SciXGen
收藏资源简介:
SciXGen是由日本国立先进工业科学技术研究所创建的大规模科学论文数据集,包含205,304篇论文,每篇论文均附有完整的参考文献和广泛使用的对象(如表格、图表、算法)。数据集旨在推动科学领域的上下文感知文本生成研究,通过提供全面标注的数据支持描述和段落生成。SciXGen的应用领域包括自动生成科学论文,特别是在需要完整上下文信息的文本生成任务中,如表格和图表的描述生成,以及基于引用的段落生成。
SciXGen is a large-scale scientific paper dataset developed by the National Institute of Advanced Industrial Science and Technology (AIST) of Japan. It contains 205,304 papers, each accompanied by complete reference lists and widely utilized academic elements including tables, figures, and algorithms. This dataset is designed to promote research on context-aware text generation in the scientific domain, by providing comprehensively annotated data to support description and paragraph generation tasks. Application scenarios of SciXGen cover automatic scientific paper generation, particularly text generation tasks requiring complete contextual information, such as generating descriptions for tables and figures, as well as citation-based paragraph generation.



