摘要
随着生成式人工智能的快速发展,其在信息检索和整合分析方面的优势为科技查新工作带来新的机遇。然而,在当前科技查新工作中,大模型的应用范围仍较为有限,普及性相对较低。为推动生成式人工智能在科技查新中的应用,以当前互联网中可获取和使用的9个开源国产大模型为研究工具,以医学类查新课题为研究对象,针对查新流程中的各个环节设计问题类型,综合使用文本比较法、对比分析法等方法,采用层次分析法量化查新各环节,研究大模型在各环节适用度,提出以Kimi为主要辅助工具嵌入查新的各环节,其他模型根据自身强项灵活搭配的人机协作新应用模式。该模式不仅能提高科技查新的效率和准确性,还能降低人力成本,为科技查新工作者提供更高效、更可靠的解决方案。
With the rapid development of generative artificial intelligence,its advantages in information retrieval and integrated analysis have brought new opportunities to the work of scientific and technical novelty research.Large models,however,still face a limited application scope and low popularity in current scientific and technical novelty search practices.To promote the application of generative artificial intelligence in scientific and technical novelty search,this study takes nine open-source domestically developed large models accessible and usable on the internet as research tools,with medical novelty search topics as the research subject.Then it designs question types for each link in the novelty search process,and explores the applicability of large models in each link with a combination of text comparison,comparative analysis and other methods,and by adopting the Analytic Hierarchy Process(AHP)to quantify each link of novelty search.Finally,it proposes a new human-machine collaboration application mode,in which Kimi serves as the main auxiliary tool embedded in all links of novelty retrieval,while other models are flexibly combined based on their respective strengths.This mode can improve the efficiency and accuracy of scientific and technical novelty search,and reduce labor costs,thereby providing a more efficient and reliable solution for professionals in this field.
作者
林淑凤
宋学芳
陈爱群
Lin Shufeng;Song Xuefang;Chen Aiqun
出处
《科技广场》
2025年第3期14-22,共9页
Science and technology Square
基金
武汉大学图书馆青年馆员科研引导基金项目“生成式人工智能在科技查新中的应用研究”(项目编号:2023-YB-02)。
关键词
科技查新
人工智能
大模型
AHP
Scientific and Technical Novelty Search
Artificial Intelligence
Large Models
AHP
