Adversarial Feature Matching for Text Generation

Yizhe Zhang, Zhe Gan, Kai Fan, Zhi Chen, Ricardo Henao, Dinghan Shen, Lawrence Carin
Proceedings of the 34th International Conference on Machine Learning, PMLR 70:4006-4015, 2017.

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.

Cite this Paper


BibTeX
@InProceedings{pmlr-v70-zhang17b, title = {Adversarial Feature Matching for Text Generation}, author = {Yizhe Zhang and Zhe Gan and Kai Fan and Zhi Chen and Ricardo Henao and Dinghan Shen and Lawrence Carin}, booktitle = {Proceedings of the 34th International Conference on Machine Learning}, pages = {4006--4015}, year = {2017}, editor = {Precup, Doina and Teh, Yee Whye}, volume = {70}, series = {Proceedings of Machine Learning Research}, month = {06--11 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v70/zhang17b/zhang17b.pdf}, url = { http://proceedings.mlr.press/v70/zhang17b.html }, 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.} }
Endnote
%0 Conference Paper %T Adversarial Feature Matching for Text Generation %A Yizhe Zhang %A Zhe Gan %A Kai Fan %A Zhi Chen %A Ricardo Henao %A Dinghan Shen %A Lawrence Carin %B Proceedings of the 34th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2017 %E Doina Precup %E Yee Whye Teh %F pmlr-v70-zhang17b %I PMLR %P 4006--4015 %U http://proceedings.mlr.press/v70/zhang17b.html %V 70 %X 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.
APA
Zhang, Y., Gan, Z., Fan, K., Chen, Z., Henao, R., Shen, D. & Carin, L.. (2017). Adversarial Feature Matching for Text Generation. Proceedings of the 34th International Conference on Machine Learning, in Proceedings of Machine Learning Research 70:4006-4015 Available from http://proceedings.mlr.press/v70/zhang17b.html .

Related Material