OPENASP
收藏资源简介:
OPENASP是一个多文档开放方面基础摘要的基准数据集,由巴伊兰大学创建。该数据集通过一种新颖且成本有效的标注协议,从现有的通用多文档摘要数据集中衍生出开放方面数据集。OPENASP包含1310个方面基础摘要,分为训练、验证和测试集,适用于任务的方法学建模。数据集的应用领域旨在解决真实场景中用户特定的信息需求,特别是在需要针对特定方面的摘要时。
OPENASP is a benchmark dataset for multi-document open aspect-based summarization, developed by Bar-Ilan University. This dataset is derived from existing general multi-document summarization datasets via a novel and cost-effective annotation protocol to generate open-aspect summarization samples. OPENASP contains 1,310 aspect-based summarization instances, which are partitioned into training, validation, and test sets, making it suitable for methodological modeling of the corresponding task. The application scope of this dataset is designed to address user-specific information demands in real-world scenarios, especially when targeted aspect-based summarization is needed.

- 1OpenAsp: A Benchmark for Multi-document Open Aspect-based Summarization巴伊兰大学 · 2023年



