1111 Transformative Research Foresight: Theoretical Model and Multi-dimensional Citation Characteristics Hot! Liang Guoqiang, Bu Yi, Hu Zhigang, Hou Haiyan DOI: 10.3772/j.issn.1000-0135.2022.11.001 Transformative research (TR), a major breakthrough in fundamental research, radically transforms our current scientific understanding and propels scientific research toward a new frontier. This study introduces a theoretical model and multi-dimensional citation characteristics of TR based on the cited-citing relationship between scientific studies. Additionally, this study explores the early characteristics of TR by using Nobel Prize-winning papers as a proxy. The results show that TR’s true nature is a potential new paradigm, as presented in Kuhn’s theory of scientific revolution; specifically, TR occurs in the beginning of a period of scientific revolution. As for the broadness of TR, the convergence of its disciplinary and early diffusion characteristics are apparent. Regarding the intensity dimension, the knowledge absorption and early disruption characteristics of TR are apparent. When considering the speed dimension, the obsolescent characteristics—as well as the early growth characteristics after its publication—are apparent. This research provides a theoretical and multi-dimensional citation framework for the quantitative analysis of TR at its early stage, which is a step toward TR forecasting. 2022 Vol. 41 (11): 1111-1123 [Abstract] ( 304 ) HTML (155 KB)  PDF (3815 KB)  ( 361 )

1124 Topic Prediction for Disruptive Technologies Based on Patent Literature—A Case Study of Artificial Intelligence Patents Hot! Chen Yuxin, Lu Jun, Han Yi DOI: 10.3772/j.issn.1000-0135.2022.11.002 As a driving force of economic development and an important starting point and breakthrough in technological innovation, disruptive technology is of great significance for a country or enterprise to optimize its R&D layout and actively seize the commanding heights of science and technology by detecting its dynamic development process and realizing the identification and prediction of disruptive technology. From the perspective of technology and market, a measurement index of disruptive potential was constructed, and a dynamic trend identification method to detect the disruptive potential of technology topics was constructed by combining a time series sliding window and latent Dirichlet allocation (LDA). Patents in the field of artificial intelligence in the United States were taken as a sample to verify the usability of this method in identifying and predicting disruptive technologies. Combined with the top 10 international patent classification (IPC) groups with high correlation intensity of identified disruptive technology topics, the contents of disruptive technology were characterized, and the practical value of the method was further tested. Deep learning and image recognition and processing are disruptive technologies in the field of artificial intelligence. They are closely related and have an obvious collaborative development trend. Deep learning technology focuses on the field of electronic digital data processing, while image recognition and processing is applied to mainstream fields such as automatic driving, medical diagnosis, and television communication. The sample empirical data shows that the multi-index fusion method has more advantages for the identification of disruptive technologies. Closely combining the multi-index fusion method of sample data with the development trend prediction and exploring the influence of historical development inertia through the iterative continuity of time series can better reveal the evolutionary details of disruptive technologies, their development trend, and the internal dependence of its various elements. 2022 Vol. 41 (11): 1124-1133 [Abstract] ( 567 ) HTML (163 KB)  PDF (1266 KB)  ( 274 )

1134 Research on Industry Emerging Technology Forecast Modeling Based on Knowledge Graph and Deep Neural Networks Hot! Wei Mingzhu, Zheng Rong, Gao Zhihao, Wang Xiaoyu DOI: 10.3772/j.issn.1000-0135.2022.11.003 Realizing accurate forecasts of emerging technologies in industry will help advance the layout of industry development, seize the technological commanding heights, and empower high-level technological self-reliance. Based on knowledge graph technology, this research defines the concept, relationship, and attributes of industrial technology patents and constructs an industrial technology patent knowledge graph, focusing on three main characteristics: novelty, social impact, and the fundamental innovation of emerging technologies. Starting from “Novelty-Application Scope-Development Ability” is used to build a complete and quantifiable indicator system for emerging technologies in the industry. Using the complex semantic information of the technology patent knowledge map, map query sentences are mapped to extract the feature values of various indicators, relying on deep neural network algorithms to train a model for predicting emerging technologies in an industry that realizes the accurate prediction of such technologies. Finally, the new energy automobile industry is taken as an example to verify the validity of the model. This research can provide a valuable reference for forecasting emerging technologies in various industries and providing decision support for industrial development. 2022 Vol. 41 (11): 1134-1148 [Abstract] ( 526 ) HTML (151 KB)  PDF (5455 KB)  ( 428 )

1149 Examining Multi-dimensional Technology Evolution Path and Technology Innovation Opportunity Identification Based on SAO Semantic Analysis Hot! Feng Lijie, Zhou Wei, Wang Jinfeng, Zhang Ke, Zhang Shibin DOI: 10.3772/j.issn.1000-0135.2022.11.004 Exploring the evolution path of key technologies in the target field and accurately identifying technological innovation opportunities represent essential steps for enterprises to stimulate technological innovation vitality. This study proposes a multi-dimensional technology evolution path and technology innovation opportunity identification method by combining SAO (subject-action-object) semantic analysis and the multi-dimensional technology innovation map. Firstly, the LDA (latent Dirichlet allocation) algorithm was used to determine the key technical issues in the target area. Secondly, the SAO semantic analysis method was utilized to accurately identify the semantic structure containing key technology elements, and the multi-dimensional technology innovation map was employed to navigate and classify key technology elements. Thereafter, by constructing the multi-dimensional technology evolution path and deeply exploring the law of technology evolution under different dimensions, the iterative transformation with innovation law was conducted to accurately judge the technology innovation opportunities. Finally, the modification technology of nano-TiO2 was taken as an example to conduct the analysis. Finally, the comparative analysis method was used to verify the effectiveness and practicability of the method. 2022 Vol. 41 (11): 1149-1160 [Abstract] ( 159 ) HTML (133 KB)  PDF (2627 KB)  ( 176 )

1161 Knowledge Convergence in Technological Innovation: Studying the Impact of Dual Knowledge of “Science-Technology” on Patents Hot! Kong Jia, Deng Sanhong, Zhang Jiarui, Kang Lele, Wu Jie DOI: 10.3772/j.issn.1000-0135.2022.11.005 Exploring the pattern of the absorption of dual knowledge of “science-technology” by patents in the process of technological innovation and analyzing its influence on patent innovation is of great significance for the management of technological innovation. As data sources, we selected 2,589,039 citing patents filed between 1979 and 2013, 4047758 papers, and 6868547 patents cited by citing patents. We conducted a field analysis of citing patents, cited papers, and cited patents, and measured the intensity and breadth of the dual knowledge of “science-technology” absorbed by citing patents. Further, we analyzed the influence of factors such as the intensity and breadth of scientific knowledge and technological knowledge on the influence of patents based on the zero-inflation negative binomial regression model. The measurement indicators proposed in this study, such as SI (science intensity), TI (technology intensity), STI (science-technology intensity), and knowledge breadth, can effectively measure the intensity and breadth of the absorption of scientific and technological knowledge by citing patents. Through our analysis, we found that the science intensity of patents in the fields of Biotechnology and Food Chemistry is greater than others, and the technology intensity of patents in the fields of IT Methods for Management and Medical Technology is greater than others. The scientific knowledge absorbed by citing patents is more distributed in the fields of Biological Sciences, Chemical Sciences, and Computer Science. The technological knowledge absorbed by citing patents is more concentrated in the fields of Computer Technology and Medical Technology. The regression model shows that there is an inverted U-shaped relationship between the influence of patent and breadth of scientific knowledge, and a positive U-shaped relationship with the breadth of technological knowledge. 2022 Vol. 41 (11): 1161-1173 [Abstract] ( 211 ) HTML (199 KB)  PDF (5754 KB)  ( 427 )

1174 Intelligent Decision-making Information System for Major Emergencies: A Holistic Approach to National Security Hot! Zhang Haitao, Luan Yu, Zhou Honglei, Zhang Xinrui, Pang Yufei, Liu Weili DOI: 10.3772/j.issn.1000-0135.2022.11.006 The management of major emergencies is a complicated task. National security risk is intertwined with emergency management, which coincides with the governance process for major emergencies. An intelligent decision-making information system for major emergencies is the key to preventing national security risks and conducting emergency management. This study is guided by the dual need for national security and emergency management from the perspective of overall national security. The feedback cycle of “information-intelligence-business” is a common theme running through the “physical-cyber-social” ternary world. This study is an in-depth investigation of the theoretical cognition and structural operation of the intelligent decision-making information system, based on three aspects: cognitive framework, system architecture, and organizational structure. The concepts of information demand, intelligence wisdom, and intelligence decision are explained. The information system architecture, driven by data and knowledge, is discussed. We put forward the idea of constructing “Large Information Business Groups” and suggest that an intelligent decision-making information system for major emergencies under overall national security, not only plays the roles of “detector, scout, and consultant,” but also guides management. 2022 Vol. 41 (11): 1174-1187 [Abstract] ( 260 ) HTML (113 KB)  PDF (4353 KB)  ( 379 )

1188 Key Technologies of Event Base Construction and Event Detection in the Field of Science and Technology Hot! Liu Yao, Fang Xiaowei, Qin Xun DOI: 10.3772/j.issn.1000-0135.2022.11.007 This study proposes a method to construct a risk event base and event detection model in the field of science and technology. By analyzing the text features of online news source data, the meta-event resource database and theme-event regenerated resource database models are constructed using crawled news in the field of science and technology, and a comprehensive evaluation model is proposed for event detection. For event detection, this study proposes a two-branch transformer model for risk events, which extracts lexical features related to risk degree from risk events and reduces the interference caused by text domain features to risk event classification, in order to identify risk events. The experimental results show that the proposed risk event detection model and the index for judging the risk propensity of meta-events are effective. This study can provide a reference for risk event base construction in the field of science and technology, and the proposed language model can provide a methodological and technical reference for the study of risk event detection. 2022 Vol. 41 (11): 1188-1198 [Abstract] ( 236 ) HTML (138 KB)  PDF (2230 KB)  ( 399 )

1199 Towards Automatic Literature Review Generation System: Research on Document Value Assessment Hot! Ding Heng, Ruan Jinglong DOI: 10.3772/j.issn.1000-0135.2022.11.008 The problem of information overload caused by “thesis explosion” has directed attention towards research on automatic review systems. How to automatically select important documents that can reflect the development of knowledge is the primary problem that the automatic review system needs to solve. In this study, starting from which factors influence review authors’ selection of references, the rules of review authors’ citation behaviors are excavated to assess the value of documents, and a document evaluation model for automatic review systems is constructed based on the ranking learning framework. This study uses Microsoft Academic Graph as the data source to construct an experimental data set and evaluates the experimental results through two indicators: ΔP@K and NDCG@K. The experimental results revealed two findings: (1) Compared with pointwise and listwise approaches, the pairwise approach is more suitable for training the optimal document evaluation model. The pairwise approach gains 0.274, 0.085, 0.738, and 0.831 on ΔP@100, ΔP@200, NDCG@100, and NDCG@200, respectively. (2) Knowledge importance, literature quality, and influence have a greater contribution to the improvement of the model and are the primary considerations for the authors of the review article to evaluate the value of the literature and choose references. 2022 Vol. 41 (11): 1199-1214 [Abstract] ( 187 ) HTML (299 KB)  PDF (4246 KB)  ( 392 )

1214 Recommendation of Patent Transaction Based on Attributed Heterogeneous Network Representation Learning Hot! He Xijun, Wu Shuangshuang, Wu Yuying, Cai Jiuran, Pang Ting, Chee Seng Chan DOI: 10.3772/j.issn.1000-0135.2022.11.009 Patent transaction recommendation is an important means of fusing heterogeneous information to facilitate transactions; however, the recommendation results are often affected by the disregard of patent attributes, which represents a research problem. This study proposes a patent transaction recommendation model based on attribute heterogeneous network (AHN) representation learning (AHNRL-PTR), which firstly filters the patent and organizational attributes affecting patent transaction, secondly constructs a patent transaction AHN, then introduces network representation learning in AHN, and finally, uses multidimensional Gaussian distribution and Kullback-Leibler divergence to solve the problems of node representation uncertainty and distance asymmetry between nodes. Finally, an empirical study with the valid invention granted patent data in the Greater Bay Area concluded that: first, compared to the metapath2vec, text-associated DeepWalk (TADW), and variant methods of the AHNRL-PTR model, the AHNRL-PTR model has the highest recommendation accuracy (more than 86%), indicating that fusing organizational and patent attributes and focusing on the solution of the uncertainty and asymmetry problem of node representation can substantially improve recommendation accuracy; second, the values of the non-accurate metrics IntraSim and Popularity of AHNRL-PTR are smaller than those of metapath2vec, AHNvec-PTR, and AHNsy-PTR methods, reflecting the diversity of this method’s recommendation results and its advantage in recommending niche cold patents; third, the organizations are clustered into the following six categories based on IntraSim and Popularity: intermediary, domain backbone, research, community, growth, and professional, which reflect the recommendation results’ interpretability and personalization level. Given the results, this study provides decision support for intelligent recommendation services for patent transactions. 2022 Vol. 41 (11): 1214-1228 [Abstract] ( 202 ) HTML (277 KB)  PDF (3588 KB)  ( 414 )