Skip to main content

Improved Training of Mixture-of-Experts Language GANs

Yekun Chai (Baidu Inc.); Qiyue Yin (Institute of Automation, Chinese Academy of Sciences); Junge Zhang (CASIA)

  • SPS
    Members: Free
    IEEE Members: $11.00
    Non-members: $15.00
07 Jun 2023

Despite the dramatic success in image generation, Generative Adversarial Networks (GANs) still face great challenges in text generation. The difficulty in generator training arises from the limited representation capacity and uninformative learning signals obtained from the discriminator. In this work, we (1) first empirically show that the multi-generator approach is able to enhance the representation capacity of the generator for sequence GANs and (2) harness the Feature Statistics Alignment (FSA) paradigm to render fine-grained learning signals to advance the generator training. Specifically, FSA forces the mean statistics of the distribution of fake data to approach that of real samples as close as possible in the finite-dimensional feature space. Empirical study on synthetic and real benchmarks shows the superior performance in quantitative evaluation and demonstrates the effectiveness of our approach to adversarial text generation.

More Like This

  • SPS
    Members: Free
    IEEE Members: $11.00
    Non-members: $15.00
  • SPS
    Members: Free
    IEEE Members: $11.00
    Non-members: $15.00
  • SPS
    Members: Free
    IEEE Members: $11.00
    Non-members: $15.00