ICML 2018oral75 citations

Model-Level Dual Learning

Yingce Xia, Xu Tan, Fei Tian, Tao Qin, Nenghai Yu, Tie-Yan Liu

Abstract

Many artificial intelligence tasks appear in dual forms like English$\leftrightarrow$French translation and speech$\leftrightarrow$text transformation. Existing dual learning schemes, which are proposed to solve a pair of such dual tasks, explore how to leverage such dualities from data level. In this work, we propose a new learning framework, model-level dual learning, which takes duality of tasks into consideration while designing the architectures for the primal/dual models, and ties the model parameters that playing similar roles in the two tasks. We study both symmetric and asymmetric model-level dual learning. Our algorithms achieve significant improvements on neural machine translation and sentiment analysis.

BibTeX
@InProceedings{pmlr-v80-xia18a,
  title = 	 {Model-Level Dual Learning},
  author =       {Xia, Yingce and Tan, Xu and Tian, Fei and Qin, Tao and Yu, Nenghai and Liu, Tie-Yan},
  booktitle = 	 {Proceedings of the 35th International Conference on Machine Learning},
  pages = 	 {5383--5392},
  year = 	 {2018},
  editor = 	 {Dy, Jennifer and Krause, Andreas},
  volume = 	 {80},
  series = 	 {Proceedings of Machine Learning Research},
  month = 	 {10--15 Jul},
  publisher =    {PMLR},
  pdf = 	 {http://proceedings.mlr.press/v80/xia18a/xia18a.pdf},
  url = 	 {https://proceedings.mlr.press/v80/xia18a.html},
  abstract = 	 {Many artificial intelligence tasks appear in dual forms like English$\leftrightarrow$French translation and speech$\leftrightarrow$text transformation. Existing dual learning schemes, which are proposed to solve a pair of such dual tasks, explore how to leverage such dualities from data level. In this work, we propose a new learning framework, model-level dual learning, which takes duality of tasks into consideration while designing the architectures for the primal/dual models, and ties the model parameters that playing similar roles in the two tasks. We study both symmetric and asymmetric model-level dual learning. Our algorithms achieve significant improvements on neural machine translation and sentiment analysis.}
}
Model-Level Dual Learning · ICML 2018