arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.11239cs.CL
收藏

基于转换器的神经机器翻译

Character-level Transformer-based Neural Machine Translation

University of Antwerp(安特卫普大学);Computational Linguistics and Psycholinguistics Research Center(计算语言学与心理语言学研究中心);Antwerp Centre for Digital Humanities and Literary Criticism(安特卫普数字人文与文学批评中心)

Nikolay Banar, Walter Daelemans, Mike Kestemont

摘要:

神经机器翻译(NMT)现在通常应用于子字级别,使用字节对编码。一种有前途的替代方法专注于字符级翻译,这大大简化了NMT中的处理管道。然而,这种方法必须考虑相对较长的序列,使得训练过程过于昂贵。在本文中,我们讨论了一种新的,基于变压器的方法,我们比较,无论是在速度和质量的Transformer在子字和字符级别,以及以前开发的字符级模型。我们在WMT'15的4个语言对上评估了我们的模型:DE-EN,CS-EN,FI-EN和RU-EN。该模型可以在单个GPU上训练,比字符级的Transformer快34%,但其结果至少与之相当。此外,该模型在FI-EN中的性能优于子字级模型,在CS-EN中的性能接近。为了促进这一领域的进一步研究并缩小与子字级NMT的差距,我们公开了所有的代码和模型。

英文摘要:

Neural machine translation (NMT) is nowadays commonly applied at the subword level, using byte-pair encoding. A promising alternative approach focuses on character-level translation, which simplifies processing pipelines in NMT considerably. This approach, however, must consider relatively longer sequences, rendering the training process prohibitively expensive. In this paper, we discuss a novel, Transformer-based approach, that we compare, both in speed and in quality to the Transformer at subword and character levels, as well as previously developed character-level models. We evaluate our models on 4 language pairs from WMT'15: DE-EN, CS-EN, FI-EN and RU-EN. The proposed novel architecture can be trained on a single GPU and is 34% percent faster than the character-level Transformer; still, the obtained results are at least on par with it. In addition, our proposed model outperforms the subword-level model in FI-EN and shows close results in CS-EN. To stimulate further research in this area and close the gap with subword-level NMT, we make all our code and models publicly available.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑