ZurichNLP/romansh-mt-evaluation
收藏资源简介:
该数据集包含对从德语翻译成六种罗曼什语变体机器翻译结果的人工评估结果。评估由相应罗曼什语方言的母语者以及专业语言学家进行。评估涵盖三个质量维度:- **文档准确性**,其中标注者评估完整文档翻译的充分性。- **片段准确性**,其中标注者为文档中的单个片段选择更准确的翻译。- **片段流畅性**,其中标注者评估单个翻译片段的流畅性。对于文档准确性任务,标注者会看到同一源文档的两个完整翻译。每个翻译获得0到6之间的分数,标注者还指示每个单独片段的优选翻译。对于片段流畅性任务,标注者独立于文档上下文评估翻译片段对,并为每个片段分配0到6之间的分数。
This dataset contains the results of a human evaluation of machine translations from German into the six Romansh varieties. The evaluations were carried out by native speakers of the respective Romansh idioms as well as professional linguists. The evaluation covers three quality dimensions: - **Document accuracy**, in which annotators assessed the adequacy of complete document translations. - **Segment accuracy**, in which annotators selected the more accurate translation for individual segments within the document. - **Segment fluency**, in which annotators assessed the fluency of individual translated segments. For the document accuracy task, annotators were presented with two complete translations of the same source document. Each translation received a score between 0 and 6, and annotators additionally indicated the preferred translation for each individual segment. For the segment fluency task, annotators evaluated pairs of translated segments independently of their document context and assigned each segment a score between 0 and 6.




