遇见数据集
官方服务:

资源简介:

该数据集名为UKP,包含了来自同一段落的组件对,同时为了增强基于对的分类,还额外加入了单一组件的“自对”。数据规模从22,000对扩大到了28,000对。该数据集的任务是进行论点挖掘。

This dataset, named UKP, includes component pairs sourced from the same paragraph. To improve pair-based classification performance, "self-pairs" made of individual components were additionally incorporated. The total number of pairs in the dataset has been expanded from 22,000 to 28,000. The core task of this dataset is argument mining.

二维码
社区交流群
二维码
科研交流群
商业服务