BriefGPT.xyz
Apr, 2018
使用多个子词候选项改进神经网络翻译模型的子词正则化
Subword Regularization: Improving Neural Network Translation Models with Multiple Subword Candidates
HTML
PDF
Taku Kudo
TL;DR
介绍了一种利用subword segmentation的噪声来提高神经机器翻译鲁棒性的正则化方法,主要应用于低资源领域。
Abstract
subword units
are an effective way to alleviate the open vocabulary problems in
neural machine translation
(NMT). While sentences are usually converted into unique subword sequences,
→