10 April 2025, Volume 45 Issue 4
-
Select all|
-
LIBRARY TRIBUNE. 2025, 45(4): 1-12.This article attempts to examine the knowledge production function of libraries,focusing on the connotation and requirements of developing new quality productive forces during the 15th Five-Year Plan period. Driven externally by the demand for developing knowledge productivity from the macro-environment and internally by the function of library knowledge production,it provides a systematic review of the evolution of the knowledge production function of libraries against the backdrop of changes in the methods of knowledge production. Emphasizing its functional expansion in cultivating new quality productive forces,the study proposes countermeasures for building new knowledge productivity in libraries during the 15th Five-Year Plan period.
-
LIBRARY TRIBUNE. 2025, 45(4): 13-23.Innovation cooperation plays an important role in the high-quality development of regional economy and comprehensive strength. With the innovation ecosystem as the background,this study used niche fitness as the measurement index to systematically analyze the factors affecting patent cooperation,and built a niche fitness measurement model corresponding to the regional innovation ecosystem. The empirical analysis and mechanism test were conducted using the actual data of cities in the Guangdong-Hong Kong-Macao Greater Bay Area from 2002 to 2021. The niche fitness of the regional innovation ecosystem has a significant positive impact on patent cooperation. The innovation subject dimension exhibits a U-shaped effect on patent cooperation,while the innovation resource dimension and the innovation environment dimension have positive impacts on patent cooperation. The coupling coordination degree among the subsystems of the innovation ecosystem plays a significant positive moderating role in the relationship between niche fitness and patent cooperation.
-
LIBRARY TRIBUNE. 2025, 45(4): 24-48.Among the large number of murals,painted sculptures,silk paintings and other works of art left in the Mogao Grottoes of Dunhuang,one can find some pictures on reading,which vividly depict Buddhist activities related to reading,such as listening to the Dharma,reciting sutras,preaching the Dharma and praying for the deceased,and artistically represent the reading and learning of people in the Middle Ages. The reading materials depicted in these images,mostly sutra scrolls written on paper,silk and other soft materials that essentially replace the bamboo and wooden slips in Han stone portraits,reflect the generational change in the form of books. The images of reading from the Dunhuang Mogao Grottoes,although clearly specific to Buddhism,demonstrate the popularity of reading and educational activities in the Middle Ages,and testify to the public reverence for knowledge and books in the Northwest region,providing an important way of understanding the history of books,reading and culture in this period.
-
LIBRARY TRIBUNE. 2025, 45(4): 49-59.Taking the Treatise on Febrile Diseases and other ancient Chinese medical books as data sources,and according to their catalog structure,this paper creates a retrieval system for the formulas（or,prescriptions）recorded in ancient books of Traditional Chinese Medicine（TCM） by extracting formula information from these books through ChatGLM and storing it in a MySQL database. By parsing the prescription information through ChatGLM,and applying data mining algorithms such as Apriori,association_rules,community_louvain and other knowledge graph tools of Echart and Pyvis,it implements data mining on the ancient books and visualizes the knowledge graphs,thus establishing an LLM-based data mining system of TCM formulas. Based on the extracted formula information,it develops a knowledge Q&A system for TCM formulas on the BISHENG platform by using the retrieval-augmented generation algorithm . The results show that the recall rate for extracting formula names from ancient TCM books in this study ranges from 99.19% to 100%;the ROUGE-L values for the accuracy of formula composition,main functions,and usage are between 84.29% and 97.78%,except for Integrating Chinese and Western Medicine · Formulas;the accuracy of Chinese medicine name and dosage identification is greater than 98.00%,and the accuracy of main function parsing is greater than 86.00%. The data mining results are consistent with those of existing research;and the performance of knowledge Q&A is as expected.
-
LIBRARY TRIBUNE. 2025, 45(4): 60-70.From the perspective of memory theory,archives,books and other documentary resources not only directly restore human memories,but also can dynamically and multiply deconstruct and reconstruct human memories through semantic organization. By analyzing the intrinsic mechanism of documentary resources involved in memory construction,this study designs a semantic organization model of documentary resources that includes three symbiotic fields and ecosystems：individual memory,social memory,and cultural memory. Considering the structure and characteristics of different documentary units,it elaborates the specific paths and methods by which documentary resources participate in memory construction during the semantic organization process of information collection and processing,data extraction and association,knowledge aggregation and dissemination. On this basis,this paper conducts an empirical study of the literature and documents on the early workers’ movement in Qingdao to demonstrate the steps of memory retrieval and deconstruction,memory analysis and reconstruction,as well as memory enhancement and dissemination,and to explore the approach of semantic organization of documentary resources for memory construction.
-
LIBRARY TRIBUNE. 2025, 45(4): 71-80.Digital inclusion is an effective approach to bridging the digital divide. As a global concern,measuring the extent of digital inclusion has become a key issue. By means of inductive content analysis,this study examines 66 research reports on Internet development released by the China Internet Network Information Center （CNNIC）,focusing on four digitally disadvantaged groups：non-Internet users,senior citizens,young people,and rural residents. From the perspective of “inclusion framework-inclusion levels-target group”,it summarizes the measurement of digital inclusion for disadvantaged groups in the reports. This paper validates the effectiveness of the three-dimensional analytical framework for measuring digital inclusion by analyzing these reports. It also demonstrates a mutually supportive relationship between digital inclusion measurement and related policy development. However,there are some challenges with current measurements,such as limited coverage of disadvantaged groups and uneven distribution of measurement dimensions.
-
LIBRARY TRIBUNE. 2025, 45(4): 81-91.Exploring the process of science and technology （S&T） innovation and effectively integrating the key information in the process can provide the theoretical basis and practical support for the formulation and adjustment of S&T innovation strategy. Based on the multilayer network theoretical framework,the paper proposes a method for developing and analyzing the knowledge graph of S&T innovation strategy and conducts an empirical study on an enterprise in the Guangdong-Hong Kong-Macao Greater Bay Area. The multilayer network-based knowledge graph empowers the formulation,analysis,and adjustment of S&T innovation strategies,and provides a new perspective for interpreting the S&T innovation process through the analysis of patent litigation,the discovery of coopetitive relationships,the tracking of technology clusters,and the study of S&T development trends.
-
LIBRARY TRIBUNE. 2025, 45(4): 92-103.By providing highly relevant search results from a relational perspective,this paper achieves a breakthrough from linked data to indexing through a knowledge graph-based indexing system specifically designed for research papers on the history of the Communist Party of China（CPC）,shedding light on the improvement of search efficiency using knowledge graphs. Powered by Lucene,a full-text search technology,this study integrates the entities and their relationships in the CPC history knowledge graph with the text of research papers,enabling the positioning and browsing of relevant entities within the literature. Using the knowledge graph and the Lucene full-text search engine,it achieves full-text and paragraph retrieval for multiple entities,significantly enhancing the relevance and accuracy of search results. The experiment shows that the system achieves an accuracy rate of 85.30% in paragraph division for 4,625 research papers. The improved algorithm incorporating multi-factor ranking has enhanced the authority and quality of search results. These findings validate the feasibility and effectiveness of the system in retrieving literature on CPC history.
-
LIBRARY TRIBUNE. 2025, 45(4): 104-118.By systematically summarizing the presentations delivered by experts in eight forums on the topic of AI govenance at the 2024 World Artificial Intelligence Conference,this paper identifies the risks to artificial intelligence. Technically,there are challenges related to data,models,and systems;at the societal level,it faces ethical,ecnomic and environmental risks;at the international level,it has the issues such as global security,geopolitical and global inequalities,and uneven distribution of resources. The paper proposes the dual governance goals for artificial intelligence,which mainly include sustainable social development and people-oriented concepts;and suggests that AI should be promoted to make greater contributions to the sustainable development of human society through research on cutting-edge technological methods,the guarantee of social ethics,policies,and laws,the coordination among multiple institutions across varies industries and disciplines,and the deepening of global governance.
-
LIBRARY TRIBUNE. 2025, 45(4): 119-128.How to evaluate the value of data has become the key issue for the circulation of data elements. Model-based data value evaluation is a new and effective intelligent way. Aiming at the shortcomings of the current value evaluation methods based on untrained data,which cannot effectively utilize the holistic scene information beyond the data content and model features,this paper proposes a training-free value assessment method with a mix of command engineering and the existing untrained methods,which uses the large models to understand and utilize the holistic information such as task type and data attributes. Through a simulation experiment,it conducts a comparative analysis of the effectiveness,robustness and generalization of different methods under the mainstream datasets. The study suggests that the untrained evaluation approach of data value can be effectively improved without adding excessive computational cost in combination with instruction engineering of large model and the no-training methods.
-
LIBRARY TRIBUNE. 2025, 45(4): 129-138.In the context of human-AI collaborative information retrieval,the performance of AI-generated content （AIGC） directly affects the efficient and accurate fulfillment of retrieval goals. Prompt reformulation is one of the crucial strategies to optimize AIGC performance. Using the questionnaire method,this paper designs a simulated conversational searching scenario focusing on medical concepts to explore the effects of four types of prompt reformulation,namely language style,explanation method,context creation,and source constraint,on the perceived performance of content generated by ChatGPT 3.5. The results show that the types of prompt reformulation significantly influence the perceived performance of AIGC. Prompts characterized by vivid language style and illustrative descriptions can significantly improve the vividness and immersion of the generated content. Prompts that feature argument-by-argument discussion and sources limited to academic literature can significantly increase the perceived performance of the generated content,in terms of accuracy and professionalism. Creating a professional scenario is the most effective way to rewrite a prompt,and the perceived AIGC performance （of the experimental group） outperforms that of the control group on all evaluation indicators. This study aims to identify prompt reformulation paradigms that can enhance the perceived performance of the generated content,and guide users to express their information needs more accurately and effectively in the context of AI search,thus shedding light on the optimization of models for AIGC systems in professional knowledge search scenarios.
-
LIBRARY TRIBUNE. 2025, 45(4): 139-148.University libraries play an indispensable and central role in AI literacy education. By using online survey and induction method,this paper thoroughly explores the experience of foreign university libraries in AI literacy education,so as to provide insights for domestic university libraries to optimize AI literacy education and create an integration model of AI literacy education. The study identifies several key characteristics of AI literacy education in foreign university libraries：strengthening cooperation between internal and external institutions （for education subjects）;focusing on user guidance and librarian training （for education objects）,highlighting the improvement of basic literacy （in the form of education）,and emphasizing ethical and moral issues （in the content of education）. Based on the responses of foreign university libraries,the paper develops an integration model for AI literacy education,which outlines the corresponding strategies for the four "small integrations" of educational subject,object,form,and content,thus facilitating a "big integration" of AI literacy education among university libraries.
-
LIBRARY TRIBUNE. 2025, 45(4): 149-161.Exploring the rules for sharing and using public data in scientific research can help reduce the friction between the order of rule design and the flexibility and innovativeness of scientific research,and promote public data to better serve scientific research and innovation. Based on grounded theory,this study reviews the public data regulations for science and research included in policies such as data strategic planning and data utilization guidelines in Europe and the United States,and summarizes the main rules and behavioral boundaries. Attaching great importance to the access and use of public data by research communities,European and American countries prioritize and optimize the rules for the sharing and utilization of public data in the field of scientific research from the perspective of the whole research process,focusing on the form of data utilization,application process,charging rules,ethics and security,storage and reuse. Taking into account the current situation in China,this paper proposes some countermeasures in terms of "rule system,policy planning,and ecological co-creation" to regulate the use of public data in scientific research.
