We built an aligned corpus from English and Chinese news articles.
我们用英文和中文新闻文章构建了一个对齐语料库。
Using an aligned corpus, the model can learn which phrases are likely translations of each other, even when the word order differs.
借助对齐语料库,模型可以学习哪些短语彼此可能是翻译关系,即使语序不同也可以。