基于提示学习混合模型的学术论文自动分类研究OACSSCICSTPCD
Automatic Classification of Academic Papers Based on Mixed Model of Prompt Learning
学术论文分类在知识管理、学术交流、研究导向和学术评估等方面都具有重要的意义.基于深度学习模型构建学术论文自动分类系统,相较于现有的文本分类方法,该系统融合提示学习思想,可较好地缩小预训练模型与下游任务的差距.结果表明,该系统较好地提高了文本分类性能和规范性,为科研工作者提供了更好的管理、利用和挖掘信息的方式.
The classification of academic papers is of great significance in knowledge management,academic exchange,research orientation,and academic evaluation.This paper builds an automatic classification system for academic papers based on a deep learning model.Compared with existing text classification methods,this system integrates the idea of prompt learning and better bridges the gap between the pre-training model and downstream tasks.The results show that this system can better improve text classification performance and standardization level,and provides better ways for researchers to manage,utilize,and mine information.
刘爱琴;贺玉斌;马茹茹
山西大学经济与管理学院,太原 030006
学术论文提示学习自动分类
Academic PaperPrompt LearningAutomatic Classification
《数字图书馆论坛》 2024 (004)
74-80 / 7
本研究得到国家社会科学基金项目"中文学术领域命名实体的知识图谱构建研究"(编号:18BTQ072)资助.
评论