TY - GEN
T1 - Learning to Select Relevant Knowledge for Neural Machine Translation
AU - Yang, Jian
AU - Wan, Juncheng
AU - Ma, Shuming
AU - Huang, Haoyang
AU - Zhang, Dongdong
AU - Yu, Yong
AU - Li, Zhoujun
AU - Wei, Furu
N1 - Publisher Copyright:
© 2021, Springer Nature Switzerland AG.
PY - 2021
Y1 - 2021
N2 - Most memory-based methods use encoded retrieved pairs as the translation memory (TM) to provide external guidance, but there still exist some noisy words in the retrieved pairs. In this paper, we propose a simple and effective end-to-end model to select useful sentence words from the encoded memory and incorporate them into the NMT model. Our model uses a novel memory selection mechanism to avoid the noise from similar sentences and provide external guidance simultaneously. To verify the positive influence of selected retrieved words, we evaluate our model on the single-domain dataset namely JRC-Acquis and multi-domain dataset comprised of existing benchmarks including WMT, IWSLT, JRC-Acquis, and OpenSubtitles. Experimental results demonstrate our method can improve the translation quality under different scenarios.
AB - Most memory-based methods use encoded retrieved pairs as the translation memory (TM) to provide external guidance, but there still exist some noisy words in the retrieved pairs. In this paper, we propose a simple and effective end-to-end model to select useful sentence words from the encoded memory and incorporate them into the NMT model. Our model uses a novel memory selection mechanism to avoid the noise from similar sentences and provide external guidance simultaneously. To verify the positive influence of selected retrieved words, we evaluate our model on the single-domain dataset namely JRC-Acquis and multi-domain dataset comprised of existing benchmarks including WMT, IWSLT, JRC-Acquis, and OpenSubtitles. Experimental results demonstrate our method can improve the translation quality under different scenarios.
KW - Neural machine translation
KW - Selective translation memory
UR - https://www.scopus.com/pages/publications/85118107605
U2 - 10.1007/978-3-030-88480-2_7
DO - 10.1007/978-3-030-88480-2_7
M3 - 会议稿件
AN - SCOPUS:85118107605
SN - 9783030884796
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 79
EP - 91
BT - Natural Language Processing and Chinese Computing - 10th CCF International Conference, NLPCC 2021, Proceedings
A2 - Wang, Lu
A2 - Feng, Yansong
A2 - Hong, Yu
A2 - He, Ruifang
PB - Springer Science and Business Media Deutschland GmbH
T2 - 10th CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2021
Y2 - 13 October 2021 through 17 October 2021
ER -