Microsoft’s Submission to the WMT2018 News Translation Task: How I Learned to Stop Worrying and Love the Data
Marcin Junczys-Dowmunt Microsoft 1 Microsoft Way Redmond, WA 98121, USA
Abstract
This paper describes the Microsoft submission to the WMT2018 news translation shared task. We participated in one language direction – English-German. Our system follows current best-practice and combines state-of-the-art models with new data filtering (dual conditional cross-entropy filtering) and sentence weighting methods. We trained fairly standard Transformer-big models with an updated version of Edinburgh’s training scheme for WMT2017 and experimented with different filtering schemes for Paracrawl. According to automatic metrics (BLEU) we reached the highest score for this subtask with a nearly 2 BLEU point margin over the next strongest system. Based on human evaluation we ranked first among constrained systems. We believe this is mostly caused by our data filtering/weighting regime.
中文速览
微软在WMT2018英德机器翻译评测中拿下自动指标和人工评测双料第一，领先第二名近2个BLEU分，而关键不在于模型有多新颖，而在于他们如何处理海量噪声数据。他们提出了一种叫"双向条件交叉熵过滤（dual conditional cross-entropy filtering）"的方法：同时训练两个互为逆方向的翻译模型，用两个方向的翻译概率是否一致、以及平均困惑度来给每个句对打质量分，从而把3200万条爬取的Paracrawl平行语料筛选到800万条高质量句子，并按分数对训练样本加权。在此基础上，他们使用标准的Transformer-big模型、结合回译数据增强和Transformer风格语言模型集成，最终比只用原始WMT数据提升了1至2.5个BLEU。这项工作说明，在神经机器翻译已高度成熟的今天，数据质量的精细把控往往比模型创新更能决定最终成绩。
原文 arXiv:1809.00196;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1809.00196v1