BriefGPT.xyz
Feb, 2022
通过最小贝叶斯风险解码识别机器翻译指标的缺陷：以COMET为例的案例研究
Identifying Weaknesses in Machine Translation Metrics Through Minimum Bayes Risk Decoding: A Case Study for COMET
HTML
PDF
Chantal Amrhein, Rico Sennrich
TL;DR
本论文探讨神经测量在机器翻译评估中的作用，提出最小贝叶斯风险解码策略，以消除评估的偏见，发现机器翻译中存在数字和命名实体的偏见，提供代码和数据以便未来的研究。
Abstract
neural metrics
have achieved impressive correlation with human judgements in the evaluation of
machine translation
systems, but before we can safely optimise towards such metrics, we should be aware of (and ideal
→