NLPCONTRIBUTIONGRAPH

NLPCONTRIBUTIONGRAPH数据集由TIB Leibniz信息中心科学技术创建，旨在从自然语言处理（NLP）学术文章中结构化贡献信息。数据集包含900条专注于贡献的句子，4,702条专注于贡献信息的短语，以及2,980个表面结构化的三元组。创建过程采用两阶段注释方法，首先定义方案，然后规范化图形模型。该数据集主要应用于学术知识图谱，帮助研究人员快速追踪学术进展，减少搜索全文中贡献信息的时间和认知劳动。

The NLPCONTRIBUTIONGRAPH dataset was created by TIB Leibniz Information Centre for Science and Technology. Its core objective is to structurally extract contribution-related information from natural language processing (NLP) academic papers. The dataset includes 900 contribution-focused sentences, 4,702 contribution-information-focused phrases, and 2,980 surface-structured triples. The construction of this dataset adopts a two-stage annotation workflow: first defining the annotation schema, then standardizing the graphical model. This dataset is primarily utilized in academic knowledge graphs, enabling researchers to rapidly track academic progress while reducing the time and cognitive labor expended when searching for contribution-related information in full-text articles.