追溯科学融合的历史轨迹:AIGC赋能交叉科学测度研究  

Tracing the Historical Trajectory of Scientific Integration:AIGC Empowers Interdisciplinarity Measurement Research

作　　者：刘江峰 张冉 王东波[3,4] 裴雷 Liu Jiangfeng;Zhang Ran;Wang Dongbo;Pei Lei(Laboratory of Data Intelligence and Cross Innovation,Nanjing University,Nanjing 210023;School of Information Management,Nanjing University,Nanjing 210023;College of Information Management,Nanjing Agricultural University,Nanjing 210095;Research Center for Humanities and Social Computing,Nanjing Agricultural University,Nanjing 210095)

机构地区：[1]南京大学数据智能与交叉创新实验室,南京210023 [2]南京大学信息管理学院,南京210023 [3]南京农业大学信息管理学院,南京210095 [4]南京农业大学人文与社会计算研究中心,南京210095

出　　处：《图书情报工作》2024年第7期137-150,共14页Library and Information Service

摘　　要：[目的/意义]生成式大语言模型改变自然语言处理研究的范式,推动人工智能赋能社会科学研究的新潮流,为从文本深层语义特征角度量化计算人文社会科学学科的交叉与融合提供新思路。[方法/过程]使用ChatGPT模型对人文社会科学学术文献进行学科判别,基于小样本学习识别模型预测结果中的学科名知识实体,从期刊知识分散分布视角衡量多学科候选分类问题,将判别结果与文献所属期刊对应学科作比较分析,提出跨学科丰富度、跨学科密切度、主体度等指标结合跨学科度的跨学科性量化研究。[结果/结论]围绕AIGC赋能交叉科学测度研究,从学科归属问题的判断、生成式模型答案集中学科名的抽取、多学科候选问题的赋权、交叉科学内容性度量指标等几方面提出一套研究框架方法,实现从内容角度充分利用AIGC赋能社会科学研究,为进一步探索各社会科学研究的内在逻辑提供借鉴和参考。[Purpose/Significance]The generative large language model has changed the paradigm of natural language processing research,promoted the new trend of AI-enabled social science research,and provided new ideas for quantitatively calculating disciplinary crossover and integration of humanities and social science disciplines from the perspective of deep semantic features of texts.[Method/Process]In this paper,it used ChatGPT to discriminate the academic literature in humanities and social sciences.Based on few-shot learning,it recognized the discipline in the model prediction,measured the multidisciplinary classification from the perspective of the knowledge decentralized distribution,made a comparative analysis between the results and the corresponding disciplines of the journals.Finally,it put forward the indexes of cross-disciplinary richness,cross-disciplinary closeness,and subjectivity,combined with the interdisciplinary degree,interdisciplinary quantitative research.[Result/Conclusion]This paper focuses on AIGC-enabled interdisciplinarity measurement,proposes a set of research framework methods from the judgment of disciplinary attribution,the extraction of disciplinary names in the answer set of the generative model,the empowerment of multidisciplinary candidate problems,and interdisciplinarity content-based metric indicators,and realizes the AIGC-enabled social sciences research from the content.It provides a reference for the internal logic of social science research.

关 键 词：ChatGPT 大语言模型 交叉科学测度 布拉德福定律 人文社会科学 时间序列预测分析 

分 类 号：G353.1[文化科学—情报学] TP18[自动化与计算机技术—控制理论与控制工程]