ColdGANs: Taming Language GANs with Cautious Sampling Strategies

Abstract

Training regimes based on Maximum Likelihood Estimation (MLE) suffer from known limitations, often leading to poorly generated text sequences that lack of coherence, factualness, and are prone to repetitions. At the root of these limitations is the mismatch between training and inference, i.e. the so-called exposure bias. Another problem lies in considering only the reference text as correct, while in practice several alternative formulations could be as good.

PDF NeurIPS Semantic Scholar

Cite

Text

Scialom et al. "ColdGANs: Taming Language GANs with Cautious Sampling Strategies." Neural Information Processing Systems, 2020.

Markdown

[Scialom et al. "ColdGANs: Taming Language GANs with Cautious Sampling Strategies." Neural Information Processing Systems, 2020.](https://mlanthology.org/neurips/2020/scialom2020neurips-coldgans/)

BibTeX

@inproceedings{scialom2020neurips-coldgans,
  title     = {{ColdGANs: Taming Language GANs with Cautious Sampling Strategies}},
  author    = {Scialom, Thomas and Dray, Paul-Alexis and Lamprier, Sylvain and Piwowarski, Benjamin and Staiano, Jacopo},
  booktitle = {Neural Information Processing Systems},
  year      = {2020},
  url       = {https://mlanthology.org/neurips/2020/scialom2020neurips-coldgans/}
}