Purpose/significance The trend of social complexity and networking poses new challenges to cognitive decision-making practice，derivation and optimization. Embedding parallel cognition into the knowledge twin system creates opportunities for the development of intelligent society and the formation of knowledge productivity，and its cognitive split path expands new perspectives for cognitive adaptation to dynamic and complex situations，cross-decision making and evolutionary iteration. Method/process Based on the theoretical connotation of parallel cognition and knowledge twin system，this paper analyzes the realization path of parallel cognition-enabled knowledge twin system，constructs a parallel cognition-enabled knowledge twin system model，explores the whole process of cognitive splitting and analyzes the specific application scenarios and value embodiment. Result/conclusion The results show that parallel cognition is supported by descriptive cognition，predictive cognition and prescriptive cognition，and empowers the knowledge twin system and jointly constructs the cognitive split realization and value twin process. Through knowledge element-state mapping，knowledge group-state prediction，and collaborative cognitive iteration，this framework facilitates knowledge mapping cloning，knowledge demand forecasting，and knowledge derivative reconstruction. This constitutes a multidimensional extension pathway of “cognitive mapping - cognitive prediction - cognitive iteration”，demonstrating emergent advantages including virtual-real interaction，efficient decision-making，and collaborative evolution. The system provides dynamic impetus for three critical functions：cognitive adjustment and balance for knowledge entities，adaptation to complex decision-making environments，and deep human-machine coupling evolution.

Purpose/significance The large-scale dissemination of rumors in online social networks is very easy to cause public panic and decision-making bias，and it is of great theoretical and practical significance to clarify its network topology and dissemination mechanism for public opinion control. Method/process This study proposes an evolutionary game online rumor propagation model incorporating hypergraph structure. First，friend and community relationships in online social networks are portrayed through hypergraphs. Second，the infected person in the traditional infectious disease SIR model is subdivided into rumor spreader （I-state） and rumor refuser （T-state） to simulate the information dissemination process in a more detailed way. Once again，considering the internal and external factors of the individual，we construct the evolutionary game payoff function when facing rumor spreading，so as to portray the dynamic game mechanism between the I-state and the T-state. Finally，the applicability and validity of the model are verified through the case study and Monte Carlo simulation of the incident of “Nuclear Sewage Discharge Affecting Salt Safety in China”. Result/conclusion The results show that the increase in the size and number of online communities significantly expands the scale of online public opinion dissemination；the increase in the influence of online communities exacerbates the phenomenon of group polarization in the process of public opinion dissemination； and the ability of individual rumor recognition has a significant effect on suppressing the rumor dissemination. The evolutionary game online public opinion information dissemination model incorporating hypergraph can not only effectively characterize the complex interaction relationship in real social networks，but also accurately simulate the process of public opinion information dissemination and evolution. ［Innovation/ limitation This study employs hypergraph theory to better characterize the topological structure of social networks，proposes an improved SITR rumor propagation model based on the traditional SIR model，and introduces evolutionary game theory to describe the game-theoretic processes of network users during public opinion events. The current model does not fully consider the heterogeneity of roles within communities or the dynamic evolution of node relationships，simplifying the influence of communities on individual game-theoretic behavior. Future research could introduce mechanisms for node connection，disconnection，and reconstruction to construct dynamic hypergraph social networks，thereby more accurately depicting the information dissemination process.

Purpose/significance Exploring research topics in the field of Information Science and Library Science （ISLS） and analyzing their evolution paths in detail is of significant theoretical and practical value for understanding the discipline’s development，revealing the progression of its knowledge system，and guiding innovative advancements in the field. Method/process This paper uses data from CSSCI-indexed journals in the ISLS field published between 2019 and 2024. Based on initial literature grouping and direction extraction using hypergraph spectral clustering and TF-IDF，we apply a dynamic topic model to analyze research topics and their evolutionary paths in different groups，further uncovering the development trends and characteristics of each research topic. Result/conclusion The research directions in the ISLS field exhibit broad and diverse characteristics，which can be classified into nine categories，including online social and health information behavior，national intelligence and strategic competition，and open science and data policy governance. Through dynamic topic modeling，40 research topics were identified，with significant attention given to topics such as the evolution of public opinion dissemination，national emergency intelligence and strategic decision-making，academic influence evaluation，social media and privacy perception，and technology identification analysis. Moreover，there are intrinsic connections and intersections among topics across different research directions. From the perspective of topic evolution，the hot topics are closely related to social development，technological advancement，and policy guidance，exhibiting strong temporal characteristics. The integrated framework combining hypergraph clustering and dynamic topic modeling not only effectively captures the complex higher-order semantic structures within knowledge networks but also enables a fine-grained dissection of each research topic. Limitation This paper only employs hypergraph spectral clustering to differentiate literature groups，and dynamic topic modeling has limitations in capturing both long-term and short-term evolutionary characteristics of topics. Further research is needed to explore these aspects.

Purpose/significance To measure and analyze the development level of China’s data elements market and its regional differences，spatial dynamic evolution characteristics，and sources of structural differences. Method/process Compound entropy value method，Dagum Gini coefficient，kernel density estimation，and variance decomposition method are used to specifically analyze the development level of China’s data elements market and regional differences. Result/conclusion The development of China’s data elements market shows a steady upward trend overall，but significant regional disparities remain. The eastern region continues to lead，while the central，western，and northeastern regions lag behind，forming a decreasing gradient of “East–Central–West” and “Coastal–Inland” Inter-regional differences are the primary source of overall disparities and show a widening trend. Among the spatial dimensions，the physical space contributes the most to regional variation，though the contributions of each dimension vary across regions，indicating that the development of the data elements market remains insufficient and imbalanced.［Innovation/ limitation This paper is the first to construct an evaluation system of data elements market development level based on the “three-dimensional spatial theory”，and conducts an in-depth study on the overall characteristics，regional differences，spatial dynamic evolution and structural decomposition，etc. However，the study does not consider non-economic factors in depth. However，the study has not considered the long-term impact of non-economic factors on regional differences，and the prediction accuracy of the dynamic evolution model needs to be further optimized.

Purpose/significance As Open Public Data is a crucial component of data elements，this paper analyzes the evolution of symbiotic modes among Open Data entities from the perspective of symbiosis theory，which can facilitate efficient circulation and effective utilization of public data，thereby maximizing the value of data resources. Method/process Based on the SDAP theoretical framework，this paper first analyzes the evolution of Open Public Data platform scenarios. Next，it sorts out the key participants on both the supply and demand sides of Open Public Data，and focuses on analyzing the driving factors of Open Data from the demand-side perspective based on demand theory. The capability dimension evolves from static empowerment to dynamic enabling，solidifying the foundation for Open Data and propelling it toward a new path of mutually beneficial symbiosis. Finally，taking the SODA Contest as an example，this paper specifically examines the scenarios，willingness，capabilities，and processes. Result/conclusion The symbiotic mode of value co-creation subjects in Open Public Data evolves from the loose and accidental point symbiosis to the more frequent but discontinuous intermittent symbiosis，and then to the stable and continuous reciprocal symbiosis. Each stage has distinct characteristics and the mechanism is gradually improved to achieve in-depth data value mining and creation. Limitation This paper preliminarily explores the evolution of the symbiotic mode of multiple stakeholders in the value co-creation of Open Public Data. When analyzing the relationships among stakeholders，it expounds on the main symbiotic relationships among them under different modes. However，there is insufficient exploration of marginal stakeholders and potential implicit relationships among stakeholders，failing to construct a complete network of stakeholder relationships.

Purpose/significance The validity of academic evaluation involves its effectiveness，low validity will lead to evaluation failure，and the validity of multi-attribute academic evaluation has not received the attention it deserves. Method/process On the basis of defining the concept of validity of multi-attribute academic evaluation，this article conducts comparative analysis based on the types of multi-attribute academic evaluation，points out that linear subjective weighting evaluation method is a relatively effective evaluation method，and takes this method as the research object to analyze its impact mechanism and existing problems. Finally，a multi-attribute academic evaluation validity evaluation system and related policy recommendations are proposed. Result/conclusion Research has found that the impact mechanism of the validity of linear subjective weighting academic evaluation methods includes indicator selection，weight assignment，weight distortion，and standardization methods. The existing problems include improper selection of evaluation methods，incorrect selection of indicators，and issues with the application of linear evaluation methods. We should attach importance to the issue of academic evaluation validity，strengthen theoretical research on academic evaluation validity，and enhance the supervision and management of academic evaluation.

Purpose/significance Taking 82 graduate education majors from 14 iSchools in China as samples，the study investigated the curriculum of Master，MLIS and Doctor degrees in order to provide reference for promoting the development of relevant disciplines of information resource management in China. Method/process By means of network research and literature research，the iSchools website was retrieved and the website of iSchools in China was entered. The research was conducted on various types of graduate degree curriculum，the educational were analyzed，course classification standards were established and course category analysis was carried out. This paper analyzes the problems existing in the curriculum setting of information resource management in iSchools，and puts forward some suggestions. Result/conclusion It is found that there are the following problems in the curriculum setting of iSchools in China： the curriculum system is not complete enough，and the content of the same type of courses is overlapping. The classification educational mechanism is not perfect，and the curriculum of different degree types is homogenous. English literature research and autonomous learning ability are not integrated into the whole process of graduate education. There are few courses on career planning and vocational training. Finally，some specific suggestions are put forward.

Purpose/significance Conducting an in-depth exploration of the mapping relationship and response gaps between social demands and policy responses holds significant importance for optimizing the development strategy of autonomous driving technology. Method/process Through constructing an analytical framework of “social demands-policy responses” mapping relationships，this study employs BERTopic and LDA text mining techniques to conduct thorough analysis of Zhihu comment texts and policy documents. Utilizing the Jaccard similarity coefficient algorithm，a mapping relationship model was established to reveal the complex interactions between social demands and policy responses. Result/conclusion The findings reveal that social demands for autonomous driving technology present multidimensional complexity，encompassing seven core dimensions including technological reliability and safety，employment impact and social equity. The policy system exhibits multidimensional governance characteristics of “industrial cultivation-standard guidance-infrastructure support-risk control”. However，the coupling degree requires improvement in three core demand dimensions：employment impact and social equity，economic costs and inclusivity. Based on these findings，systematic countermeasures and suggestions are proposed from seven dimensions，including technical reliability，employment governance，and legal ethics.

Purpose/significance The quantitative analysis of China’s artificial intelligence（AI） data governance policies aims to improve AI data governance policy system in China and promote compliance and efficiency in the application of AI data. Method/process By using content analysis and statistical analysis，a three-dimensional analysis framework of “policy tools-data governance tasks-data life cycle" was constructed to quantify AI data governance policies issued by central and local government，analyze policy priorities and deficiencies and put forward suggestions for improvement. Result/conclusion There are some problems in relevant policies，such as the imbalanced usage of policy tools，the relatively concentrated deployment of governance tasks，and the uneven coverage of data life cycle. In the future，it is suggested to coordinate the proportion of policy tool usage，promote the rational deployment of governance tasks，and achieve a balanced layout of the data life cycle.

Purpose/significance The issue，as revealed by the public’s difficulty in locating specific policy content on government portals and traditional policy websites，reflects not only the low perceived accessibility but also the limited social value of disclosed government information. To address this issue，this paper proposes a policy knowledge graph framework based on multi-factor semantic relevance to improve data organization and enhance the social value of government information disclosure. Method/process Utilizing a knowledge graph as the technical framework，this study employs models such as BERTopic，KeyBERT，and BERT-BiLSTM-CRF to refine the granularity of policy text information. Semantic fine-grained information is then used to calculate the similarity and relevance of policy texts to identify related policies. The knowledge graph is constructed and stored using Neo4j. Result/conclusion Visualization and policy relevance tests demonstrate that the proposed semantically relevant knowledge graph provides more relevant and accurate policy information compared to traditional policy databases，meeting users’ needs for high-value information retrieval. Limitations The method is currently limited to policy text data. Future research could focus on improving the processing of multimodal government data，strengthening cross-domain data integration and application capabilities，and enhancing the applicability and generalizability of the constructed knowledge graph.

Purpose/significance Current research lacks a measurement tool for generative AI’s dialogue management capabilities from a user experience perspective. This study develops a specialized scale to address the limitations of technical performance metrics and generic user satisfaction questionnaires，providing scientific support for user experience optimization and product selection. Method/process Based on existing research，we developed an initial scale through semi-structured interviews and expert consultation. Using two-phase data collection （n=205/406），we validated and refined the scale through exploratory factor analysis and confirmatory factor analysis. Result/conclusion The final validated scale contains 20 items across four dimensions： context continuity，intent understanding，content generation，and module coordination，demonstrating good reliability and validity. This integrated measurement tool captures multi-dimensional dynamic attributes from a user experience perspective，enabling systematic evaluation of generative AI’s dialogue management capabilities，while offering standardized instruments for academic research and industrial applications to enhance human-AI collaboration effectiveness.

Purpose/significance To provide theoretical support and quantitative analysis tools for understanding the impact of social bots on human users’ opinions by investigating the mechanism of opinion evolution under human-bot collaboration on open platforms. Method/process Firstly，based on the interaction characteristics of social bots，their decision-making model is simplified into a constant function. Then，a group opinion evolution model is constructed using Markov chains. Finally，the “echo chamber effect” is tested through the group opinion heterogeneity index，and the impact of different factors of social bots is explored by analyzing collaborative interaction data from Sina Weibo related to “BYD Han”，thereby verifying the model’s validity. Result/conclusion The model can accurately simulate the evolution of group opinions. There is a negative correlation between the proportion of social bots and the heterogeneity of group opinions. An increasing number of positive sentiments reviews from social bots further reduces heterogeneity. The involvement of social bots reduces heterogeneity，and earlier intervention is more likely to facilitate the formation of an “echo chamber effect”.

Purpose/significance In order to better grasp market dynamics，predict future trends，and guide corporate product development，and strategic decision-making，it is necessary to analyze the influencing factors of the evolution of product weak demand signal. Method/process This research is based on the stochastic resonance model and constructs a model of the influencing factors of the evolution of weak demand signal by taking the influencing factors such as signal cognition，regional factors，potential difference，netizen attitude values，and noise as variables. Then，simulation experiments are conducted for each influencing factor respectively，and the rationality and feasibility of the model of the influencing factors of the evolution of weak demand signal are verified using data to confirm the role of each influencing factor in the evolution of weak demand signal. Result/conclusion The empirical results show that there is a positive relationship between signal cognition and weak demand signal evolution resonance. Within a certain range，the regional heat will dramatically affect the evolution of demand weak signal，but after reaching a certain extent，the increase of regional heat will not have much impact on the peak resonance of demand weak signal evolution. There is a positive relationship between potential difference and weak demand signal evolution resonance. The attitude value of netizens is positively correlated with the resonance evolution of weak demand signal. The relationship between noise and weak demand signal evolution resonance is first positive and then reverse.

Purpose/significance By accurately identifying derivative topics within extensive public opinion data and interpretably analyzing their evolutionary process，it can effectively explore dangerous topics and achieve targeted control over public opinion. Method/process This paper develops a Graph-BERTopic topic model by integrating deep learning and complex network methods. Initially，by enhancing the output correlation of the model，the performance of topic clustering is significantly improved. Next，we deploy topic embedding vectors，grounded in semantic similarity，to build a clustering graph that reflects the intricate relationships between topics. Finally，the derived topics in the community discovery detection graph are used，and the shortest path algorithm is employed to capture the evolutionary relationships between topics. Result/conclusion Experimental validation was conducted on the “China Eastern Airlines MU5735 Aircraft Crash” dataset，yielding superior clustering performance compared to diverse benchmark models，with NPMI and TD reaching 0.187 and 0.873，respectively. The quality function of the derived topic structure，divided by the clustering graph，reaches 0.831. The constructed model can accurately mine derived topics from large-scale texts and interpretably capture the evolution process between public opinion topics.

Purpose/significance Based on the SAO-QFD model，this study constructs the technological evolution path of emerging industries，deeply explores the driving factors of efficacy in the process of technological evolution，and reveals the internal laws of technological development. Method/process Firstly，the SAO-QFD model was used to extract technical information and efficacy characteristics，and the key technologies were identified according to the technology-efficacy association. Secondly，combined with the technical similarity and efficacy quantitative analysis，the efficacy-driven technology evolution path is constructed. Finally，the method is verified by taking lithium-ion battery as an example. Result/conclusion Cathode materials have undergone an evolution from lithium cobalt oxide to lithium iron phosphate，ternary materials and high-nickel ternary materials，which have promoted the improvement of energy density，service life and safety. The anode material has been developed from graphite to silicon-based composites，which has improved capacity and stability. In general，the technology evolution path driven by efficiency not only reflects the iterative upgrading of the technology itself，but also reveals the continuous role of power efficiency characteristics in promoting the development of technology，and the two synergistically promote the optimization of the overall performance of the battery. Limitations In this study，a quantitative mapping relationship between technological progress and actual efficacy changes has not been established，and it is difficult to visually present the specific impact of technological changes on efficacy indicators.

Purpose/significance The formation of critical core technologies relies on the R&D investment of innovation entities and the value validation through industrial applications. Therefore，identifying critical core technologies from the perspectives of technological formation and value enhancement is of significant importance. Method/process A three-dimensional system is constructed by integrating the R&D，technology，and application layers into a multi-layer network. The dynamic value of nodes within each layer is calculated using a time-value function，and the value-driven intensity between layers is measured through dynamic transfer entropy. The ML-Rank algorithm is applied to facilitate value transmission. Key metrics for identifying critical core technologies are established based on the value of technological nodes，the value contribution of the R&D and application layers，and their distribution trends，enabling precise technology identification. Result/conclusion The results indicate that the technologies identified by the proposed method align with the characteristics of key core technologies and exhibit significant potential for technological advancement and practical applications.

Purpose/significance Analyzing the dissemination effect of provincial libraries’ WeChat official accounts and understanding their current development status can provide valuable insights and targeted suggestions for the future development of provincial and municipal library official accounts. Method/process This study evaluates the communication effectiveness of provincial libraries’ WeChat official accounts in 2024 using the Super-Efficiency SBM model. By analyzing high-readership articles with NewRank and Qingbo Index data，it identifies key influencing factors through Bootstrap regression，ultimately proposing targeted strategies to enhance communication efficiency. Result/conclusion The communication effectiveness of provincial libraries’ WeChat official accounts in China exhibits pronounced regional disparities，with southeastern and central developed regions demonstrating superior performance，while central-western and remote areas face communication challenges. We recommend adopting tailored approaches to optimize resource allocation，enhance technical capabilities and content development，ultimately establishing targeted communication frameworks that adapt to local characteristics. Limitations This study has limitations in sample hierarchy and metric dimensions：Provincial-level library data inadequately reflect regional disparities，while efficiency evaluation overemphasized dissemination metrics without incorporating critical variables like operational team competencies. Future research should establish multi-tier data networks and longitudinal research frameworks to enhance explanatory power.

Purpose/significance Freedom to Operate （FTO） analysis runs through the process of product development to market launch and is an important means of identifying patent infringement risks. This paper proposes a multidimensional knowledge reorganization method driven by situational flow and constructs an FTO analysis mechanism based on the interaction and dynamic adaptation between situational flow and knowledge reorganization，aiming to overcome the limitations of the existing methods. Method/process Taking the project of emergency rapid evaluation of slow cracking resistance of polyethylene pipes as an example，focusing on the pre-development and in-development situations，the paper demonstrates the process of formation and evolution of situational flow by mining the dynamic changes of demand，information，and decision-making elements in the situations and constructs an adapted knowledge reorganization system in order to enhance the systematic and holistic nature in the automated analysis of FTO. Result/conclusion The multi-dimensional knowledge reorganization method driven by situational flow connects multi-situation knowledge management and integration，solves the problem of systematic intelligence demand and efficient intelligent analysis in the whole process of FTO，and at the same time expands the application of the situational flow theory and knowledge reorganization method in the research of intellectual property intelligence. Limitations Taking part of the FTO situations as examples，the paper has not yet covered the whole process or other domains of multi-situation system analysis，and subsequent research can be extended to validate the generalizability of the methodology.

Purpose/significance In response to the high annotation costs and poor domain adaptability of traditional knowledge graph construction methods，this research aims to explore an efficient approach for knowledge extraction and structured representation of red cultural resources under low-resource conditions. Method/process This study proposes a method for constructing a knowledge graph of red historical figures based on large language models （LLMs） and LoRA fine-tuning. Utilizing parameter-efficient fine-tuning strategies，we designed appropriate information extraction prompt templates and constructed instruction datasets to perform few-shot LoRA fine-tuning on four mainstream foundation models，GLM4-9B，DeepSeek-R1-Distill-Qwen-7B，Meta-Llama-3.1-8B-Instruct，and Qwen2.5-7B，comparing their effectiveness in joint entity and relation extraction. Result/conclusion Experimental results demonstrate that Qwen2.5-7B achieves optimal performance in joint entity and relation extraction，reaching a precision of 95.10% with 100 training samples. The fine-tuned Qwen2.5-7B was then applied to information extraction，knowledge graph construction，and visualization analysis for the Northeast Counter-Japanese United Army. By proposing a novel few-shot information extraction method for red historical figures，this study aims to provide reference and guidance for efficient knowledge graph construction in related fields under low-resource and low-cost conditions.

Purpose/significance Addressing issues such as the disconnection between theory and practice in red resource digital storytelling，weak interactive experience，and insufficient integration of digital technologies，this paper proposes an innovative approach to red resource digital storytelling by combining dual-layer retrieval-augmented generation technology. It provides a research solution and technical roadmap for “telling good red stories and inheriting the red gene.” Method/process Firstly，a large language model is used to construct a text index based on a graph structure，converting red resources into structured graph data. Secondly，user needs are mapped to the graph data，and a dual-layer retrieval system is applied to the red resource knowledge base，enabling both low-level fine-grained and high-level macro-level retrieval to enhance the application of large language models in historical resource storytelling. Finally，by comparing the dual-layer retrieval-augmented generation technology with traditional retrieval-augmented generation technology and graph retrieval-augmented generation technology，the effectiveness of this implementation in red resource digital storytelling is validated. Result/conclusion By integrating large language models with dual-layer retrieval-augmented generation techniques，this approach optimizes the content of digital storytelling and innovates narrative structuring through generative logic. It enables interactive online storytelling and demonstrates the practical value and application prospects of generative AI technologies in the digital storytelling of red resources.