RaFoLa

RaFoLa数据集是由曼彻斯特大学计算机科学系创建，专注于检测强制劳动指标的英文语料库。该数据集包含989篇从专业数据源检索的新闻文章，每篇文章根据国际劳工组织（ILO）定义的风险指标进行标注。数据集不仅标注了强制劳动的分类标签，还标注了支持这些标签决策的文本片段，即‘理性标注’。此数据集旨在支持多类别和多标签文本分类的研究，特别是在解释性方面，为开发对抗现代奴役的人工智能工具提供数据支持。

The RaFoLa dataset is an English-language corpus developed by the Department of Computer Science at the University of Manchester, specifically for detecting forced labor indicators. It comprises 989 news articles retrieved from professional data sources, with each article annotated in accordance with the risk indicators defined by the International Labour Organization (ILO). The dataset provides not only classification labels for forced labor, but also annotates the text segments that justify the decisions behind these labels, which are termed "rational annotations". This dataset is intended to support research on multi-class and multi-label text classification, especially in the domain of interpretability, and offers data support for the development of AI tools aimed at combating modern slavery.