MRP 2019 and 2020 Shared Tasks Sample

该数据集包含了87个句子，其中在UCCA和AMR之间有重叠的部分。具体来说，前17个句子被用于训练，而剩余的70个则用于测试。此外，该数据集还包括了为分析而修改的图形表示，突显了AMR和UCCA图形在结构上的差异。该数据集的任务是从AMR重写为UCCA的图形重写。

This dataset comprises 87 sentences that have overlapping representations between UCCA and AMR. Specifically, the first 17 sentences are utilized for model training, while the remaining 70 are set aside for testing. Additionally, the dataset also includes graph representations modified for analytical purposes, which underscore the structural disparities between AMR and UCCA graphs. The core task of this dataset is graph rewriting from AMR to UCCA.