图书情报工作

图书情报工作 >

2024 , Vol. 68 >Issue 5: 74 - 86

DOI: https://doi.org/10.13266/j.issn.0252-3116.2024.05.008

情报研究

基于离群点视角的颠覆性专利预测研究

  • 王丹 ,
  • 周潇 ,
  • 赵捧未 ,
  • 樊嘉逸
展开
  • 西安电子科技大学经济与管理学院 西安 710126
王丹,博士研究生;赵捧未,教授,博士,博士生导师;樊嘉逸,硕士研究生。

收稿日期: 2023-08-18

  修回日期: 2023-10-20

  网络出版日期: 2024-03-27

基金资助

本文系国家自然科学基金面上项目“重组创新视角下新兴共性技术识别及突破路径预测研究”（项目编号：72374165）研究成果之一。

收起

Research on Disruptive Patent Prediction Based on Outlier Perspective

  • Wang Dan ,
  • Zhou Xiao ,
  • Zhao Pengwei ,
  • Fan Jiayi
Expand
  • School of Economics and Management, Xidian University, Xi'an 710126

Received date: 2023-08-18

  Revised date: 2023-10-20

  Online published: 2024-03-27

Supported by

This work is supported by National Natural Science Foundation of China project titled “Research on emerging general-purpose technology identification and breakthrough Pathway forecast from the perspective of technology re-combination” (Grant No. 72374165).

Fold

摘要

[目的/意义]在日趋激烈的国际竞争背景下，颠覆性技术被认为是引领技术和产业发展方向、助推企业和产业实现“弯道超车”的绝佳机会窗口。为此，预测和部署颠覆性技术对于国家抢占科技制高点、重塑价值链均具有重大战略意义。[方法/过程]结合深度学习和离群点检测算法，构建基于离群点视角的颠覆性专利预测框架。该研究框架包括五个关键步骤：首先，利用BERT模型和TF-IDF算法将专利文本和专利分类号转化为可计算的高维向量表示，并结合PCA算法进行降维和特征融合；其次，采用三种离群点检测算法，以增量迭代的方式识别离群专利；再者，通过数据集修正，从离群专利中保留新技术专利；在此基础上，通过深度剖析新技术形式颠覆性专利的核心特征，构建有效的测度指标体系；最后，利用深度学习DNN模型拟合专利指标和颠覆性专利标签之间的关联关系，从而实现从大量的新技术专利中对潜在颠覆性专利的有效预测。[结果/结论]以人工智能为例，验证了该方法的有效性。结果共预测出411条颠覆性专利，这些专利主要涉及六大颠覆性方向：多模态预训练大模型、增强现实、生成式AI、自动驾驶、图像识别与处理和智能通信。这些技术的推广和应用，将对未来的科技和产业发展产生重大影响。研究结果可为国家政策制定和企业技术布局提供重要的决策参考。

关键词： 颠覆性技术; 离群点检测; 预测研究; 深度学习; 专利分析

本文引用格式

王丹 , 周潇 , 赵捧未 , 樊嘉逸 . 基于离群点视角的颠覆性专利预测研究[J]. 图书情报工作, 2024 , 68(5) : 74 -86 . DOI: 10.13266/j.issn.0252-3116.2024.05.008

Abstract

[Purpose/Significance] In the fierce international competition, disruptive technology is an important opportunity to lead the development of technology and industry, and help enterprises and industries achieve “overtaking on curves”. Therefore, to predict and deploy it is of great strategic significance for the country to seize the height of technology and reshape the value chain. [Method/Process] This paper used deep learning and outlier detection algorithms to construct a disruptive patent prediction framework. This research framework included five key steps. Firstly, it adopted the BERT model and TF-IDF algorithm to convert patent text and classification number into computable high-dimensional vector, and took the dimensionality reduction and feature fusion with the PCA algorithm. Secondly, it used three outlier detection algorithms to identify outlier patents through incremental iteration. Furthermore, by modifying the dataset, it retained new technology patents from outlier patents. On this basis, by deeply analyzing its core characteristics of disruptive patents in the form of new technologies, it constructed an effective measurement index system. Finally, it took the deep learning DNN model to fit the correlation between patent indicators and disruptive patent labels to effectively predict potential disruptive patents from a lot of new technology patents. [Result/Conclusion] Taking artificial intelligence as an example, it verifies the effectiveness of this method. It predicts 411 disruptive patents in six disruptive directions: multimodal pre-trained large models, augmented reality, generative AI, autonomous driving, image recognition and processing, and intelligent communication, which will have a significant impact on future technological and industrial development. The research results provide important decision-making references for national policy formulation and enterprise technology layout.

Key words： disruptive technology; outlier detection; predictive design; deep learning; patent analysis

参考文献

[1] 胡登峰, 黄紫微, 冯楠, 等.关键核心技术突破与国产替代路径及机制——科大讯飞智能语音技术纵向案例研究[J].管理世界, 2022, 38(5):188-209.(HU D F, HUANG Z W, FENG N, et al.Path and mechanism of core technology breakthrough and domestic substitution:a longitudinal case study of IFLY TEK intelligent speech technology[J].Journal of management world, 2022, 38(5):188-209.)
[2] 邓建军, 刘安蓉, 曹晓阳, 等.颠覆性技术早期识别方法框架研究——基于科学端的视角[J].中国科学院院刊, 2022, 37(5):674-684.(DENG J J, LIU A R, CAO X Y, et al.Methodological framework of identifying disruptive technologies on emerging stage:based on science[J].Bulletin of Chinese Academy of Sciences, 2022, 37(5):674-684.)
[3] 陈育新, 卢俊, 韩毅.基于专利文献的颠覆性技术识别研究——以人工智能为例[J].情报学报, 2022, 41(11):1124-1133.(CHEN Y X, LU J, HAN Y.Topic prediction for disruptive technologies based on patent literature:a case study of artificial intelligence patents[J].Journal of the China Society for Scientific and Technical Information, 2022, 41(11):1124-1133.)
[4] LEE C, JEON D, AHN J M, et al.Navigating a product landscape for technology opportunity analysis:a word2vec approach using an integrated patent-product database[J].Technovation, 2020, 96:102140.
[5] CAPPONI G, MARTINELLI A, NUVOLARI A.Breakthrough innovations and where to find them[J].Research policy, 2022, 51(1):104376.
[6] 许佳琪, 汪雪锋, 雷鸣, 等.从突破性创新到颠覆性创新:内涵、特征与演化[J].科研管理, 2023, 44(2):1-13.(XU J Q, WANG X F, LEI M, et al.From breakthrough innovation to disruptive innovation:their connotations, characteristics and evolution[J].Science research management, 2023, 44(2):1-13.)
[7] ARTHUR W B.The nature of technology:what it is and how it evolves[M].New York:Simon and Schuster, 2009.
[8] ARTS S, APPIO F P, VAN L B.Inventions shaping technological trajectories:do existing patent indicators provide a comprehensive picture?[J].Scientometrics, 2013, 97(2):397-419.
[9] CHRISTENSEN C M.The innovator's dilemma:when new technologies cause great firms to fail[M].Boston:Harvard Business School Press, 1997.
[10] 曲冠楠, 陈凯华, 陈劲.颠覆性技术创新:理论源起、整合框架与发展前瞻[J].科研管理, 2023, 44(9):1-9.(QU G N, CHEN K H, CHEN J.Disruptive technovation:origins, integrated framework, and prospects[J].Science research management, 2023, 44(9):1-9.)
[11] 苏成, 赵志耘, 赵筱媛, 等.颠覆性技术新阐释:概念、内涵及特征[J].情报学报, 2021, 40(12):1253-1262.(SU C, ZHAO Z Y, ZHAO X Y, et al.New Explanation on disruptive technology:concept, connotation and features[J].Journal of the China Society for Scientific and Technical Information, 2021, 40(12):1253-1262.)
[12] 王康, 陈悦.基于异质性专利的颠覆性技术早期识别研究[J].科学学研究, 2023, 41(8):1364-1375.(WANG K, CHEN Y.Research on early identification of disruptive technologies based on heterogeneous patents[J].Studies in science of science, 2023, 41(8):1364-1375.)
[13] SUN B, KOLESNIKOV S, GOLDSTEIN A, et al.A dynamic approach for identifying technological breakthroughs with an application in solar photovoltaics[J].Technological forecasting and social change, 2021, 165:120534.
[14] ZHANG N, YOU D, TANG L, et al.Knowledge path dependence, external connection, and radical inventions:evidence from Chinese Academy of Sciences[J].Research policy, 2023, 52(4):104738.
[15] DAHLIN K B, BEHRENS D M.When is an invention really radical? defining and measuring technological radicalness[J].Research policy, 2005, 34(5):717-737.
[16] 苏敬勤, 刘建华, 王智琦, 等.颠覆性技术的演化轨迹及早期识别——以智能手机等技术为例[J].科研管理, 2016, 37(3):13-20.(SU J Q, LIU J H, WANG Z Q, et al.The evolution trajectory and early identification of disruptive technology by taking smartphones and other technologies as an example[J].Science research management, 2016, 37(3):13-20.)
[17] 梁镇涛, 毛进, 李纲.融合"科学-技术" 知识关联的高颠覆性专利预测方法[J].情报学报, 2023, 42(6):649-662.(LIANG Z T, MAO J, LI G.Integrating science-technology knowledge linkage to predict disruptive patents[J].Journal of the China Society for Scientific and Technical Information, 2023, 42(6):649-662.)
[18] ZHOU Y, DONG F, LIU Y, et al.A deep learning framework to early identify emerging technologies in large-scale outlier patents:an empirical study of CNC machine tool[J].Scientometrics, 2021, 126:969-994.
[19] 孔德婧, 董放, 陈子婧, 等.离群专利视角下的新兴技术预测——基于BERT模型和深度神经网络[J].图书情报工作, 2021, 65(17):131-141.(KONG D J, DONG F, CHEN Z J, et al.Prediction of emerging technologies from the perspective of outlier patents-based on BERT model and deep neural networks[J].Library and information service, 2021, 65(17):131-141.)
[20] JEON D, AHN J M, KIM J, et al.A doc2vec and local outlier factor approach to measuring the novelty of patents[J].Technological forecasting and social change, 2022, 174:121294.
[21] 罗素平, 寇翠翠, 金金, 等.基于离群专利的颠覆性技术预测——以中药专利为例[J].情报理论与实践, 2019, 42(7):165-170.(LUO S P, KOU C C, JIN J, et al.Disruptive technology prediction based on outlier patents:traditional Chinese medicine patents as an example[J].Information studies:theory & application, 2019, 42(7):165-170.)
[22] ARTS S, HOU J, GOMEZ J C.Natural language processing to identify the creation and impact of new technologies in patent text:code, data, and new measures[J].Research policy, 2021, 50(2):104144.
[23] 马荣康, 王艺棠.基于专利相似度的突破性技术发明识别研究——以纳米技术为例[J].科研管理, 2021, 42(5):153-160.(MA R K, WANG Y T.Identification of breakthrough inventions based on patent-to-patent similarity:a case study of nanotechnology[J].Science research management, 2021, 42(5):153-160.)
[24] 刘小玲, 谭宗颖.基于专利多属性融合的技术主题划分方法研究[J].数据分析与知识发现, 2022, 6(2/3):45-54.(LUI X L, TAN Z Y.Clustering technology topics based on patent multiattribute fusion[J].Data analysis and knowledge discovery, 2022, 6(2/3):45-54.)
[25] 赵雪峰, 胡瑾瑾, 吴德林, 等.一种基于特征拼接、标签迁移及深度学习组合的专利价值评估方法[J].情报学报, 2023, 42(6):663-680.(ZHAO X F, HU J J, WU D L, et al.Patent valuation method based on a combination of feature stitching, label migration, and deep learning[J].Journal of the China Society for Scientific and Technical Information, 2023, 42(6):663-680.)
[26] LI X, WU Y, CHENG H, et al.Identifying technology opportunity using SAO semantic mining and outlier detection method:a case of triboelectric nanogenerator technology[J].Technological forecasting and social change, 2023, 189:122353.
[27] KWON S, LIU X, PORTER A L, et al.Research addressing emerging technological ideas has greater scientific impact[J].Research policy, 2019, 48(9):103834.
[28] LIANG Z, MAO J, LU K, et al.Combining deep neural network and bibliometric indicator for emerging research topic prediction[J].Information processing & management, 2021, 58(5):102611.
[29] 李乾瑞, 郭俊芳, 黄颖, 等.基于专利计量的颠覆性技术识别方法研究[J].科学学研究, 2021, 39(7):1166-1175.(LI Q R, GUO J F, HUANG Y, et al.Research on the method of disruptive technology identification based on patent bibliometrics[J].Studies in science of science, 2021, 39(7):1166-1175.)
[30] HIGHAM K, DE R G, JAFFE A B.Patent quality:towards a systematic framework for analysis and measurement[J].Research policy, 2021, 50(4):104215.
[31] 梁国强.变革性研究的引文特征分析与早期识别模型构建[D].大连:大连理工大学, 2021.(LIANG G Q.Citation patterns analysis and the early identification model construction of transformative research[D].Dalian:Dalian University of Technology, 2021.)
[32] BENSON C L, MAGEE C L.Quantitative determination of technological improvement from patent data[J].Plos one, 2015, 10(4):e0121635.
[33] COZZENS S, GATCHAIR S, KANG J, et al.Emerging technologies:quantitative identification and measurement[J].Technology analysis & strategic management, 2010, 22(3):361-376.
[34] VAN Z N, VAN P D L P B.The vulnerability of patent value determinants[J].Economics of innovation and new technology, 2011, 20(3):283-308.
[35] MA Z, LEE Y.Patent application and technological collaboration in inventive activities:1980-2005[J].Technovation, 2008, 28(6):379-390.
[36] MEYER M.Are patenting scientists the better scholars? an exploratory comparison of inventor-authors with their noninventing peers in nano-science and technology[J].Research policy, 2006, 35(10):1646-1662.
[37] FORMAN G.An extensive empirical study of feature selection metrics for text classification[J].Journal of machine learning research, 2003, 3:1289-1305.
[38] TSAY M Y, LIU Z W.Analysis of the patent cooperation network in global artificial intelligence technologies based on the assignees[J].World patent information, 2020, 63:102000.
[39] 达摩院.达摩院十大科技趋势[EB/OL].[2024-01-05].https://damo.alibaba.com/techtrends/2023?lang=zh.(Alibaba DAMO Academy.Top ten technology trends of Damo academy[EB/OL].[2024-01-05].https://damo.alibaba.com/techtrends/2023?lang=zh.)
[40] 艾瑞咨询研究院.2023年中国增强现实(AR)行业研究报告[EB/OL].[2024-01-05].https://www.iresearch.com.cn/Detail/report?id=4138&isfree=0.(iResearch.2023 China augmented reality (AR) industry research report[EB/OL].[2024-01-05].https://www.iresearch.com.cn/Detail/report?id=4138&isfree=0.)
[41] GARNTER.Generative AI, machine customers and AR/VR are expected to transform sales in the next five years[EB/OL].[2024-01-05].https://www.gartner.com/en/newsroom/pressreleases/2022-10-10-gartner-identifies-seven-technologydisruptions-that-will-impact-sales-through-2027.
[42] 飞浆.百度研究院2023科技趋势预测发布|AI向实而生, 智能技术构筑科技变革主线[EB/OL].[2024-01-05].https://www.paddlepaddle.org.cn/support/news?action=detail&id=3267.(PaddlePaddle.Baidu Research Institute 2023 technology trend forecast release|AI emerging from reality, intelligent technology building the mainline of technological change[EB/OL].[2024-01-05].https://www.paddlepaddle.org.cn/support/news?action=detail&id=3267.)
[43] 欧阳晔, 王立磊, 杨爱东, 等.通信人工智能的下一个十年[J].电信科学, 2021, 37(3):1-36.(OU Y Y, WANG L L, YANG A D, et al.Next decade of telecommunications artificial intelligence[J].Telecommunications science, 2021, 37(3):1-36.)
Options
文章导航

/

〈 〉