摘要
以大模型为核心的新一代人工智能生成内容(AIGC)技术(以下简称生成式人工智能技术)正代表着AI技术新的发展方向,而当前对于大模型的发展共识(即堆积算力和高质量数据可以继续创造“奇迹”)同时也促使人工智能科技创新转入“万卡”时代。由AI芯片所构筑的算力“护城河”成为支撑AI迈向通用人工智能(AGI)的必备需求。为研判我国AI芯片技术创新的发展态势,本文对我国专利数据进行了检索和分析,梳理和描绘了AI芯片技术的创新现状与趋势,并探索构建了代表性创新主体的创新潜力专利评估因子,以期形成对未来发展势能的理解与判断。最后,基于专利数据解读和分析,形成了助推我国AI芯片创新发展的有关建议。
The new generation of AI generated content(AIGC)technology(hereinafter referred to as generative AI technology),which is centered on big models,is representing a new direction of development of AI technology,The current consensus on the development of large models(i.e.,the accumulation of arithmetic power and high-quality data can continue to create"miracles")has also prompted AI technological innovation to shift into the era of"Wanka".The arithmetic power"moat"constructed by AI chips has become a necessary requirement to support AI to move towards general artificial intelligence(AGI).In order to study the development trend of AI chip technology innovation in China,this paper searches and analyzes China's patent data,combs and depicts the innovation status and trend of AI chip technology,and explores the construction of patent assessment factors of innovation potential of the representative innovation subject,with a view to forming the understanding and judgment of the future develop-ment potential.Finally,based on the patent data interpretation and analysis,relevant suggestions for boosting the innovative development of AI chips in China are formed.
作者
高雪松
李慧颖
GAO Xue-song;LI Hui-ying(China industrial control systems cyber emergency response team)
出处
《中国集成电路》
2024年第6期17-20,36,共5页
China lntegrated Circuit
关键词
生成式人工智能
AI芯片
专利分析
创新因子
创新态势
generative artificial intelligence
AI chips
Patent analysis
innovation factors
innovation dynamics