ReAct
收藏arXiv2022-10-02 更新2024-06-21 收录
下载链接：
https://github.com/gtmdotme/ReAct
下载链接
链接失效反馈官方服务：
资源简介：
ReAct数据集由Adobe研究创建，专注于识别和分类学术文档中的评审评论。该数据集包含1250条已标注的评论，用于识别评论的行动性和类型。数据来源于OpenReview平台，通过众包方式进行标注，旨在帮助作者快速理解和响应评审意见。数据集的创建过程涉及从大量评论中筛选和分类，最终形成一个细粒度标注的资源，适用于文本分类研究，特别是学术文档的评审过程优化。
The ReAct dataset, developed by Adobe Research, focuses on identifying and classifying review comments within academic documents. It contains 1,250 annotated comments designed to recognize the actionability and category of review comments. Sourced from the OpenReview platform, these comments were annotated via crowdsourcing, with the goal of helping authors quickly understand and respond to reviewer feedback. The dataset creation process involved screening and categorizing a large volume of comments, ultimately forming a fine-grained annotated resource applicable to text classification research, especially for optimizing the academic document review workflow.
提供机构：
Adobe研究
创建时间：
2022-10-02
