BriefGPT.xyz
Nov, 2021
基于神经度量的最小贝叶斯风险解码：高质量而非高模型概率
Minimum Bayes Risk Decoding with Neural Metrics of Translation Quality
HTML
PDF
Markus Freitag, David Grangier, Qijun Tan, Bowen Liang
TL;DR
本篇论文研究神经机器翻译，提出使用最小贝叶斯风险解码优化翻译质量度量作为一种替代推论策略来优化自动翻译质量度量，实验结果表明，使用BLEURT作为质量度量指标产生的质量比传统Beam-search输出更好。
Abstract
This work applies Minimum Bayes Risk (MBR) decoding to optimize diverse
automated metrics
of
translation quality
. Automatic metrics in machine translation have made tremendous progress recently. In particular, ne
→