1 repo5732017

Air-Decoding: Attribute Distribution Reconstruction for Decoding-Time Controllable Text Generation

82% match1 repo92023

A Block Metropolis-Hastings Sampler for Controllable Energy-based Text Generation

82% match2023

Copy is All You Need

81% match1 repo1902023

ZYN: Zero-Shot Reward Models with Yes-No Questions for RLAIF

81% match1 repo342023

Cite this paper

@article{xu2018dpgan,
  title  = {DP-GAN: Diversity-Promoting Generative Adversarial Network for Generating Informative and Diversified Text},
  author = {Jingjing Xu and Xuancheng Ren and Junyang Lin and Xu Sun},
  year   = {2018},
  eprint = {1802.01345},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1802.01345}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1802.01345.svg)](https://codeofpaper.com/paper/1802.01345)

Discussion