摘要
本文旨在探索国产生成式人工智能(AIGC)大模型在信息科学论文初审阶段的应用效果,以期解决传统人工初审效率低、主观性强的问题。通过模拟编辑部的真实收稿情况,选取213篇论文作为试验稿件,利用国产大模型进行初审,并采用人工智能中常用的评价指标(准确率和精确率)进行综合评估。结果显示,文心一言和通义千问不仅能提供明确建议和详尽分析,而且在多个关键指标上表现较好,可显著减少人工工作量。最后,构建“AI+初审”工作流程,为实现初审工作的智能化转型提供实施方案。虽然AIGC不能完全替代人工初审,但在人机协同的初审模式下,国产大模型可辅助提高信息科学期刊的初审工作效率。
This article aims to explore the application effects of domestic generative artificial intelligence(AIGC)large models in the initial review stage of information science papers,in order to address the issues of low efficiency and strong subjectivity associated with traditional manual initial reviews.By simulating the real submission process of editorial departments,213 papers were selected as experimental manuscripts for initial review using domestic large models,and comprehensive evaluations were conducted using commonly used evaluation metrics in artificial intelligence.The results show that both WenxinYiyan and Tongyi Qianwen can not only provide clear suggestions and detailed analyses but also perform well on multiple key indicators,significantly reducing the workload of manual labor.Finally,the“AI+initial review”workflow constructed based on this provides a specific implementation plan for the intelligent transformation of initial review work.Although AIGC cannot completely replace manual initial reviews,in the human-machine collaborative initial review mode,domestic large models can assist in improving the efficiency of the initial review of information science journals.
作者
刘俏亮
刘东亮
张洁
LIU Qiaoliang;LIU Dongliang;ZHANG Jie(Journal of Jilin University(Information Science Edition),130012,Changchun,China)
出处
《编辑学报》
CSSCI
北大核心
2024年第5期548-552,共5页
Acta Editologica
关键词
生成式人工智能(AIGC)
国产大模型
论文初审
人机协同
“AI+初审”工作流程
artificial intelligence generated content(AICC)
domestic large model
paper preliminary review
human-computer collaboration
“AI+preliminary review”workflow
