AI芯片专利技术研发态势

The R&D Trend of AI Chip Patent Technology

近年来,人工智能在基础研究、技术研发和产业应用方面均取得了快速发展,也对计算芯片提出了新的需求和挑战。深度学习算法需对海量数据进行训练,对大规模并行计算能力有很高的要求,而传统计算架构无法支撑大规模计算需求,因此新架构的AI芯片得以迅速发展。文章以专利文献为分析对象,从专利态势、专利来源国/地区、专利申请人、高价值专利四个角度对AI芯片及GPU、FPGA、CPLD等分支领域专利进行分析,揭示全球AI芯片专利技术布局态势和竞争格局,以期为我国人工智能及AI芯片技术研发布局提供情报参考和支撑。

In recent years,Artificial Intelligence has made rapid progress in R&D and industrial application,which has also put forward new demands and challenges for computing chips.Deep learning algorithms need to train massive data and have a high requirement on the ability of large-scale parallel computing.However,the traditional computing architecture cannot support the demand of large-scale computing,so the AI chip of the new architecture develops rapidly.Taking patent literature as the analysis object,this paper analyzes the patents of AI chip,GPU,FPGA and other sub-fields from four perspectives,namely,patent situation,country/region of origin,patent applicant and high-value patent,to reveal the global AI chip patent technology layout situation and competition pattern,and provide reference and support for China’s Artificial Intelligence and AI chip technology research and development layout.