ACL 2021long50 citations

Consistency Regularization for Cross-Lingual Fine-Tuning

Bo Zheng, Li Dong, Shaohan Huang, Wenhui Wang, Zewen Chi, Saksham Singhal, Wanxiang Che, Ting Liu

Abstract

Fine-tuning pre-trained cross-lingual language models can transfer task-specific supervision from one language to the others. In this work, we propose to improve cross-lingual fine-tuning with consistency regularization. Specifically, we use example consistency regularization to penalize the prediction sensitivity to four types of data augmentations, i.e., subword sampling, Gaussian noise, code-switch substitution, and machine translation. In addition, we employ model consistency to regularize the models trained with two augmented versions of the same training set. Experimental results on the XTREME benchmark show that our method significantly improves cross-lingual fine-tuning across various tasks, including text classification, question answering, and sequence labeling.

BibTeX
@inproceedings{zheng-etal-2021-consistency,
    title = "Consistency Regularization for Cross-Lingual Fine-Tuning",
    author = "Zheng, Bo  and
      Dong, Li  and
      Huang, Shaohan  and
      Wang, Wenhui  and
      Chi, Zewen  and
      Singhal, Saksham  and
      Che, Wanxiang  and
      Liu, Ting  and
      Song, Xia  and
      Wei, Furu",
    editor = "Zong, Chengqing  and
      Xia, Fei  and
      Li, Wenjie  and
      Navigli, Roberto",
    booktitle = "Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)",
    month = aug,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.acl-long.264/",
    doi = "10.18653/v1/2021.acl-long.264",
    pages = "3403--3417"
}
Consistency Regularization for Cross-Lingual Fine-Tuning · ACL 2021