NEWSCLAIMS

NEWSCLAIMS是由伊利诺伊大学厄巴纳-香槟分校创建的一个新型基准数据集，专注于新闻领域的属性感知声明检测。该数据集包含889个声明，分布在143篇新闻文章中，旨在评估声明检测系统在处理新兴场景中的性能，特别是针对训练数据有限或无训练数据的情况。NEWSCLAIMS不仅关注声明本身，还扩展到提取与每个声明相关的额外属性，如声明者及其主要对象。数据集的应用领域主要集中在新闻理解和自动事实核查，旨在解决新闻中错误信息和虚假信息的识别问题。

NEWSCLAIMS is a novel benchmark dataset developed by the University of Illinois Urbana-Champaign, dedicated to attribute-aware claim detection in the news domain. It comprises 889 claims distributed across 143 news articles, and is intended to evaluate the performance of claim detection systems in emerging scenarios, especially when training data is limited or entirely unavailable. Beyond centering on the claims themselves, NEWSCLAIMS also enables the extraction of supplementary attributes related to each individual claim, such as the claimant and its core subjects. The primary application scenarios of this dataset focus on news understanding and automated fact-checking, with the objective of resolving the issue of identifying misinformation and disinformation in news content.