阿拉伯语事实核查数据集

阿拉伯语事实核查数据集是由麻省理工学院计算机科学与人工智能实验室创建的，旨在解决互联网上大量虚假新闻和谣言的问题。该数据集包含422条经过验证的阿拉伯语新闻标题，分为真实和虚假两类。数据集的创建过程涉及从VERIFY和REUTERS网站提取新闻标题，并通过人工和自动方法进行筛选和标注。该数据集的应用领域主要集中在自动事实核查和立场检测，通过分析文档对特定声明的立场来预测声明的真实性。

The Arabic Fact-Checking Dataset was developed by the MIT Computer Science and Artificial Intelligence Laboratory (MIT CSAIL) to address the widespread proliferation of fake news and rumors on the Internet. This dataset contains 422 verified Arabic news headlines, categorized into two classes: genuine and fake. The dataset construction process involved extracting news headlines from the VERIFY and REUTERS websites, followed by screening and annotation via both manual and automated methods. Its primary application domains focus on automated fact-checking and stance detection, where the veracity of a given claim is predicted by analyzing the stance of documents towards that specific claim.