A bibliometric network and content analysis of deep learning research within artificial intelligence from 2019 to 2025

Abstract The rapid development of artificial intelligence (AI) and deep learning has driven innovation across multiple fields. This study aims to provide a comprehensive overview of deep learning research within AI from 2019 to 2025. A total of 1214 relevant articles from the Web of Science were systematically analyzed using bibliometric, network, and content analysis methods. VOSviewer and R Studio (Biblioshiny) supported bibliometric and network analyses to examine publication trends, influential authors, journals, countries, institutions, keywords, and key publications, while NVivo 14 was used for content analysis to categorize core AI concepts, technical methods, and domain-specific applications. The results indicate substantial growth in research output, particularly in computer vision, natural language processing, and medical diagnostics. Collaborative studies dominate the field, with key contributors including Kaiming He, Yann LeCun, and Ramprasaath R. Selvaraju. Leading journals such as IEEE Access and Scientific Reports , and top institutions including the National University of Singapore and Mayo Clinic play a central role in knowledge dissemination. The United States and China emerge as the most influential countries. Keyword and content analyses reveal six research clusters, emphasizing methodological innovation, practical application, and cross-domain problem-solving. Overall, this study provides a detailed overview of the development, influential contributors, and thematic trends in deep learning research within AI, offering insights to guide future investigations and highlight emerging opportunities in both theory and application.