[Purpose/Significance] The world is undergoing a new round of deep restructuring of scientific and technological revolution and industrial transformation, and scientific and technological innovation has become the core strategic variable that determines the reconstruction of the international competition order.Therefore, strengthening research on innovative intelligence will help us take the lead in the future competition in science and technology and seize the new commanding heights of science and technology. [Method/Process] From a theoretical perspective, the study revealed the theoretical sources of innovation intelligence in innovation opportunity identification, innovation knowledge integration, and innovation cooperation networks, and proposed the conceptual connotation of innovation intelligence.On this basis, from a practical perspective, the study analyzed the correlation and practical characteristics between innovation intelligence work and intellectual property intelligence, technology intelligence, and competitive intelligence. [Result/Conclusion] Based on the theoretical and practical needs of innovative intelligence, this study conducts research on the research approach of innovation intelligence, with a focus on addressing the theoretical and methodological issues of innovation intelligence research; The second focus is on exploring how innovation intelligence can promote high-level innovation, empower the national innovation system, and promote the improvement of new quality productivity.

[Purpose/Significance] This study aims to explore the mechanism of value co-creation in innovation intelligence services involving multiple stakeholders, so as to enhance the efficiency and quality of such services, and facilitate the realization of innovation intelligence value. [Method/Process] Based on the theory of value co creation and TOE theory, the paper explored the formation mechanism of innovation intelligence value co-creation from three aspects: environmental factors, technological factors, and organizational factors.Then the study employed system dynamics modeling to simulate and analyze the dynamic interactions between key factors and their various combinations. [Result/Conclusion] The results reveal that innovation intelligence product standards and implicit demand significantly influence the supply-demand matching level of innovation intelligence.Furthermore, the coordinated development of variables such as innovation intelligence institutions, innovation environment, intelligence technology investment, algorithm optimization, and innovation intelligence product standards has a significant impact on the value co-creation of innovation intelligence services.

[Purpose/Significance] A sound national innovation system is of great practical significance and era value for China to cope with the external scientific and technological blockade and master the leading power of the future high-tech industry in the new era. [Method/Process] The paper analyzed the concepts and research of innovation intelligence, national innovation system, etc., combed the evolution process of the construction of national innovation system enabled by intelligence work, and based on the innovation chain theory, analyzed the theoretical logic of national innovation system enabled by innovation intelligence from three aspects: innovation subject, innovation activity, and innovation process, thus putting forward the implementation path of national innovation system enabled by innovation intelligence. [Result/Conclusion] The pathways for innovation intelligence to empower the national innovation system primarily manifests in three aspects: fundamental research, development and application, and commercial industrialization.This pathway facilitates the optimization of knowledge production, dissemination, and application within the national innovation system, invigorating scientific and technological innovation vitality, thereby collectively advancing the high-quality development of China's sci-tech innovation.

[Purpose/Significance] Existing co-word networks are confined to capturing associations of knowledge entities within individual documents and struggle to model cross-document associations of knowledge entities.Furthermore, they lack semantic information.Current cross-document knowledge entity networks also exhibit certain shortcomings in modeling cross-document associations of knowledge entities.Therefore, it is necessary to explore more refined construction methods for cross-document knowledge entity co-occurrence networks. [Method/Process] The study integrated research from citation content analysis to propose a novel method for constructing cross-document knowledge entity co-occurrence networks based on citing text-cited span pairs.And the method distinguished between intra-document and cross-document associations of knowledge entities by introducing micro citation importance.Subsequently, the study conducted a comparative analysis of the constructed cross-document knowledge entity co-occurrence networks, traditional co-word networks and knowledge entity co-occurrence networks based on citing text. [Result/Conclusion] Empirical research in natural language processing indicates that the proposed networks exhibit a larger scale, enrich associations among knowledge entities, and demonstrate characteristics of sparsity and"small-world"to a certain extent.In terms of knowledge entity importance evaluation, the weighted degree centrality method applied to the constructed networks encapsulates richer information, demonstrates a higher correlation with the frequency-based method, and exhibits stronger discriminatory capability.Additionally, research topics identified through the constructed networks are more concise and cohesive.This study integrates methodologies stemmed from knowledge networks, entitymetrics, and citation content analysis and so on, thereby pushing the frontiers of these domains and fostering interdisciplinary fusion, and offers novel methodological support for relevant research.Furthermore, this study holds significant implications for advancing knowledge discovery and utilization within scientific literature, as well as boosting researchers' cognitive efficiency.

[Purpose/Significance] With the rapid growth of social media platforms, multimodal named entity recognition(MNER)has become a prominent research area in recent years.Although recent studies indicate that vision-and-language models using vision transformers outperform traditional methods based on object detectors, there is still a lack of systematic research on vision-and-language transformer models in MNER. [Method/Process] To address this gap, the paper introduced a novel end-to-end framework for designing and training of a fully transformer-based vision-and-language MNER model.The framework systematically explored key aspects of model design, including multimodal feature extraction, fusion modules, and decoding architectures. [Result/Conclusion] Experimental results show that the model surpasses all baseline models, including those using large language models, and achieves the highest overall metrics across two datasets.Specifically, the model achieved overall F1 scores of 80.06% on the Twitter-2015 dataset and 94.27% on the Twitter-2017 dataset, representing improvements of 1.34% and 3.80% over the state-of-the-art models, respectively.Additionally, the model demonstrates stronger generalization capabilities in cross-dataset evaluations compared to baseline models.

[Purpose/Significance] This study proposes a method to construct an emotional classification dataset for online government-public interaction texts and develop corresponding sentiment classification approaches based on the arousal-valence theory of emotion using large language models. [Method/Process] First, the paper established an emotion classification system with 8 emotional intervals and 56 emotion labels based on the arousal-valence theory.Next, the study utilized large language models to perform fine-grained emotional annotation on government-public interaction texts, constructing a high-quality sentiment classification dataset.Finally based on this, the paper proposed and systematically evaluated an emotion recognition model for this type of sentiment classification system. [Result/Conclusion] Experimental results validate the quality of the constructed dataset and the effectiveness of the sentiment classification method.This research provides reproducible data resources and technical solutions for fine-grained sentiment analysis in the field of online government-public interaction.

[Purpose/Significance] This study aims to uncover the evaluation orientations for regional scientific and technological innovation capacity by systematically analyzing previous studies and recent policy documents, in order to develop a comprehensive and scientifically robust theoretical framework that offers precise guidance for conducting evaluations, formulating policies, and promoting the development of regional scientific and technological innovation. [Method/Process] Using grounded theory as the primary research method, this study analyzed a total of 276 policy documents related to regional scientific and technological innovation, sourced from the official websites of various government levels in China, covering the period from 2019 to 2023. [Result/Conclusion] The paper identifies key evaluation orientations present within these policy documents and constructs a theoretical framework emphasizing four core orientations: foundational capability orientation, demand orientation, enterprise development orientation, and innovation environment orientation.A detailed analysis of these orientations is also provided.

[Purpose/Significance] Smart city construction relies on coordinated efforts across various domains, but faces policy coordination challenges stemming from a convergence of multiple interacting factors during transformation and upgrading processes. [Method/Process] This study explored and quantified textual topics within national and three provincial smart city related policies.Through the application of data visualization techniques and statistical methods, the study conducted an evolutionary analysis on the strength and structure of policy collaboration in smart cities. [Result/Conclusion] 27 policy topics contribute to coordination in smart city construction through top-level design, basic guarantees, integration and mutual promotion, public governance, and public services.The analysis reveals four stages of coordination strength: the blank period, initiation period, active period, and transformation period.The proportion of topics between central and local governments is generally consistent, but there are differences between regions.This study addresses gaps in quantitative research, integrity and systematicness of related studies, and also innovates research methods for analyzing the evolution of policy coordination intensity and structure.

[Purpose/Significance] The detection of key nodes in the spread of online rumors is of great significance for maintaining a clear cyberspace and promoting social stability.Addressing the issue of current research on key node detection overlooking the significance of nodes in the dissemination of multiple rumors, which often leads to false or missed detection of key nodes.This paper introduces a method for detecting key node groups in rumor propagation using a directed joint graph. [Method/Process] Firstly, by integrating the propagation trees of multiple rumors, a directed rumor propagation joint graph was constructed, providing a network structure diagram encompassing the spread of various rumors.Secondly, by analyzing the three-dimensional indicators of rumor publication, spread, and influence of nodes in the graph, this paper could quantify the importance of nodes.Finally, by calculating the coverage of node rumor information sorted by importance, key node groups in rumor propagation were detected. [Result/Conclusion] Empirical research is conducted on public datasets.Annotating and visualizing the dataset constructs a directed rumor propagation joint graph.The empirical results show that compared to the existing methods, the proposed method has a certain improvement in Precision, Recall and F1 value.

[Purpose/Significance] The study aims to reveal the formation mechanisms of online public opinion reversal in emergencies, assisting governments in effectively responding to sudden disasters and enhancing governance efficiency in public opinion management. [Method/Process] Focusing on natural disasters and accident disasters, this paper constructed a public opinion reversal formation model based on the WSR (Wuli—Shili—Renli) theory.The study employed NCA and fsQCA to conduct empirical analyses on 40 high-profile cases, and further uses the PSM method to analyze the impact of different configurations on public opinion reversal. [Result/Conclusion] The study categorizes three distinct formation pathways for online public opinion reversal during emergencies: object absence pattern, opinion leader-led pattern, and public-driven pattern.These pathways exhibit significantly different influence patterns.Specifically, analysis demonstrates that natural disaster scenarios predominantly experience opinion reversal through inherent event severity coupled with sustained public engagement, whereas industrial accident contexts show greater vulnerability to interventions from institutional authorities and key opinion leaders.

[Purpose/Significance] In the field of academic research, measuring the impact of academic achievements is a complex and important issue.Traditional assessment methods mainly rely on the number of citations, that is, "citation".However, with the rise of social media and online academic platforms, the"attention"of academic achievements has also become an important dimension for evaluation. [Method/Process] This study reviewed the current indicators based on citation and attention as well as research on the evaluation of academic influence, and constructed a comprehensive paper academic influence evaluation model by combining citation and attention indicators. [Result/Conclusion] This study suggests that the paper academic influence evaluation model is composed of four dimensions-potential influence, professional influence, disseminative influence and societal influence, and the model is operable to a certain extent, which has reference value for improving our country's scientific research evaluation system.

[Purpose/Significance] This study aims to identify the different roles of scientific papers within citation networks and analyze their mechanisms and impacts on knowledge diffusion from the perspectives of network structure and scale. [Method/Process] Based on papers published between 2011 and 2015 in five journals within the field of Information Resource Management(seed paper set)and their bidirectional citation data(reference paper set and citing paper set), a multi-layer citation network was constructed.The BERT algorithm was used to calculate the abstract similarity between the target paper set and the citing paper set, as well as between the reference paper set and the citing paper set.Papers were categorized into four roles—disseminators, initiators, intermediaries, and outsiders—based on differences in similarity scores. [Result/Conclusion] The roles of papers within citation networks significantly differ in their influence on knowledge diffusion.Disseminators and intermediaries play substantial roles in shaping the citation network structure and promoting knowledge diffusion.Disseminators, often review papers, act as knowledge navigators in the diffusion process, while intermediaries, typically theoretical or methodological papers, provide traceable knowledge pathways for subsequent research.Initiators, mainly representing theoretical or methodological innovations, inspire follow-up knowledge innovation and drive scientific progress.Outsider papers exhibit weak connections with their references and citing papers.

[Purpose/Significance] Measuring patent novelty is crucial for evaluating the quality of patent technology.Traditional citation-based methods for measuring patent novelty focus solely on the external characteristics of physical citations, failing to adequately reflect the genuine process of knowledge absorption during invention creation.Incorporating semantic citation information into the measurement of patent novelty can help reveal the knowledge and technological foundation of patents more comprehensively and accurately, thus providing support for fine-grained evaluation of patent novelty. [Method/Process] The study first used rules and syntactic analysis to extract knowledge elements from relevant patents and papers.Next, the study applied Sentence-BERT and Word2vec models to vectorize these knowledge elements and calculated the cosine similarity of vectors to determine semantic citations.Then, the study measured patent novelty from the dimensions of quantity, quality, and breadth of scientific and technical knowledge absorption, as well as the quantity and quality of technological impact.Finally, an empirical study was conducted in the field of quantum computing. [Results/Conclusion] Experimental results show that the proposed method can improve the accuracy and effectiveness of measuring patent novelty, thereby supporting patent examination and evaluation.

[Purpose/significance] This paper conducted quantitative evaluation and analysis of national open science policies formulated by European countries, with a view to providing necessary reference for China to formulate and improve its open science policies. [Method/process] This paper firstly selected national open science policies formulated by 11 European countries as research objects, secondly conducted quantitative analysis from three perspectives: policy themes, policy tools and policy effectiveness, thirdly constructed the policy evaluation index system and PMC index model according above analysis results, then carried out quantitative evaluation and comprehensive analysis of open science policy texts of European countries, and finally put forward suggestions for promoting the practice progress of open science policy in China. [Result/conclusion] The overall evaluation result of national open science policies of 11 European countries is excellent, and the shortcomings and references are found through above policy evaluation analysis.It was suggested that China should refer to and learn from the advanced policy practices of European countries as soon as possible, strengthen the top-level design of open science policy at the national strategic level, coordinate multiple policy subjects to formulate an open science policy system, and comprehensively use various policy tools to promote the progress of open science policy practice.

[Purpose/Significance] At present, no consistent conclusion has been reached on the influencing factors of government data service quality.The meta-analysis and fuzzy set QCA combination analysis are tried to draw the influencing factors of government data service quality, in order to provide reference for the high-quality development of government data service. [Methods/Process] Influencing factors of government data service quality were summarized through systematic literature screening, and the influence of different factors on the service quality of government data is revealed by using meta- analysis method.On this basis, 22 online provincial government data open platforms were selected as research cases, and the three configuration paths of government data service quality improvement were proved with the help of fuzzy set QCA method. [Results/Conclusion] The research results show that institutional support, personalized service, accuracy, infrastructure, convenience, perceived value and user trust are the key elements that affect the quality of government data service.System-trust driven, personalized-trust driven, system-personalized driven is the configuration path of service quality improvement.