Abstract: With the aid of interdisciplinary index and science overlay technology, this paper takes 197 394 papers in the field of artificial intelligence in Web of science from 1964 to 2015 as the research sample and explores the interdisciplinary evolution process of artificial intelligence in the past 50 years including the identification of core disciplines and the analysis of interdisciplinary evolution process. The results show that the core disciplines of the artificial intelligence field are mainly concentrated in computer science, engineering electrical, medical and imaging, mathematics and other disciplines, and these core disciplines play key role in the development of the interdisciplinary evolution process in the field of artificial intelligence; with the passage of time, the interdisciplinary features of artificial intelligence field are becoming more and more obvious, and the interdisciplinary evolution process of this field has experienced embryonic period, rapid development period, stable period and slow-down period. The results also show that the scope of interdisciplinary begins with the adjacent core disciplines and spread to the more distant potential disciplines and the marginal disciplines in recent years, and some new cross disciplines may appear with the gradual appearance of these disciplines.

Key words: artificial intelligence, interdisciplinary, Rao-Stirling index, Brillouin index, science overlay

摘要： 以Web of Science中人工智能领域1964年至2015年包含的197 394篇文献为研究样本，借助于跨学科指数和科学叠加图技术，对近五十年来国际上人工智能领域的跨学科演变过程进行了探究，包括核心学科的识别及跨学科演变过程的分析。研究结果表明探寻到的核心学科主要集中在计算机类、工程电气类、医学及成像类、数学类等学科中，这些学科在人工智能领域跨学科的发展演变过程中起着关键性的作用；随着时间的推移人工智能领域的跨学科特征越来越明显，依次经历了萌芽期、迅速发展期、平稳期和缓慢降低期。研究结果同时表明了人工智能领域的跨学科范围起始于人工智能领域邻近的核心学科，近年来逐渐扩散到更加遥远的潜在学科和边缘学科中，随着这些学科的出现，可能会形成一些新的交叉领域。

关键词: 人工智能, 跨学科, Rao-Stirling指数, 布里渊指数, 科学叠加图