TY - GEN
T1 - Adversarial text generation via feature-mover's distance
AU - Chen, Liqun
AU - Dai, Shuyang
AU - Tao, Chenyang
AU - Shen, Dinghan
AU - Gan, Zhe
AU - Zhang, Haichao
AU - Zhang, Yizhe
AU - Carin, Lawrence
N1 - Generated from Scopus record by KAUST IRTS on 2021-02-09
PY - 2018/1/1
Y1 - 2018/1/1
N2 - Generative adversarial networks (GANs) have achieved significant success in generating real-valued data. However, the discrete nature of text hinders the application of GAN to text-generation tasks. Instead of using the standard GAN objective, we propose to improve text-generation GAN via a novel approach inspired by optimal transport. Specifically, we consider matching the latent feature distributions of real and synthetic sentences using a novel metric, termed the feature-mover's distance (FMD). This formulation leads to a highly discriminative critic and easy-to-optimize objective, overcoming the mode-collapsing and brittle-training problems in existing methods. Extensive experiments are conducted on a variety of tasks to evaluate the proposed model empirically, including unconditional text generation, style transfer from non-parallel text, and unsupervised cipher cracking. The proposed model yields superior performance, demonstrating wide applicability and effectiveness.
AB - Generative adversarial networks (GANs) have achieved significant success in generating real-valued data. However, the discrete nature of text hinders the application of GAN to text-generation tasks. Instead of using the standard GAN objective, we propose to improve text-generation GAN via a novel approach inspired by optimal transport. Specifically, we consider matching the latent feature distributions of real and synthetic sentences using a novel metric, termed the feature-mover's distance (FMD). This formulation leads to a highly discriminative critic and easy-to-optimize objective, overcoming the mode-collapsing and brittle-training problems in existing methods. Extensive experiments are conducted on a variety of tasks to evaluate the proposed model empirically, including unconditional text generation, style transfer from non-parallel text, and unsupervised cipher cracking. The proposed model yields superior performance, demonstrating wide applicability and effectiveness.
UR - http://www.scopus.com/inward/record.url?scp=85064807771&partnerID=8YFLogxK
M3 - Conference contribution
SP - 4666
EP - 4677
BT - Advances in Neural Information Processing Systems
PB - Neural information processing systems foundation
ER -