首页|期刊导航|情报杂志|面向美国出口管制的关键核心技术识别与特征测度

面向美国出口管制的关键核心技术识别与特征测度OA

Research on the Identification and Characterization of Key Core Technologies Oriented to U.S.Export Controls:A Case Study of Artificial Intelligence

中文摘要英文摘要

[目的]实现我国关键核心技术突破的前提是关键核心技术的识别.从美国对华出口管制视角出发,研究我国受制企业的关键核心技术,为打赢关键核心技术攻坚战提供借鉴参考.[方法]以美国关键新兴技术清单、商业管制清单、实体清单为依据,运用Word2vec模型计算清单描述文本和我国受制企业专利文本的相似度,从而识别出我国人工智能关键核心技术,并利用XGBoost算法和SHAP解释方法分析我国人工智能关键核心技术的重要特征.[结果/结论]从228 家我国受制企业2018-2024 年申请的 50 445 件专利中识别出 6 882 件人工智能关键核心技术.技术影响力、保护力度、科学基础、市场前景、技术融合度是我国人工智能关键核心技术的高辨识度特征.

[Purpose]The prerequisite for achieving the breakthrough of key core technologies in China is the identification of key core technologies.From the perspective of U.S.export control on China,this paper researches on the key core technologies of our country's constrained enterprises,to provide reference for our country to win the battle of key core technologies.[Method]The paper uses Word2vec model to calculate the similarity between the description text of the list and the patent text of China's controlled enterprises,based on the list of key emerging technologies,the business control list and the entity list of the United States,so as to identify the key core technologies of China's AI.The paper uses XGBoost algorithm and SHAP interpretation method to analyze the important characteristics of the key core technologies of AI in China.[Result/Conclusion]The paper identifies6882 AI key core technologies from 50445 patents applied by 228 Chinese controlled enterprises between 2018 to 2024.Technology influence,protection,scientific basis,market prospect and technology integration are the high differentiation characteristics of the key core technologies of AI in China.

戚湧;程驭

南京理工大学知识产权学院 南京 210094南京理工大学知识产权学院 南京 210094

社会科学

出口管制关键核心技术人工智能Word2vecXGBoostSHAP

export controlskey core technologiesartificial intelligenceWord2vecXGBoostSHAP

《情报杂志》 2026 (1)

64-74,11

国家社会科学基金重点项目"新型举国体制下打赢关键核心技术攻坚战的目标、主攻方向与对策研究"(编号:23AZD038)研究成果.

10.3969/j.issn.1002-1965.2026.01.009

评论