GUM
收藏arXiv2025-09-30 收录
下载链接：
https://gucorpling.org/gum/
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一个涵盖多种体裁的语篇树库，其中包括访谈、新闻故事、学术论文以及对话等12种不同类型。与RST-DT相似，该数据集并没有明确的人工标注的主题段落边界。其任务是进行话语解析。
This dataset is a discourse treebank that covers 12 distinct types of discourse, including interviews, news stories, academic papers, dialogues and others. Similar to RST-DT, this dataset does not have explicitly manually annotated topic paragraph boundaries, and its core task is discourse parsing.
提供机构：
GUM
