BriefGPT.xyz
May, 2021
神经机器翻译中最小贝叶斯风险解码的特性理解
Understanding the Properties of Minimum Bayes Risk Decoding in Neural Machine Translation
HTML
PDF
Mathias Müller, Rico Sennrich
TL;DR
本论文探讨神经机器翻译中的偏差以及在域偏移和样本干扰下的弱点，并找到应用最小贝叶斯风险解码对抗这些问题的方案。结果表明，这个方法虽然仍有长度和频率偏差，但同样增加了模型的鲁棒性，对样本干扰和域偏移具有更好的适应能力
Abstract
neural machine translation
(NMT) currently exhibits biases such as producing translations that are too short and overgenerating frequent words, and shows poor
robustness
to copy noise in training data or
→