SPACE-IDEAS
收藏arXiv2024-03-26 更新2024-06-21 收录
下载链接：
https://github.com/expertailab/SPACE-IDEAS
下载链接
链接失效反馈官方服务：
资源简介：
SPACE-IDEAS数据集由Expert.ai的语言技术研究实验室创建，专注于空间创新领域的显著信息检测。该数据集包含176个非保密的想法，总计1733个句子，49420个单词，主要来源于Open Space Innovation Platform。数据集的创建过程包括人工标注和使用大型生成语言模型进行扩展标注。SPACE-IDEAS的应用领域包括帮助读者聚焦于文本中最重要部分，解决信息过载问题，并支持多任务学习以训练更优的分类器。
The SPACE-IDEAS dataset was created by the Language Technology Research Laboratory at Expert.ai, focusing on salient information detection in the domain of space innovation. This dataset contains 176 non-confidential ideas, totaling 1,733 sentences and 49,420 words, and is primarily sourced from the Open Space Innovation Platform. The dataset creation process includes manual annotation and extended annotation using large generative language models. The application scenarios of SPACE-IDEAS include helping readers focus on the most critical parts of texts, addressing the problem of information overload, and supporting multi-task learning to train more optimal classifiers.
提供机构：
Expert.ai, 语言技术研究实验室
创建时间：
2024-03-26
