ODSum

ODSum是一个针对开放域多文档摘要任务的新型数据集，由浙江大学、耶鲁大学和艾伦人工智能研究所合作创建。该数据集包含1422条数据，主要来源于SQuALITY和QMSum两个现有数据集，通过规则方法处理形成。ODSum数据集的特点是文档索引相互依赖且经常相互关联，适用于测试多文档摘要系统的性能。该数据集的应用领域主要集中在提高信息检索和多文档摘要的效率，帮助用户从大量文档中快速提取关键信息，无需逐一阅读。

ODSum is a novel dataset for the open-domain multi-document summarization task, jointly created by Zhejiang University, Yale University, and the Allen Institute for AI. It comprises 1422 data instances, which are primarily derived from two existing datasets, SQuALITY and QMSum, and processed and formulated via rule-based methods. The ODSum dataset features interdependent and often interrelated document indexes, making it ideal for evaluating the performance of multi-document summarization systems. Its core application scenarios center on enhancing the efficiency of information retrieval and multi-document summarization, enabling users to quickly extract key information from a vast number of documents without having to read each one individually.