研究提出运用客观数据和系统的方法来识别颠覆性技术并提高识别精准度的方法。基于颠覆性技术理论，重新修正颠覆性技术特征，从创新性、扩散性和转轨性三方面构建一种采用搜索路径统计数（SPC）算法、专利吸收率和专利扩散率测度的新方法对颠覆性技术进行识别和判断；并利用1970－2020年间的1 985件专利的数据对智能语音领域进行实证分析，分析识别出强化学习和神经网络技术是该领域的颠覆性技术，而端到端的神经网络算法是该领域未来发展的方向。

The research proposes to use objective data and systematic methods to identify disruptive technologies and improve the accuracy of identification. Based on the theory of disruptive technology, this paper revises the characteristics of disruptive technology, and constructs a new method to identify and judge disruptive technology by using search path count (SPC) algorithm, patent absorption rate and patent diffusion rate measurement from the three aspects of innovation, diffusion and transition; This paper makes an empirical analysis on the field of intelligent speech, Use the data of 1985 patents from 1970 to 2020, and analyze and recognize that reinforcement learning and neural network technology are subversive technologies in this field, and end-to-end neural network algorithm is the future development direction of this field.

G255；F224；G301

国家社会科学“组织模块化驱动的企业颠覆性创新生态系统建构与管理机制研究”(19BGL045)