TimelineQA
收藏arXiv2023-06-02 更新2024-06-21 收录
下载链接：
https://github.com/facebookresearch/TimelineQA
下载链接
链接失效反馈官方服务：
资源简介：
TimelineQA是一个用于评估时间线问答系统的基准数据集，由Meta Cornell University创建。该数据集包含1.28亿条记录，覆盖从日常活动到重大生活事件的广泛情境。数据集通过合成生成，模拟不同年龄、性别和教育背景的虚拟人物的生活记录。TimelineQA旨在推动问答技术在处理复杂时间线和结构化数据方面的进步，特别适用于开发个人助理应用，以帮助用户在特定情境下获取有用的生活信息。
TimelineQA is a benchmark dataset for evaluating timeline-focused question answering systems, created by Meta and Cornell University. This dataset contains 128 million records covering a wide range of scenarios from daily activities to major life events. It is synthetically generated, simulating life records of virtual individuals with diverse ages, genders and educational backgrounds. TimelineQA aims to advance the development of question answering technologies in handling complex timelines and structured data, and is particularly suitable for developing personal assistant applications to help users retrieve useful life information in specific scenarios.
提供机构：
Meta Cornell University
创建时间：
2023-06-02
