Transfer Learning for Sequence Generation: from Single-source to Multi-source

Xuancheng Huang, Jingfang Xu, Maosong Sun, Yang Liu


Abstract
Multi-source sequence generation (MSG) is an important kind of sequence generation tasks that takes multiple sources, including automatic post-editing, multi-source translation, multi-document summarization, etc. As MSG tasks suffer from the data scarcity problem and recent pretrained models have been proven to be effective for low-resource downstream tasks, transferring pretrained sequence-to-sequence models to MSG tasks is essential. Although directly finetuning pretrained models on MSG tasks and concatenating multiple sources into a single long sequence is regarded as a simple method to transfer pretrained models to MSG tasks, we conjecture that the direct finetuning method leads to catastrophic forgetting and solely relying on pretrained self-attention layers to capture cross-source information is not sufficient. Therefore, we propose a two-stage finetuning method to alleviate the pretrain-finetune discrepancy and introduce a novel MSG model with a fine encoder to learn better representations in MSG tasks. Experiments show that our approach achieves new state-of-the-art results on the WMT17 APE task and multi-source translation task using the WMT14 test set. When adapted to document-level translation, our framework outperforms strong baselines significantly.
Anthology ID:
2021.acl-long.446
Volume:
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:
August
Year:
2021
Address:
Online
Venues:
ACL | IJCNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
5738–5750
Language:
URL:
https://aclanthology.org/2021.acl-long.446
DOI:
10.18653/v1/2021.acl-long.446
Bibkey:
Cite (ACL):
Xuancheng Huang, Jingfang Xu, Maosong Sun, and Yang Liu. 2021. Transfer Learning for Sequence Generation: from Single-source to Multi-source. In 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), pages 5738–5750, Online. Association for Computational Linguistics.
Cite (Informal):
Transfer Learning for Sequence Generation: from Single-source to Multi-source (Huang et al., ACL 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.acl-long.446.pdf
Video:
 https://aclanthology.org/2021.acl-long.446.mp4
Code
 THUNLP-MT/TRICE