Intelligence Theories and Methods

509 Construction of Science and Technology Security Information Service System Empowered by Digital Intelligence Hot! Wang Kaile, Chen Yunwei DOI: 10.3772/j.issn.1000-0135.2025.05.001 Amid profound and complex transformations in the global landscape, the dual imperatives of advancing scientific innovation and safeguarding technological security have become increasingly urgent. As a pivotal variable reshaping international dynamics, science and technology (S&T) security has increased the strategic importance of S&T security information. To address the pressing need for intelligent transformation in S&T security information services, this study considered existing literature and systematically examined the conceptual foundations, defining characteristics, evolutionary trends, and risks inherent to S&T security. Building upon this analysis, this paper proposes and constructs an S&T security information service system empowered by digital intelligence from the aspects of service subject and object, service purpose and content, service methods and means, service processes and strategies, and service mechanisms. The proposed system not only enriches the theoretical framework of S&T security information but also offers actionable insights for delivering high-impact information services in this critical domain. 2025 Vol. 44 (5): 509-521 [Abstract] ( 11 ) HTML (1 KB)  PDF (1970 KB)  ( 95 )

522 Emerging-Technology Identification Based on Technology Structure-Function Semantic Association Hot! Jiang Man, Yang Siluo, Wei Jiaze DOI: 10.3772/j.issn.1000-0135.2025.05.002 Unearthing the deep semantic information of emerging technologies is crucial for accurately capturing technology development opportunities and enhancing the strategic orientation of emerging technologies. This study focuses on the technological element level, which encompasses the technology structure and function. Initially, a model for extracting technological structure-function elements is constructed. Next, an improved “subject-action-object” method is used to extract the internal semantic structure of the technological elements. Subsequently, a comprehensive multidimensional index is established based on the four attributes of the emerging technology to evaluate its emergence degree. Finally, a multidimensional association diagram is constructed to illustrate the internal and external semantic structures of the emerging technological elements and their semantic associations and trends. Empirical evidence from digital medical technology shows that this method effectively identifies emerging technological structures and functions, as well as reveals their internal and external semantic associations. The results indicate that while new technological structures do not necessarily result in new functions, most new functions depend on innovative structures. This method proficiently reveals the internal and external semantic associations and can present the content of emerging technologies with fine granularity and multiple dimensions. 2025 Vol. 44 (5): 522-534 [Abstract] ( 14 ) HTML (1 KB)  PDF (3692 KB)  ( 48 )

535 Impact of Data Factor Marketization on Disruptive Technologies Hot! Wang Haisen, Li Gang DOI: 10.3772/j.issn.1000-0135.2025.05.003 Under the backdrop of an era characterized by accelerating global transformation, the technological competition among major powers has become increasingly intense. Disruptive technologies have ascended to become a core component of the national security strategy of China and serve as a critical driver for building an innovative nation and modernizing national governance capabilities and systems in this new era. This study focuses on the institutional practices of market-oriented reforms for data factors in China, and it utilizes the construction of data trading platforms as a quasi-natural experimental setting. Based on data from Shanghai and Shenzhen A-share listed companies (spanning from 2014 to 2023), we constructed multi-period difference-in-differences models and causal mediation models to systematically analyze the impact of data factor marketization on corporate disruptive technologies and its underlying mechanisms. The findings reveal that data factor marketization significantly promotes corporate disruptive technologies. In this process, marketization breaks down data silos and monopoly barriers, expands the “data resource pool” for enterprises to acquire new knowledge, enhances the technological capabilities of corporate data, optimizes the efficiency of data factor allocation among enterprises, and creates greater resource opportunities for disruptive technologies. Further analysis identifies that “hidden barriers” formed by data protectionism remain a significant challenge. Although current market mechanisms can reduce the cost of acquiring existing knowledge, they provide limited incentives for high-end disruptive technologies that require breakthroughs in cognitive boundaries. The conclusions of this study offer new mechanisms for incentivizing disruptive technologies and provide novel insights for improving data factor markets. 2025 Vol. 44 (5): 535-548 [Abstract] ( 10 ) HTML (1 KB)  PDF (980 KB)  ( 67 )

549 Research on Patent Tradability Prediction Based on Weight Balancing Algorithm Hot! Ran Congjing, Ding Qunzhe, Li Wang, Song Yonghui, Liu Shuang DOI: 10.3772/j.issn.1000-0135.2025.05.004 As a crucial link between innovation and the realization of market value, patent transactions play a significant role in supporting national strategic goals, driving technological progress, and facilitating technology transfer and collaboration between enterprises and research institutions. Therefore, considering the importance of identifying patents with high market potential to foster innovation and collaboration, predicting patent tradability is crucial. To address this, this paper proposes a patent tradability prediction method based on a weighted balancing algorithm. To support the prediction of patent tradability, an initial dataset was constructed by integrating data from the incoPat patent database and China Patent Information Service Platform. The initial patent transaction dataset was further refined based on patent transfer records, transfer and assignee addresses, and stakeholder information, using a series of rules and algorithms to construct the final patent transaction dataset. Using this dataset, patent tradability prediction is framed as a supervised binary classification task. The input variables include the multidimensional technical features of the patents prior to the transaction, while the target variable indicates whether the patent is transacted before its expiration. Using this methodology, the proposed patent tradability prediction model based on the weighted balancing algorithm outperformed baseline models in terms of overall performance, and its effectiveness was validated through empirical results. In addition, model interpretability techniques revealed that key technical features, such as applicant country, applicant type, number of family patents, and number of family countries, significantly influence patent transactions. Despite these advancements, predicting patent transactions remains challenging. Future research could explore the incorporation of multidimensional features, such as patent text and images, to further enhance the predictive performance of the model. 2025 Vol. 44 (5): 549-561 [Abstract] ( 8 ) HTML (1 KB)  PDF (2929 KB)  ( 101 )

Intelligence Technology and Application

562 High-Quality Answer Identification in Online Health Communities Based on a Time-Series Graph Convolutional Neural Network Hot! Sun Wenjing, Ma Jie, Hao Zhiyuan DOI: 10.3772/j.issn.1000-0135.2025.05.005 The consultation modules in online health platforms are important channels for users to seek health consultations and obtain health information under the “internet+medical treatment” framework. The accurate identification of high-quality answers is important for guiding users in making informed health decisions. This study takes consultation texts from the “Good Doctor Online” platform as the research object and proposes a high-quality answer identification model based on a time-series graph convolutional neural network. Based on dual-process and graph theories, this model designs a feature system to measure the quality of answers, thereby creating an experimental dataset. Moreover, this model uses the graph sample and aggregate (GraphSAGE) graph convolutional neural network (CNN) to integrate the refined indicators of a quality measurement feature system. This model also integrates the gated recurrent unit (GRU) in GraphSAGE and constructs a “doctor-question” network graph as the model input and ultimately forms the GraphSAGE-GRU model to predict the quality of answers and identify high-quality answers. In this study, support vector machine (SVM), decision tree (DT), k-nearest neighbor (kNN), CNN, GRU, and graph convolutional network (GCN) are selected as baseline methods for control experiments. The results show that the proposed model has a higher accuracy of 93.2% and exhibits the best performance in identifying high-quality answers. Furthermore, nearly 95% of high-quality answer samples can be identified from experimental samples. 2025 Vol. 44 (5): 562-576 [Abstract] ( 8 ) HTML (1 KB)  PDF (3162 KB)  ( 93 )

577 Interdisciplinary Event Knowledge Fusion Research for Library Digital Collections Hot! Wang Zhongyi, Wang Zeren, Li Zhipeng, Zhang Jiexin DOI: 10.3772/j.issn.1000-0135.2025.05.006 In the current big science era, the research problems encountered are increasingly complex. Thus, interdisciplinary research has gradually developed as a tool for solving these complex problems, which has led to a growing demand for interdisciplinary knowledge services. Therefore, librarians must gather the knowledge required to solve complex problems from different disciplines and provide comprehensive, interdisciplinary knowledge services. The fusion of interdisciplinary event knowledge has become crucial in addressing this challenge. Starting with the fusion of interdisciplinary event knowledge, this study first proposes a model for extracting interdisciplinary event knowledge based on argumentative semantic associations to extract relevant event knowledge from digital library collections across various disciplines. Subsequently, methods are introduced to fuse and generate interdisciplinary linear and nonlinear event knowledge. Finally, this study conducts empirical research in the field of climate change to demonstrate the feasibility and effectiveness of this approach. 2025 Vol. 44 (5): 577-591 [Abstract] ( 11 ) HTML (1 KB)  PDF (7035 KB)  ( 15 )

592 The Effects of Task Dependencies on Human-Bot Collaborative Patterns in Online Knowledge Communities: An Empirical Study from a Machine Behavior Perspective Hot! Zuo Min, Qiu Jiangnan DOI: 10.3772/j.issn.1000-0135.2025.05.007 Bots have been introduced into online knowledge communities (OKC) to achieve human-bot intelligence enhancement, and the key is to understand the interdependence between humans and bots performing tasks, as well as the human-bot collaborative pattern that coordinates teamwork. From the perspective of machine behavior, this study uses process mining methods to identify and subdivide human-bot collaboration patterns into two categories: automated and enhanced assistance. Based on coordination theory, a dual fixed-effects model was used to empirically analyze the impact of three basic task dependencies-flow, integration, and sharing-on human-bot collaborative patterns, as well as the moderating effect of task types. The results indicate that the enhanced assistance collaborative pattern of bot-assisted humans effectively manages task-dependent human-bot team coordination problems, whereas the role of the automated assistance collaborative pattern that is independently executed by bots is limited. Notably, the above relationships change depending on task type. This study expands the application of coordination theory and machine behavior, enriches the empirical research on the factors influencing human-bot collaborative patterns, and provides useful guidance for managing human-bot processes and designing tasks on the OKC platform. 2025 Vol. 44 (5): 592-608 [Abstract] ( 9 ) HTML (1 KB)  PDF (2170 KB)  ( 45 )

Intelligence Users and Behavior

609 Perceived Values, Attitudes, and Behavioral Responses of Knowledge Creators towards Generative AI in the Context of Human-AI Competition Hot! Jia Mingxia, Zhao Yuxiang, Zhang Yan, Zhang Xiaoyu, Song Shijie DOI: 10.3772/j.issn.1000-0135.2025.05.008 Generative artificial intelligence (GAI) is reshaping traditional content and knowledge creation paradigms while intensifying competition between human knowledge creators and the evolving capabilities of GAI. Previous research has focused on the broad economic impacts of GAI on labor markets; however, it has largely ignored the personal experiences of creators adapting to this technology. This study addresses this gap by exploring the values, attitudes, and behaviors of creators toward GAI using a value-attitude-behavior framework and innovation diffusion theory. Through a grounded theory approach and data from interviews and online texts, the study reveals the following findings. (1) GAI leads to both substitution and complementarity effects, with creators progressing through phases from initial contact to competitive symbiosis. (2) Competitive pressure is influenced by professional, technical, and organizational factors but can be mitigated by intrinsic satisfaction and a sense of technological identity; however dependency on GAI may increase this pressure. (3) The attitude of creators toward GAI, which affects their behavioral responses, evolves over time. (4) Individual traits, organizational support, and social influence drive changes in the experiences and responses of creators. This study offers insights into the co-evolution of human knowledge creators and GAI as well as provides practical guidance for designing human-AI interactions and managing knowledge services. 2025 Vol. 44 (5): 609-628 [Abstract] ( 10 ) HTML (1 KB)  PDF (2652 KB)  ( 483 )

Intelligence Reviews and Comments

629 Application Characteristics and Development Trends of Data-Intelligence Integration Methods in Scientific and Technological Information Scenarios Hot! Liu Bowen, Xia Yikun, Ba Zhichao DOI: 10.3772/j.issn.1000-0135.2025.05.009 Both the data and technology environments in the data-intelligence era have undergone profound changes, and the “data-intelligence” transformation of scientific and technological information research is extensively investigated. Organizing existing related studies from a scenario-driven perspective can not only provide insights into the evolution of data-intelligence integration methods based on application demands but also provide a reference for scientific and technological information research to better accommodate uncertain, dynamic, and complex scenarios. Based on the analysis and definition of “data-intelligence integration,” this article analyzes the profound impact of the application of data-intelligence integration methods on scientific and technological information work, reviews relevant domestic and foreign research papers published in 2014-2023, and reveals data-intelligence integration in scientific and technological information scenarios. The application characteristics and development trends of the method are analyzed, and the scientific and technological information scene is structurally reshaped based on the data-information-knowledge-wisdom concept chain. By constructing an evolution diagram of the data-intelligence integration method, the main problems presented in current studies are summarized and discussed. Subsequently, four frontier scenarios of future development and their response strategies are proposed. This study demonstrates that research methods in scientific and technological information scenarios show a significant “data-intelligence integration” trend, whereas related theories and methods present challenges that warrant further investigation. 2025 Vol. 44 (5): 629-644 [Abstract] ( 6 ) HTML (1 KB)  PDF (3393 KB)  ( 150 )