WMT2020

WMT 2020 是第五届机器翻译会议共享任务中使用的数据集集合。该会议建立在一系列关于统计机器翻译的年度研讨会和会议的基础上。 会议共有十项共同任务： 新闻翻译任务， 生物医学翻译任务， 类似的语言翻译任务， 无监督且资源极少的翻译任务， 自动后期编辑任务， 一个度量任务（在给定参考翻译的情况下评估 MT 质量）， 质量评估任务（评估 MT 质量，无需访问任何参考）， 并行语料库过滤和对齐任务， 终身学习 MT 任务， 聊天翻译任务。

WMT 2020 is the dataset collection utilized in the shared tasks of the 5th Conference on Machine Translation. This conference builds upon a series of annual workshops and conferences focused on statistical machine translation. The conference features ten shared tasks in total: News Translation Task, Biomedical Translation Task, Similar-Language Translation Task, Unsupervised and Extremely Low-Resource Translation Task, Automatic Post-Editing Task, a Metrics Task (evaluating MT quality given reference translations), Quality Estimation Task (evaluating MT quality without access to any reference translations), Parallel Corpus Filtering and Alignment Task, Lifelong Learning MT Task, and Chat Translation Task.