Intelligence Theories and Methods

1395 Early Identification of Disruptive Technology Using Weak Signals Hot! Liu Junwan, Pang Bo, Xu Shuo DOI: 10.3772/j.issn.1000-0135.2023.12.001 An efficient set of techniques and ideas for the early identification of disruptive technologies is provided by combining a patent-based weak signal detection model and a disruptive technology potential measuring system. First, the topics of patent contents are extracted using from domain patent information using the latent Dirichlet allocation (LDA) topic model. We then use the weak function to filter the topics to collect the topics containing weak signals. The terms contained in the weak-signal topics are further filtered by a prognosis function to obtain the set of weak-signal terms and their corresponding patent collections to be related to their respective domains. Second, the technology disruptive potential measurement index system is applied to measure the disruptive potential of patents containing weak-signal terms. Finally, we obtain the collection of technologies with disruptive potential, which can serve as a reference for the early identification of disruptive technologies in the target domain. We apply this procedure to patent data in the incoPat patent database in gene editing research during the period 2008-2019 to identify the technologies with disruptive potential in this domain. After comparing the results with the weak-signal results based on a keyword text mining procedure, the above identification process and the results of weak signals are verified using CRISPR/Cas9 technology, demonstrating the feasibility and effectiveness of this method. 2023 Vol. 42 (12): 1395-1411 [Abstract] ( 242 ) HTML (214 KB)  PDF (2285 KB)  ( 340 )

1412 Dual-Dimension Scientist Evaluation Framework Based on the Disruptive and Consolidating Impact of Their Papers Hot! Yang Alex J., Kong Jia, Zhang Yiwei, Wang Hao, Deng Sanhong DOI: 10.3772/j.issn.1000-0135.2023.12.002 Using the rich mutual citation information in citation networks, the citation links of scientific papers can be categorized into two citation types: disruptive and consolidating. This study adopted a dual perspective to examine a scientist's disruptive and consolidating impact and proposes a measurement framework that divides scientists into two academic characteristics. We can obtain valuable insights by condensing various types of scientists into these two categories. To evaluate the consistency and effectiveness of this framework, we conducted an empirical analysis using the American Physical Society (APS) dataset, which comprises 463,348 papers, 9,370,286 citation links, and 234,086 post-disambiguated scholars. We employed Kendall's tau correlation, the identification proportion, and the average rank as evaluation metrics. The results indicate that disruptive citations and the disruptive h-index indicate a high consistency with traditional indices while surpassing benchmark indices in terms of the convergence efficiency. The disruptive influence of scientists serves as an indicator of their innovation level and potential. The dual measurement framework accurately and effectively captures the influence of scientists, providing a means for the early identification of future innovative researchers, reforming performance and reward systems, reviewing and evaluating funding projects, and formulating incentive policies for scientific research. 2023 Vol. 42 (12): 1412-1423 [Abstract] ( 182 ) HTML (151 KB)  PDF (3174 KB)  ( 207 )

1424 Quantitative Analysis of Data Governance Policies in China: Policy Evolution Path, Policy Subject, Sources of Policy, and Policy Tools Hot! Huo Fanfan, Huo Chaoguang, Ma Haiqun DOI: 10.3772/j.issn.1000-0135.2023.12.003 Data governance policy, by establishing the rules and criteria for data governance, is crucial for the development of China’s digital economy, especially regarding data production factors. This article integrates policy external structural elements and policy content elements to propose a policy analysis framework including the evolution path, policy subjects, sources of policy, and policy tools for the analysis of the 1097 central data governance policies. We use key policies and important events to depict the evolution path, analyze the policy subjects involved in policy-making, trace the sources of policy, and characterize the policy tool structure based on the policy coding. We find that some core departments of data governance policy making need to strengthen the awareness of data governance. The construction of National Data Bureau is an urgent necessity and a significant action in top-level design to which all departments should actively respond. Policy-making should ensure that authoritative policies strengthen the long-term effectiveness and impact of policy. Our coding analysis, based on the two dimensions of tool dimension and content dimension, identifies imbalances and deficiencies in the policy tools structure, showing that not only leading and controlling tools but also strong basic support and demonstration promotion measures are needed to ensure the health and vitality of data economic development. 2023 Vol. 42 (12): 1424-1437 [Abstract] ( 236 ) HTML (139 KB)  PDF (5147 KB)  ( 259 )

1438 The Influence of Fund Support on the Research Direction of Outstanding Young Scholars Hot! Chen Xueyi, Pan Yuntao, Ma Zheng, Zhang Guilan, Zhang Baizhen, Ren Qian DOI: 10.3772/j.issn.1000-0135.2023.12.004 Young people are the driving force behind scientific and technological innovations. The science fund influences the research direction of researchers. The science fund assists young talents in making breakthroughs in innovation and focusing on the research frontier. From the perspective of “environment, motivation, and behavior,” this study conducted research in three dimensions, including theme cohesion, cooperation continuity, and team stability. This study selected seven indicators to quantify the research direction. An empirical study was conducted using the winners of the “Beijing Outstanding Youth Science Fund” from 2018 to 2020 as an example. The following findings were obtained: research topics of the scholars supported by the foundation had become more extensive, collaboration with scholars in the same field had been continuously strengthened, and the internal academic team had become more stable. By observing the comprehensive score of outstanding youth in three stages and analyzing the change in research direction, it was found that 95% of scholars tended to focus on research under the funding. The focus types included gradual focus, slow focus, steady focus, and delayed focus. The rest tended to branch out, which was closely related to the evolution and continuous integration of disciplines. From the perspective of fund project management, scholars should be guided to conduct interest-driven and target-oriented research through fund projects. Moreover, scholars should be provided with stable and continuous funding projects to ensure their continuous, in-depth research. 2023 Vol. 42 (12): 1438-1447 [Abstract] ( 126 ) HTML (112 KB)  PDF (1644 KB)  ( 282 )

1448 Research on the Construction of National Security Intelligence System from an Interdisciplinary Perspective Hot! Wang Bing, Xu Fangting DOI: 10.3772/j.issn.1000-0135.2023.12.005 The national security intelligence system, a core component of national security, is its primary support mechanism. Studying this system from an interdisciplinary perspective not only highlights its multifaceted nature but also offers crucial insights for its modernization. First, this study highlights the interdisciplinary nature of the national security intelligence system and offers an understanding of it from an interdisciplinary perspective, based on the level of the discipline. Second, this study extracts the elements of the national security intelligence system from the perspectives of national security science, intelligence science, and system science. Finally, the study proposes paths for modernizing the national security intelligence system based on an interdisciplinary perspective. According to the research, from an interdisciplinary perspective, the elements of the national security intelligence system primarily include “the national security issue element, national security objective element, and national security work element from a perspective of national security science”, “the national security intelligence ontology element, national security intelligence management element, and national security intelligence technology element from a perspective of intelligence science” and “the subject element, relation element, and environment element in the national security intelligence system from a perspective of system science”. The paths for modernizing the national security intelligence system primarily include its totalization, digital intellectualization, collaboratization, and scientization. 2023 Vol. 42 (12): 1448-1457 [Abstract] ( 249 ) HTML (92 KB)  PDF (1103 KB)  ( 298 )

Intelligence Technology and Application

1458 Do Distance Factors Affect International Collaboration in Different Subfields Equally? Evidence from Computer Science Hot! Zhao Yi, Zhang Chengzhi, Xi Haixu DOI: 10.3772/j.issn.1000-0135.2023.12.006 Understanding the distance factors that influence international scientific collaboration is crucial for enhancing these collaborative efforts. However, previous studies have mostly focused on the investigation of top-level disciplines, ignoring the heterogeneity of distance factors affecting international collaboration in different subfields; therefore, the results of these studies cannot provide a basis for formulating refined policies. From a comparative perspective, this study analyzed the temporal and spatial patterns of international collaboration in the different subfields of computer science based on computer science publication records published by 187 countries between 1990 and 2019 from the database of DBLP computer science bibliography. Subsequently, we used a zero-inflated beta regression model to reveal the influence of six factors on international collaboration for a large set of countries, by different subfields, and over time. From the perspective of the spatio-temporal distribution of international collaboration, using artificial intelligence, a representative subfield of computer science, as an example, this study shows that the early stage of high-intensity collaborative relationships is mainly dominated by the United States. Countries, including China and Singapore, have become increasingly involved in artificial intelligence research over time, and the pattern of international collaboration has shifted from “one strong and multipolar” to “multipolar collaboration”. In addition, from the results of our regression analysis, we identify that geographical distance, cognitive distance, and economic distance obstruct international collaboration in all subfields, with cognitive distance having the highest impact, while cultural distance, degree of company participation, and political distance have significant negative effects only in some subfields. From the temporal dimension, the marginal effect of geographical and cognitive distances on international collaboration has decreased over time, whereas the impact of economic distances has increased. 2023 Vol. 42 (12): 1458-1476 [Abstract] ( 151 ) HTML (248 KB)  PDF (9755 KB)  ( 75 )

1477 Research on Fake News Detection Based on Multimodal Transformer Hot! Wang Zhenyu, Zhu Xuefang DOI: 10.3772/j.issn.1000-0135.2023.12.007 Fake news detection has been an essential area in natural language processing to reduce the negative impact of misinformation on society. Most existing multimodal fake news detection methods use pre-trained models to act as feature extractors; however, these methods have the following shortcomings: (1) Pre-trained model parameters are typically frozen during model training. However, it is crucial to note that these pre-trained models are not flawless; (2) CNN-based image feature extractor structures are typically more complex than Transformer-based text feature extractor structures, and because image features are typically stored in advance, the shortcomings of these models are negligible. Therefore, this study proposes a multimodal end-to-end Transformer, unifies the feature extraction process for different modalities by extracting image features using a vision Transformer rather than a CNN, achieves cross-fusion of image features and text features using a co-attention module, and conducts comparative experiments on three public datasets. The experimental results show that the performance of the model proposed in this study outperforms other baseline models. 2023 Vol. 42 (12): 1477-1486 [Abstract] ( 224 ) HTML (140 KB)  PDF (2622 KB)  ( 511 )

1487 ChpoBERT: A Pre-trained Model for Chinese Policy Texts Hot! Shen Si, Chen Meng, Feng Shuyang, Xu Qiankun, Liu Jiangfeng, Wang Fei, Wang Dongbo DOI: 10.3772/j.issn.1000-0135.2023.12.008 With the rapid development of deep learning and the accumulation of domain data, domain-based pre-trained models play an increasingly important supporting role in knowledge organization and mining. Aimed at massive Chinese policy texts, the pre-trained model of Chinese policy texts combined with the corresponding pre-trained strategies not only helps to improve the level of intelligent processing of Chinese policy texts, but also lays a solid foundation for the refinement, multi-dimensional analysis, and exploration of policy texts driven by data. For the national, provincial, and municipal policy texts, 131,390 policy texts with a total number of 305,648,206 Chinese words were obtained through the combination of automatic capture and manual assistance by removing non-policy text. This study develops a Chinese policy text pre-training model (ChpoBERT) for the constructed Chinese policy text corpus, which is based on the Chinese-RoBERTa-wwm-ext and BERT-base-Chinese. The model is open source and is available on Github. In terms of the evaluation indices of perplexity and downstream tasks of automatic word segmentation, automatic part-of-speech tagging, and named entity recognition of policy texts, the constructed ChpoBERT models showed better performance, which can provide basic computing resource support for the domain of intelligent knowledge mining of policy texts. 2023 Vol. 42 (12): 1487-1497 [Abstract] ( 234 ) HTML (155 KB)  PDF (1022 KB)  ( 395 )

1498 Research on Ancient Book Citations from the Perspective of Computational Humanities and the Construction of Full-Text Knowledge Base Hot! Liu Liu, Qi Yue, Liu Chufei, Li Wenqi, Wang Dongbo DOI: 10.3772/j.issn.1000-0135.2023.12.009 The citation of ancient books plays an important role in computational humanities research. They provide valuable insight into the context of traditional cultural inheritance. This paper explores the connotation of ancient book citations from the perspective of computational humanities, describes the citations of four types of ancient books, including catalogs, notes and commentaries, historical books, and poetries, and discusses the construction of a full-text knowledge base of ancient book citations. This paper describes the process of creating a full-text knowledge base of ancient book citations, which includes knowledge representation, annotation, completion, and disambiguation. Attempts are made to build knowledge bases for four types of ancient books, and the future application of the knowledge base is discussed. 2023 Vol. 42 (12): 1498-1512 [Abstract] ( 171 ) HTML (147 KB)  PDF (2402 KB)  ( 134 )