世界科技研究与发展 ›› 2024, Vol. 46 ›› Issue (4): 497-510.doi: 10.16507/j.issn.1006-6055.2023.09.001

• 人工智能 • 上一篇    下一篇

医学人工智能领域专利技术主题发展态势研究

周隽如 刘智勇   

  1. 华中科技大学同济医学院医药卫生管理学院,武汉 430030
  • 出版日期:2024-08-25 发布日期:2024-09-03
  • 基金资助:
    中央高校基本科研业务费项目(5003516062)

Research on the Development Trends of Patent Technology Themes in the Field of Medical Artificial Intelligence

ZHOU Junru   LIU Zhiyong   

  1. School of Medicine and Health Management, Tongji Medical College of Huazhong University of Science &Technology, Wuhan 430030, China
  • Online:2024-08-25 Published:2024-09-03

摘要:

对医学人工智能领域1971—2022年的专利文本进行文本挖掘,揭示其技术主题内容和演化趋势,分析技术研发热点和发展态势,有助于为科研人员梳理技术发展脉络,进而为未来研究和应用提供参考和借鉴。本文首先获取德温特专利数据库中14184条医学人工智能领域相关专利数据,并进行数据清洗,结合生命周期理论将其划分为三个阶段(萌芽期、发展期、快速发展期)。随后,使用BERTopic主题模型对专利文本进行主题识别,并通过计算技术主题对之间的语义相似度和主题过滤的方法对技术主题进行演化分析。结果显示,共识别出萌芽期主题5个、发展期主题9个、快速发展期主题29个,医学人工智能技术正处于快速发展阶段。其中,基础技术(信号处理与分析技术、图像处理与计算机视觉技术、数据挖掘技术)产生于萌芽期和发展期,并在发展期和快速发展期逐渐成熟,形成了分支领域技术;需求与技术的进步推动了基础技术的分化,技术之间的融合将会产生新的技术。最后,提出未来应加强对医学人工智能领域基础关键技术的研究和创新,并持续关注医学人工智能领域重要实际需求背后的技术问题,以助力技术创新。此外,还应强调医学人工智能领域的技术融合发展的重要性,鼓励多领域科研人员组建交叉学科人才团队,为促进多领域技术的深度融合和前沿多维度探索奠定基础。

关键词: 医学人工智能, 文本挖掘, 专利分析, 发展态势, BERTopic模型

Abstract:

The text mining of patent texts in the field of medical artificial intelligence (AI) from 1971 to 2022 reveals the content and evolution trend of its technical topics, and analyzes the hot spots and development trends of technology research and development, which is helpful to provide reference and reference for scientific researchers in the context of technological development, and then for future research and application. In this paper, 14184 patent data related to the field of medical artificial intelligence in the Derwent patent database are obtained, and the data is cleaned, and it is divided into three stages (embryonic stage, development stage, and rapid development stage) based on the life cycle theory. Subsequently, the BERTopic topic model was used to identify the subject of patent texts, and the evolution of technical topics was analyzed by calculating the semantic similarity between technical topic pairs and topic filtering. The results showed that a total of 5 embryonic themes, 9 development themes, and 29 rapid development topics were identified, indicating that medical artificial intelligence technology is in the rapid development stage. Foundational technologies such as signal processing and analysis, image processing and computer vision, and data mining emerged during the early and development stages, gradually maturing during the development and rapid development stages, leading to the formation of specialized technological domains. Both demand and technological advancements drove the differentiation of foundational technologies. The convergence of technologies is expected to generate novel advancements. Medical artificial intelligence technology is currently in a rapid stage of development. Finally, it is proposed that the research and innovation of basic and key technologies in the field of medical artificial intelligence should be strengthened in the future, and the technical issues behind the important practical needs in the field of medical artificial intelligence should be continuously paid attention to help technological innovation. Additionally, the significance of technological convergence and development in the field of medical AI should be emphasized, encouraging multidisciplinary teams comprising researchers from various fields, to lay the foundation for promoting deep integration and multidimensional exploration of interdisciplinary technologies at the forefront.

Key words: Medical Artificial Intelligence, Text Mining, Patent Analysis, Development Trend, BERTopic Model