10 August 2026, Volume 46 Issue 8
-
Select all|
-
LIBRARY TRIBUNE. 2026, 46(8): 1-14.In the digital and intelligent era,artificial intelligence （AI） technologies have become deeply integrated into the field of scientometrics. A systematic review of current research on AI-empowering scientometrics is of practical significance for promoting innovative development in the field and provides a reference for future studies. Based on journal articles from the Web of Science and CNKI databases,this study employs bibliometric and content analysis methods to examine,from a global perspective,the current state of AI-empowered scientometrics and to project its development trends. It presents a theoretical framework to advance AI-empowered scientometrics research and proposes six major research prospects：optimizing measurement techniques and methods,enhancing full-text metric accuracy,advancing intelligent academic evaluation,improving scientific prediction reliability,increasing academic communication efficiency,and standardizing risk governance.
-
LIBRARY TRIBUNE. 2026, 46(8): 15-25.The 15th Five-Year Plan identifies the development of an integrated nationwide data market as a core strategy for unlocking the potential of data elements and building new competitive advantages. It is imperative to systematically examine the theoretical logic,practical challenges,and implementation pathways for establishing such a market. Integrating institutional economics,the data value chain theory,and multi-level governance frameworks,this paper proposes a three-dimensional theoretical framework —centered on property rights definition,value circulation,and order maintenance— to reveal the intrinsic logic and systemic interdependence in building a nationally integrated data market. Analysis reveals three structural challenges currently facing China's data market：high transaction costs stemming from an ineffective property rights system,broken value chains caused by fragmented governance structures,and constrained market vitality due to an imbalance between security and regulatory effectiveness. These challenges are intertwined and collectively hinder the optimal allocation and value realization of data elements across the country. Accordingly,this study proposes a three-phase breakthrough strategy：implementing practical measures for separating the three types of data property rights to strengthen the institutional foundation for market operations;innovating mechanisms that combine technology empowerment with incentive compatibility to enhance market vitality;and establishing a transaction supervision and risk prevention system to optimize the market ecosystem and strike a dynamic balance between security and development.
-
LIBRARY TRIBUNE. 2026, 46(8): 26-36.Research project proposals represent the crystallization of scientific exploration undertaken by researchers,contain specialized content and critical method knowledge elements across various fields. Since the advent of the data-intensive scientific research paradigm,theoretical logic,data models,implementation technologies,and application verification have together constituted the core of disciplinary breakthroughs. From a data-driven perspective,this paper explores the characteristics of disciplinary distribution and combination patterns of method knowledge elements within proposals submitted to the National Natural Science Foundation of China （NSFC）. Firstly,knowledge elements are extracted and classified,followed by an analysis of the statistical characteristics of their disciplinary distribution. Next,a co-occurrence network of such knowledge elements is established to identify their combination preferences. Finally,combination patterns and effects are uncovered by modelling node and edge attribute data with graph convolutional networks. The results reveal that basic-class knowledge elements account for a higher proportion and possess prominent scientific value. The co-occurrence of method knowledge elements exhibits a radiation pattern centered on the dual core of basic- and computational-classes. Basic-class knowledge elements can effectively and consistently enhance combinatorial performance;and there is a significant information complementary effect among different classes of method knowledge elements. The combination effect of the basic- and technical-classes shows obvious disciplinary heterogeneity,with the most distinctive differentiated performance observed in the Earth Sciences.
-
LIBRARY TRIBUNE. 2026, 46(8): 37-49.The lack of datasets to support the study of the overall circulation of ancient Chinese books arises because traditional bibliographies only record metadata such as titles,authors,editions,seals,and colophons. Reconstructing the complete transmission history of ancient books requires meticulous examination of each seal and colophon,a task that demands both philological expertise and digital skills. This remains a major challenge for which no significant breakthroughs have yet been made. To address the data-scarce predicament,this study establishes a specialized annotation platform and standardizes the data annotation workflow. It develops a dataset on ancient book circulation history to better meet both book-centered and collector-centered research needs,and creates a database capturing the relationships among collectors,books,and seals,thereby effectively supporting the intelligent management of ancient book resources.
-
LIBRARY TRIBUNE. 2026, 46(8): 50-64.This study proposes a method for representing and serving knowledge from ancient Chinese texts that integrates an event logic graph with a large language model,with the aim of promoting the revitalization and utilization of ancient texts while lowering reading barriers for users. First,a knowledge representation model for ancient texts that incorporates an event logic graph is developed to identify entities,events,and their complex relationships in the texts. Then,the event logic graph is integrated with a large language model,and an external event-logic knowledge base is introduced using the retrieval-augmented generation （RAG） technique to establish the ELG-LLM model. Finally,the knowledge service performance is comprehensively evaluated through a case study of the Records of the Grand Historian：The Twelve Basic Annals （Shiji·Shier Benji）. The automatic evaluation results show that ELG-LLM significantly outperforms general-purpose large language models,traditional knowledge-graph-enhanced models,and domain-specific models fine-tuned on large-scale classical text corpora across multiple tasks;moreover,its performance is statistically more significant than that of knowledge-graph-enhanced models. The results of manual evaluation further indicate that ELG-LLM demonstrates clear advantages in content accuracy,sentence fluency,and answer completeness. In addition,a comparative experiment between “static fact retrieval” and “dynamic logical reasoning” reveals that the event logic graph is superior to the knowledge graph in addressing historical questions involving complex logic and evolutionary processes. By employing event logic graphs for retrieval augmentation,this study provides a feasible technical approach for the revitalization of ancient textual resources and their contemporary dissemination.
-
LIBRARY TRIBUNE. 2026, 46(8): 65-77.To address the challenges faced by traditional methods in developing large-scale knowledge graphs for ancient texts,such as reliance on manual annotation,limited semantic parsing capabilities,and low processing efficiency,this study integrates large language models with AI Agent technology to propose an automated knowledge graph generation framework for ancient books. Based on AI agents capable of autonomous planning,memory,and acting,the framework presents a core workflow comprising content parsing,knowledge extraction,data cleaning and transformation,knowledge mapping and visualization. An empirical study using the entry “Zhong” （loyalty） in volume 481 of the Yongle Encyclopedia successfully achieved the end-to-end automatic conversion from unstructured text to structured knowledge graphs. The results show that this method significantly improves the efficiency and accuracy of ancient book knowledge processing,enables natural language interaction and dynamic adjustment of user-defined triples,and provides a feasible technical pathway for the efficient and automated development of ancient book knowledge graphs.This study not only provides a scalable solution for the digital preservation of the Yongle Encyclopedia,but also proposes an agent-driven paradigm and a hierarchical prompt engineering system with strong transferability,which can significantly reduce the cost of knowledge graph development across various fields and promote the transformation of digital humanities research towards dynamic cognitive systems.
-
LIBRARY TRIBUNE. 2026, 46(8): 78-90.Artificial intelligence has acquired the technical capability to partially replace human experts in knowledge-intensive fields,sparking complex reactions in online public opinion. From the perspective of online public opinion and based on recent survey data of Chinese and American netizens,this paper employs various statistical methods to systematically examine the overall attitudes,structural characteristics,and related factors regarding “AI replacing experts” among netizens in both countries. The findings reveal that：netizens in the two countries generally hold a moderate level of attitudes toward AI replacing experts,with similar internal structures;Chinese netizens have a stronger attitude toward“AI replacing experts,showing distinct differences in attitudes between humanities and science experts;the notion that liberal arts are useless significantly influences the attitudes of both Chinese and American netizens;information behaviors such as media usage demonstrate complex and heterogeneous effects. Based on these empirical results,this study suggests strengthening AI public opinion governance and expert image building,while promoting human-machine collaboration,specifically the application scenarios and enhancement pathways of the “Expert + AI” model.
-
LIBRARY TRIBUNE. 2026, 46(8): 91-105.With regard to the application of digital human technology to public policy communication,this paper compares the effectiveness of human and digital streamers and examines the roles of psychological distance,anthropomorphism,and credibility endorsement,providing guidance for optimizing policy communication in the digital era. Through an online experiment using a between-subjects design,we systematically manipulated the type of communicator （human vs. digital human）,the degree of avatar anthropomorphism （high vs. low）,and the level of credibility endorsement （high vs. low）. Participants were recruited online,and their psychological distance and policy compliance intentions were measured after watching policy videos. Mediation and moderation models were tested using the PROCESS macro and ANOVA. Results show that human anchors are more effective than digital anchors at positively fostering the public's intention to comply with policies;psychological distance serves as a partial mediator;high anthropomorphism reduces the psychological distance gap between the two types of anchors;strong credibility endorsements can bring highly anthropomorphic digital anchors closer to human anchors in terms of psychological distance,but those with low anthropomorphism remain at a disadvantage. The effectiveness of digital anchors in policy communication is the result of a complex interplay among communicator characteristics,external empowerment,and audience psychology.
-
LIBRARY TRIBUNE. 2026, 46(8): 106-116.The rapid development of AI-generated video large models provides key technological support for the digital and intelligent transformation of libraries in China. How to harness the benefits of this technology while effectively managing risks,based on the actual conditions of Chinese libraries,has become an imperative issue demanding immediate attention. Taking Seedance 2.0 as the primary research object and Sora as the technological benchmark,this paper employs a comparative analysis method to reveal the differences between the two and assess their adaptability to Chinese libraries. It also develops application scenarios for Seedance 2.0 and proposes risk governance pathways. The study finds that,with localized advantages in Chinese context understanding,multimodal instruction integration,ease of use,and deployment requirements,Seedance 2.0 can play a significant role in scenarios such as activating collection resources,promoting reading,enhancing user literacy,creating digital spaces,and providing inclusive services. To address the potential risks related to knowledge authenticity and copyright ownership,it is recommended that governance measures be implemented across various areas—including video quality assessment,institutional safeguards,technical prevention,ethical guidance,and ecosystem development—to achieve a balance between technology application and risk prevention.
-
LIBRARY TRIBUNE. 2026, 46(8): 117-129.With the rapid development of artificial intelligence （AI） and multimodal technologies,visualized reading tools have gradually become an important supplementary measure for children's reading education. This study focuses on the behavioral mechanisms of children's use of such tools. Drawing on the push-pull transfer theory （PPM） and the stimulus-organism-response theory （SOR）,this paper presents an integrated theoretical framework encompassing environmental stimuli,psychological mechanisms,and behavioral intentions,and conducts a comparative analysis between typically developing children and those with visual impairments. User comments and questionnaire data from parents were collected;sentiment analysis was conducted using the ERNIE 3.0 model,and hypothesis testing was performed via regression analysis. The results show that pull factors （perceived usefulness and content richness） significantly promote learning engagement,which in turn enhances continuance intention;push factors （commercial interference and technical issues） significantly increase cognitive load,thereby reducing continuance intention;and accessible design significantly reduces cognitive load for visually impaired children and indirectly promotes reading behavior by increasing learning engagement. Analysis of group differences shows that visually impaired children are more sensitive to commercial interference,while typically developing children have a lower tolerance for technical issues. The study validates the effectiveness of the PPM-SOR model in researching children's usage behavior of AI reading tools,reveals the mediating mechanism of learning engagement and cognitive load,and provides a basis for developing differentiated strategies to optimize content,improve business models,enhance technical stability,and ensure accessibility for AI reading tools.
-
LIBRARY TRIBUNE. 2026, 46(8): 130-145.With the wide application of artificial intelligence,AI-based fraud schemes are constantly evolving. Due to differences in cognitive abilities,emotional trust tendencies,and behavioral habits,the elderly face greater risks when obtaining and responding to anti-fraud information. To address the issue of insufficient precision in AI-driven anti-fraud information services for the elderly,this paper employs text analysis and empirical analysis to characterize the information needs of the elderly regarding fraud prevention. By combining survey data with the Analytic Hierarchy Process,the study analyzes the weights of different characteristics and generates AI-driven anti-fraud information recommendation rules for the elderly based on IF-THEN logic. The results indicate that behavioral characteristics dominate the demand for anti-fraud information,emotional characteristics play a significant moderating role,and cognitive characteristics serve as a fundamental constraint. The resulting recommendation rules provide practical guidance for the design,dissemination formats,and channel selection of AI anti-fraud information content for the elderly.
-
LIBRARY TRIBUNE. 2026, 46(8): 146-160.This study explores the influencing factors and configuration paths of users' willingness to pay attention to cultural heritage elements in commercial video games,providing insights for facilitating the high-quality dissemination of fine traditional Chinese culture. Based on the stimulus-organism-response （SOR） model,the study employs structural equation modeling （SEM） and fuzzy-set qualitative comparative analysis （fsQCA） to examine,from a cognitive psychology perspective,the key influencing factors and configuration paths affecting users' attention to cultural heritage elements in video games. According to SEM analysis,symbol design,content design,topicality of elements,and system presentation all influence users' willingness to pay attention by affecting their aesthetic enjoyment and cultural identity,with aesthetic enjoyment and cultural identity acting as full mediators. Moreover,the effect of aesthetic enjoyment is stronger than that of cultural identity. The fsQCA reveals six configurational pathways and two trigger patterns leading to users' willingness to pay attention,in which content design,system presentation,aesthetic enjoyment,and cultural identity play important roles.
