基于NLP的辅助审稿系统设计与开发实践

Design and development practice of NLP-assisted review system

[目的]利用自然语言处理(Natural Language Processing,NLP)技术设计研发辅助审稿系统,提高编辑审稿工作的效率和精度。[方法]分析在科技论文审稿过程中各环节的审查关键点,梳理分析辅助审稿系统的需求,从搭建审稿专家画像、文本挖掘和审稿专家推荐3个方面入手,基于已有审稿专家库数据信息,设计并开发出一套基于NLP的辅助审稿系统。[结果]该系统建立了较为完善的审稿专家画像库,并基于该专家画像库,实现初步送审意见生成、相关论文匹配、审稿专家推荐等辅助功能。[结论]将人工智能领域的技术应用到审稿系统构建是未来的发展方向之一,能够在一定程度上减少编辑的工作量,同时能够更精准地匹配“小同行”,给予作者更专业的服务,从而更有效地促进学术交流。

[Purposes]This study aims to design and develop an auxiliary review system using natural language processing(NLP)technology to improve the efficiency and accuracy of editing and review.[Methods]We analyzed the key points of review in each stage of the scientific paper review to sort out the requirements of the auxiliary review system.Starting with review expert profile building,text mining,and review expert recommendation,we designed an NLP-assisted review system based on the existing review expert database's information.[Findings]A relatively complete database of review experts has been established in the review system.Based on this database,auxiliary functions such as preliminary review opinion generation,relevant article matching,and reviewers'recommendations are achieved.[Conclusions]Applying artificial intelligence to the establishment of review systems is a way to develop in the future,as this technology can reduce the heavy workload of editors and provide relevant review experts more accurately,offering authors more professional services and promoting academic exchanges effectively.