基于文献关联的生成式人工智能技术演化分析  

Research on the Evolution of Generative Artificial Intelligence Technology Based on Literature Association

作　　者：赛秋玥 徐峰[1] 雷孝平[1] SAI Qiuyue;XU Feng;LEI Xiaoping(Institute of Scientific and Technical Information of China,Beijing 100038,China)

机构地区：[1]中国科学技术信息研究所,北京100038

出　　处：《情报工程》2024年第5期18-28,共11页Technology Intelligence Engineering

基　　金：科技创新2030—“新一代人工智能”重大项目、“新一代人工智能风险防范与治理手段研究”课题、“新一代人工智能伦理风险评估与应对策略研究”(2023ZD0121701)。

摘　　要：[目的/意义]生成式人工智能(AIGC)作为一种前沿技术带来了新的工业革命,全球竞相推进AIGC技术的发展。深入理解AIGC的内涵是把握其发展趋势的关键。[方法/过程]从多维视角进行AIGC技术分析,针对前沿技术的识别和路径演化设计了YAKE-Apriori算法展示领域内技术发展规律和路径。首先,基于YAKE算法对AIGC相关技术进行识别和可视化;其次,基于Apriori算法分析AIGC关键技术的关联技术,揭示技术发展规律;最后,基于关键技术—技术基础分析刻画技术演化路径,识别技术发展层级关系。[结论/结果]以AIGC为对象,成功识别出GAN、Transformer和Diffusion等AIGC相关技术和4个主要发展阶段,研究有助于厘清AIGC的演化路径,同时验证了YAKE-Apriori算法对前沿技术的识别和路径演化分析的有效性。[Objective/Significance]The advent of Generative Artificial Intelligence(AIGC)marks a new industrial revolution,with nations around the world engaging in the competitive race to advance AIGC technology.Understanding the essence and characteristics of AIGC technology is vital for responding to its growth.[Methods/Processes]This study adopts a multidimensional perspective to identify the evolutionary pathways of AIGC technology.The YAKE-Apriori algorithm is designed to identify cutting-edge technologies and path evolution,illuminating the directions of technological development.First,the YAKE method is applied to extract key technologies and visualize literature.Then,the Apriori algorithm is utilized to analyze the associations between key technologies,thereby revealing the underlying patterns of technological evolution.Finally,the study delineates the trajectory of technology evolution through an analysis of key technologies and their foundational bases,pinpointing the hierarchical relationships within technological development.[Results/Conclusions]Focusing on AIGC,this study successfully identifies relevant technologies such as GAN,Transformer,and Diffusion,along with four primary developmental stages.These insights are crucial for grasping AIGC’s trajectory and affirm the effectiveness of YAKE-Apriori method in tracking the progress of emerging technologies.

关 键 词：生成式人工智能 技术特征 发展脉络 路径演化 热点分析 

分 类 号：G35[文化科学—情报学] G203