launch/gov_report
收藏资源简介:
GovReport数据集由美国政府研究机构(如国会研究服务局和美国政府问责办公室)编写的报告及其摘要组成。与其他长文档摘要数据集相比,该数据集的摘要和文档更长,需要更多的上下文阅读来覆盖需要总结的关键词。数据集包含三个配置:plain_text(默认)、plain_text_with_recommendations和structure,分别对应不同的数据格式。数据集的语言为英语,大小为10K到100K之间,许可证为CC BY 4.0。
The GovReport dataset comprises reports and their abstracts authored by U.S. government research institutions including the Congressional Research Service and the Government Accountability Office. Compared with other long-document summarization datasets, this dataset features longer documents and their corresponding abstracts, requiring more contextual reading to cover all key terms that need to be summarized. The dataset includes three configurations: plain_text (default), plain_text_with_recommendations, and structure, each corresponding to distinct data formats. The dataset is in English, with a size ranging from 10K to 100K, and is licensed under CC BY 4.0.
数据集概述
数据集名称
- GovReport
数据集摘要
- 内容来源:政府报告数据集包含由政府研究机构(如国会研究服务部和美国政府问责局)编写的报告及其相关摘要。
- 特点:与其他长文档摘要数据集相比,政府报告数据集具有更长的摘要和文档,需要更多的上下文阅读来涵盖要摘要的关键词。
版本信息
- 版本:1.0.1(默认),去除多余的空格;1.0.0,原始论文中使用的数据集。
支持的任务
- 任务:摘要生成
语言
- 语言:英语
数据集结构
- 配置:
- plain_text(默认):原始论文中使用的文本到文本摘要设置。
- plain_text_with_recommendations:包含“GAO建议”的文本到文本摘要设置。
- structure:包含部分结构的数据。
数据实例
- 示例结构:
- plain_text & plain_text_with_recommendations:包含
id,document,summary字段。 - structure:包含
id,document_sections(包含title,paragraphs,depth),summary_sections(包含title,paragraphs)字段。
- plain_text & plain_text_with_recommendations:包含
数据字段
- plain_text & plain_text_with_recommendations:
id: 字符串类型。document: 字符串类型。summary: 字符串类型。
- structure:
id: 字符串类型。document_sections: 字典类型,包含title,paragraphs,depth列表。summary_sections: 字典类型,包含title,paragraphs列表。
数据分割
- 训练集:17519
- 验证集:974
- 测试集:973
许可证
- 许可证:CC BY 4.0
引用信息
@inproceedings{huang-etal-2021-efficient, title = "Efficient Attentions for Long Document Summarization", author = "Huang, Luyang and Cao, Shuyang and Parulian, Nikolaus and Ji, Heng and Wang, Lu", booktitle = "Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies", month = jun, year = "2021", address = "Online", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2021.naacl-main.112", doi = "10.18653/v1/2021.naacl-main.112", pages = "1419--1436", abstract = "The quadratic computational and memory complexities of large Transformers have limited their scalability for long document summarization. In this paper, we propose Hepos, a novel efficient encoder-decoder attention with head-wise positional strides to effectively pinpoint salient information from the source. We further conduct a systematic study of existing efficient self-attentions. Combined with Hepos, we are able to process ten times more tokens than existing models that use full attentions. For evaluation, we present a new dataset, GovReport, with significantly longer documents and summaries. Results show that our models produce significantly higher ROUGE scores than competitive comparisons, including new state-of-the-art results on PubMed. Human evaluation also shows that our models generate more informative summaries with fewer unfaithful errors.", }




