Analysis of Research Hotspots and Frontiers in Deep Learning in the Field of Artificial Intelligence Based on CiteSpace

人工智能的发展已对人类及其未来产生了深远的影响，深度学习源于神经网络的研究，是机器学习的一个全新领域，日益成为驱动社会智能化转型的核心引擎。本文借助CiteSpace 6.4.R1软件，对中国知网核心数据库中人工智能领域深度学习研究文献进行可视化分析，按年发文量、机构合作、关键词共现、关键词聚类、时区图谱及突现词分析对2010~2024年的文献进行梳理。研究结果表明，年发文量在2016年后快速增长。研究机构间整体合作关系较为松散，从已有的合作网络来看，存在“一超多强”的现状，以中国科学院大学为核心节点的合作网络最为紧密，以武汉大学、清华大学、北京工业大学、复旦大学等为核心节点也形成了小范围的合作网络。该领域研究的热点为人工智能、深度学习、机器学习、大数据和神经网络等，未来研究着眼于各领域的预测、可解释性、隐私保护、边缘计算等，深度学习研究正从单一的技术突破转向“算法–场景–伦理”的系统性创新，未来将深度重构医疗、制造、交通等核心产业的价值链。The development of artificial intelligence (AI) has had a profound impact on humanity and its future. Originating from research on neural networks, deep learning, as a brand-new field of machine learning, has increasingly become the core engine driving the intelligent transformation of society. This paper uses CiteSpace 6.4.R1 software to conduct a visual analysis of literature on deep learning in the field of AI from the core database of the China National Knowledge Infrastructure (CNKI), sorting out the literature from 2010 to 2024 through annual publication volume, institutional collaboration, keyword co-occurrence, keyword clustering, time-zone mapping, and burst term analysis. The research results show that the annual publication volume has grown rapidly since 2016. The overall cooperative relationships among research institutions are relatively loose. From the existing cooperation network, there is a current situation of “one superpower and multiple strong players”: the cooperation network with the University of Chinese Academy of Sciences as the core node is the closest, while small-scale cooperation networks have also formed with Wuhan University, Tsinghua University, Beijing University of Technology, Fudan University, etc., as core nodes. The research hotspots in this field include artificial intelligence, deep learning, machine learning, big data, and neural networks. Future research will focus on predictions in various fields, such as interpretability, privacy protection, edge computing, etc. Deep learning research is shifting from single-technique breakthroughs to systematic innovation of “algorithm-scene-ethics”, which will deeply reconstruct the value chains of core industries such as healthcare, manufacturing, and transportation in the future.