Special Topics

1493 Information Science in the Era of Data Intelligence: Integration, Innovation, and Development Hot! Li Gang, Sun Jie DOI: 10.3772/j.issn.1000-0135.2025.12.001 Drawing on the development and transformation of Information Science since the implementation of the “14th Five-Year Plan,” and in response to the new requirements of the forthcoming “15th Five-Year Plan,” this study systematically examines the innovative development of Information Science in the era of data intelligence. First, it addresses several long-standing issues within the academic community by exploring the conceptual meaning of intelligence, disciplinary challenges, evolutionary trajectories, and the evolution of principal contradictions. On this basis, focusing on the innovation needs of the data intelligence era, this study proposes a “mission-driven” development logic and systematically assesses future trends in the discipline from aspects including mission deepening, paradigm integration, knowledge space, predictive intelligence, and value co-creation. Furthermore, this study outlines development suggestions and action strategies for disciplinary construction during the “15th Five-Year Plan,”centering on key agendas such as building a macro-intelligence system serving national strategies and governance, constructing an autonomous knowledge system of information science with Chinese characteristics, conducting scenario-based application services in vertical domains, and cultivating interdisciplinary talent adapted to the needs of the era. Ultimately, this study envisions that Chinese information science in the data-intelligence era should foster a distinct development paradigm. It presents the argument that, grounded in a robust autonomous knowledge system, the discipline should contribute a “Chinese Solution” with international exemplary significance for addressing global complexity challenges. 2025 Vol. 44 (12): 1493-1502 [Abstract] ( 5 ) HTML (83 KB)  PDF (797 KB)  ( 21 )

Intelligence Theories and Methods

1503 Research on the Identification of High-Value Patents in Industries Based on Technological Distance Measurement Hot! Ran Congjing, Jiang Yunlong, Li Wang, Jia Zhixuan, Cheng Fan DOI: 10.3772/j.issn.1000-0135.2025.12.002 The identification of high-value patents, which is crucial for seizing the global strategic high ground of industries and facilitating their sustained, efficient, and healthy development, can offer significant cues for the excavation of key core technologies within industries. This study builds its analysis from the perspective of patent technology distance measurement, and based on the extraction of upper-level themes through topic clustering, it proposes a method for identifying high-value patents under the dual-dimensional influence of knowledge contribution distance and connection degree. For the knowledge contribution distance, a hierarchical patent citation network among patents is constructed, and the continuous knowledge contribution value of each patent to the theme is computed. The dynamic time warping distance between themes is calculated based on the time series of knowledge contributions to form a theme knowledge contribution distance matrix. For the connection degree, a bipartite graph network of themes and patents is established, and the initial strength and global logic are calculated by integrating the co-occurrence frequency and citation relationship strength of patents to form a theme connection degree matrix. The patent technology distance matrix is constructed by fusing the dual-dimensional matrices. The absolute technology distance of each patent is calculated based on the technology distance matrix. Patents with high absolute technology distances within the threshold range are selected as high-technology-value patents within the field. Performance verification of the verified dataset shows that the accuracy and F1 score of the proposed method reaches 0.8218 and 0.8014, respectively. This study conducts an empirical study of the field of “generative artificial intelligence.” It identifies 1437 patents with high value in the industry, and found that these patents had a high transfer ratio of 58.59%. This study quantifies the technological gap between patents from the perspective of technological essence, overcoming the limitations of previous approaches that merely judged patent value based on external features or simple statistical data, enhancing the accuracy of identification. It also proposes a dual-dimensional technological distance influence mechanism to enhance the interpretability of identification. 2025 Vol. 44 (12): 1503-1522 [Abstract] ( 12 ) HTML (310 KB)  PDF (3516 KB)  ( 19 )

1523 Interaction Between Value and Risk of Cross-Border Data Flow Studied Through Dynamic Simulations Hot! Lu Can, Gao Hui, Yang Jianlin DOI: 10.3772/j.issn.1000-0135.2025.12.003 Analyzing the interaction between the risks and values of cross-border data flow is crucial for identifying the measures to maintain data sovereignty and promote digital economic development. In this study, data value chain theory was employed to analyze the value realization pathways of cross-border data flow and identify the associated risks. Subsequently, a “Value-Risk” dual unified simulation model was constructed using system dynamics. Finally, a multi-scenario analysis was conducted by simulating a complex real-world situation through a combination of technical, commercial, and government-related factors. This study revealed that the interaction between values and risks of cross-border data flow exhibits high-value, high-risk characteristics with increasing interaction intensity over time, which significantly impacts the cross-border data flow industry. Based on the multi-scenario analysis, expanding technological investment and engaging in technological competition can achieve a rapid increase in value but may also lead to increased risks in the absence of regulatory oversight. Active integration into the global data ecosystem can steadily promote the rise of value and reduce risks over the long term, by enhancing the right to formulate rules, encouraging private innovation while achieving effective regulation, and improving public data literacy, which makes it more suitable as a national strategy in promoting the medium and long term development of cross-border data flow industry. 2025 Vol. 44 (12): 1523-1538 [Abstract] ( 6 ) HTML (181 KB)  PDF (3395 KB)  ( 22 )

1539 Research on the Theoretical Model of Fundamental Data Semantic Knowledge Representation Based on Semantic Triangles and Related Theories Hot! Yuan Jingshu, Zhai Kexin, Yuan Man DOI: 10.3772/j.issn.1000-0135.2025.12.004 The ISO/IEC 11179:2023 Information technology—metadata registry (MDR) standard provides data semantic governance and management. Its research and application have triggered in-depth thinking about the intention of fundamental data semantic knowledge and its constituent elements. Based on this, this study adopts the traceability method to systematically identify the basic theories and standards related to the semantic organization and representation of data. Although Ogden and Richard’s semantic triangle and Dalberg’s conceptual triangle are classic core theories, they mainly explain the logical relationships between elements from a higher-order perspective, making it difficult to meet the current demand for fine-grained and rich data semantic knowledge representation. To this end, this study first takes two types of triangle theories as the core and constructs three fundamental conceptual semantic knowledge representation theoretical models and the conceptual semantic triangle integration model from a multidisciplinary theoretical perspective in the conceptual world, revealing human cognition of real-world things and the mechanisms of complex semantic organization and representation. Second, in the computer world, three types of fundamental metadata semantic knowledge representation theoretical models, semantic pyramid theoretical models, and basic semantic knowledge representation models based on MDR are constructed in sequence. The mapping, organization, and representation processes of conceptual semantics to metadata semantics in the computer are systematically explored, and the standardized representation of metadata semantics is achieved based on the MDR standard. Finally, the validity of the models is verified by constructing educational resource metadata and semantic description cases. The results of this study can provide theoretical references for the research and governance of data semantic standards in fields such as knowledge organization and data modeling. 2025 Vol. 44 (12): 1539-1553 [Abstract] ( 7 ) HTML (115 KB)  PDF (3469 KB)  ( 16 )

1554 Research on the Role of Science and Technology Policies in the Development of Disruptive Technologies Hot! Fu Junying, He Defang, Zhang Xu, Zheng Jia DOI: 10.3772/j.issn.1000-0135.2025.12.005 Exploring the role of science and technology policies in the development of disruptive technological innovations and effective approaches is of great significance. Taking DNA sequencing technology as an example, this study found that the full chain policy system for scientific and technological activities can play a key role in early high-risk research, in the development and application stages of disruptive technologies through policy instruments of the government by developing prospect predictions, preparing strategic plans, innovating funding modes, adjusting technology directions, configurating R&D performers, accelerating technology transfers and achievement transformations, implementing project evaluations, which result in a leapfrog development of disruptive technologies. Corresponding policy recommendations are accordingly proposed. 2025 Vol. 44 (12): 1554-1565 [Abstract] ( 8 ) HTML (110 KB)  PDF (2895 KB)  ( 20 )

Intelligence Technology and Application

1566 Research on Disruptive Technology Identification and Prediction Based on Deep Semantic Information Mining of Patent Texts Hot! Wu Lei, Zhou Shufa, Lin Chaoran DOI: 10.3772/j.issn.1000-0135.2025.12.006 This study addresses the limitations of existing methods in quantifying dynamic topic strength within deep semantic analysis, to enhance the accuracy and prediction of disruptive technology identification. We propose a multi-dimensional, dynamic model that couples bidirectional encoder representations from transformers topic (BERTopic) with dynamic topic modeling (DTM) and introduces “topic strength” to capture the emergence of technology topics. The model leverages the powerful semantic representation of the BERT language model to analyze inter-topic relationships and quantify their novelty using an adaptive thresholding algorithm based on Otsu’s method and quantiles. An empirical analysis of patent data from the integrated circuit (IC) industry demonstrates that our model identifies nascent and potentially disruptive technologies with greater sensitivity than traditional methods. Moreover, the integrated attention-based long short-term memory (LSTM) prediction module has a distinct advantage in terms of forecasting accuracy. This study offers a new paradigm for intelligence analysis of disruptive technologies by integrating semantic depth with a dynamic evolutionary perspective, thereby holding significant theoretical and practical value. 2025 Vol. 44 (12): 1566-1579 [Abstract] ( 8 ) HTML (148 KB)  PDF (5760 KB)  ( 23 )

1580 Analysis of the Influence Mechanism and Opportunities for the Convergence of Artificial Intelligence Technology and Traditional Industries from the Perspective of Multilayer Networks Hot! Wang Tao, Wang Jiajie, Kang Lele DOI: 10.3772/j.issn.1000-0135.2025.12.007 Exploring the influence mechanism and potential convergence opportunities of artificial intelligence technology and traditional industries is important for promoting the construction of modernized industrial systems and the convergence development of digital and real economies. Based on patent data in the field of convergence of artificial intelligence technology and traditional industries after in-depth screening, this study constructs a multilayer network of organizational technology and measures relevant indices, takes the field of unmanned aerial vehicles as a sample, and utilizes the exponential random graph model to empirically explore the influence mechanism of the convergence of artificial intelligence technology and traditional industries. It further identifies and analyzes the convergence opportunities based on the probability of connecting the edges of the fitted results of the model. The results of the study show that (1) the formation of a convergence network between AI technology and traditional industries is jointly driven by organizational cooperation and technological fusion, and the contribution of organizational cooperation may be effective only for innovation-advantageous organizations; (2) the mediating effect has a significant negative impact on the formation of convergence networks between AI technology and traditional industries; and (3) the probabilistic prediction method based on the results of the exponential stochastic graph model fitting can effectively identify convergence opportunities, specifically including the three technical themes of precision agriculture, intelligent inspection, and special operations. This study provides a reference for the construction of a convergent innovation ecology of artificial intelligence technology and traditional industries. 2025 Vol. 44 (12): 1580-1595 [Abstract] ( 7 ) HTML (181 KB)  PDF (3237 KB)  ( 26 )

1596 Research on Weak Signal Detection Models for Key Derivative Technologies along the Innovation Chain of Critical Core Technologies Hot! Ye Guanghui, Tu Kai, Han Li, Hu Lina, Xiong Bingqiao DOI: 10.3772/j.issn.1000-0135.2025.12.008 Continuously tracking the key derivative innovation technologies along the innovation chain of critical core technologies requires technological foresight. Only in this way can enterprises avoid having their research and development (R&D) agendas and scientific and technological (S&T) resource-allocation decisions of relevant agencies sink into a fog of technological information. Drawing on Hiltunen’s triadic model of the future sign and integrating complementary lenses, such as fitness landscape and information foraging theories, we constructed a multi-theoretically grounded weak-signal detection model for key derivative innovation technologies. Within this model, weak signal detection is operationalized as a three-stage process: signal visibility measurement, signal diffusion measurement, and signal sense-making. To address issues such as the isolated interpretation of weak signals and signal disappearance, a weak-signal detection model was applied at the thematic level to validate the model’s effectiveness and explore the evolutionary patterns of specific instance technologies. The critical core technology of Transformer in the field of artificial intelligence, was selected for empirical analysis. By acquiring and fusing heterogeneous data, the weak signals of key derivative innovation technologies were unearthed at different stages and their subsequent evolutionary trajectories were forecasted. From the second half of 2018 to the first half of 2020, Transformer developments clustered in natural-language processing, and several spin-offs, such as GPT-style models, were detected as weak signals. Between the second half of 2020 and the first half of 2022, Transformer advances shifted to computer vision, with weak signals detected for spin-offs, such as image classification. From the second half of 2022 to the first half of 2024, weak signals emerged in multimodal applications, indicating substantial future potential. The weak signal detection model advanced in this study not only enriches the theoretical landscape of technology foresight but also furnishes firms and agencies with a strategic compass for situating their R&D portfolios. 2025 Vol. 44 (12): 1596-1609 [Abstract] ( 9 ) HTML (184 KB)  PDF (4006 KB)  ( 26 )

1610 Multi-Source Data Fusion for Identification of Key Generic Technology: A Knowledge Graph and Deep Learning-Based Approach Hot! Zhong Yule, Yao Zhanlei, Xu Xin DOI: 10.3772/j.issn.1000-0135.2025.12.009 The precise identification of key generic technology is a pivotal component in the innovation and development of technology and plays a significant role in the strategic layout of industrial technology and investment decision-making. In this study, a novel method was developed for identifying key generic technology by integrating multi-source data. Specifically, a quantifiable key generic technology identification index system was developed based on the fundamental characteristics and features of key generic technology, encompassing five dimensions: universality, associativity, benefit, foundationality, and criticality. By leveraging the semantic representation and retrieval advantages of knowledge graphs, we trained a key generic technology identification model using bidirectional long short-term memory (BiLSTM) attention and employed bidirectional encoder representations from transformers (BERT) topic clustering to formulate a list of candidate key generic technology themes. Subsequently, by incorporating multi-source heterogeneous data, such as news and social media, we introduced technology-society interaction factors to scientifically determine key generic technology. Finally, an empirical study in the field of electrochemical energy storage was conducted to validate the effectiveness of our method. The results demonstrate that our approach provides valuable references for the identification of key generic technology across various domains, scientifically promoting strategic industrial planning and decision-making, which fosters the formation and development of new quality productive forces. 2025 Vol. 44 (12): 1610-1620 [Abstract] ( 8 ) HTML (140 KB)  PDF (2083 KB)  ( 25 )

Intelligence Users and Behavior

1621 Research on Information Cocoon Identification and Cocoon Breaking Topic Recommendation in Research Collaboration Groups Hot! Chen Xiang, Huang Lu, Cao Xiaoli, Ren Hang DOI: 10.3772/j.issn.1000-0135.2025.12.010 As collaborative relationships among researchers become increasingly entrenched, the emergence of information cocoons within scientific collaboration groups may hinder interdisciplinary integration and limit advancements in collaborative scientific innovation. This paper proposes a method for identifying information cocoons and recommending breakthrough research topics in scientific collaboration groups. First, a time-series, two-layer network comprising co-authorship and keyword semantic similarity was constructed. An incremental community detection algorithm is applied to extract the evolving community structure in the cumulative co-authorship network over time. Each author’s research topic vector was calculated based on the correspondence between the authors and keywords across a two-layer network. Information cocoons are identified by jointly considering topic homogeneity and novelty metrics. Second, an information dissemination influence measurement model is constructed to measure the potential of author nodes to break out of a cocoon. Then, a co-authorship - keyword semantic two-layer network considering the ranking of potential to break out of the cocoon is generated, and an author topic recommendation algorithm based on restarted random walk (ATR_RWR) is proposed to help researchers break out of the cocoon. An empirical analysis was conducted in the field of computer science to validate the effectiveness of the proposed method. 2025 Vol. 44 (12): 1621-1636 [Abstract] ( 7 ) HTML (241 KB)  PDF (2143 KB)  ( 13 )