基于人工智能专利图谱的技术热点发现以及演化分析

Technology Focus and Its Evolution Detection Based on Patent Map of Artificial Intelligence

人工智能在近几年快速发展并成为最热门的技术之一,如能快速了解人工智能技术热点与发展态势,对抢抓人工智能发展的重大战略机遇与构筑先发优势具有重要意义。本文提出了一种基于专利可视化图谱发现技术热点的方法,即在连续时间窗口上绘制图谱,在此基础上运用密度分布变化来识别该领域技术热点。为了提升专利图谱的准确性,本文使用海量专利文本训练了基于深度学习的doc2vec模型,形成了专利文本特征抽取模型。经过实验对比发现该模型在测试数据集中表现远超经典的词袋模型与主题模型。在实例分析中使用了2012—2019年10457件三方人工智能专利进行热点发现,共发现研究热点7个,并对7个热点中关键概念词、专利申请人所属国家进行深入分析。

Artificial intelligence has been developing rapidly in recent years and has become one of the hottest technologies.It is important to quickly understand the hot spots and development trends of artificial intelligence to seize the major strategic opportunities and build first-mover advantages in the development of artificial intelligence.This study proposes a method to detect technology focus and its evolution,10457 triadic AI patents were used as a case study dataset.Based on the doc2vec model and t-SNE visualization algorithm,the patent maps were created with three-time window.After that,the area with higher density distribution in the map is calculated and regarded as a technology hotspot.Finally,hot focuses and trends were determined according to focuses location changing on the three-time window maps.A total of seven research focuses were identified,and in-depth analysis of key concept words and countries of patent applicants in the seven focuses was conducted.