Pub Date : 2025-08-26DOI: 10.26599/IJCS.2025.9100005
Hang Liu;Xuan Liu;Baowen Sun;Jiayu Wang
In the digital economy era, the rapid expansion of internet platforms has resulted in highly concentrated market structures in online markets, thereby eliciting intensified scrutiny from regulatory authorities. In this paper, we aim to explore the key factors that shape the boundaries of platform firms by extending the transaction cost theory. We first define the boundaries of platform enterprises and provide specific measurement methods for their boundaries. By analyzing the distinctions between platform enterprises and manufacturing firms, we adapt the classical transaction cost theory to identify the key determinants of platform enterprise boundaries across three dimensions: data assets and digital technology, network effects, and organizational models. Finally, we offer policy recommendations to foster the healthy development of the platform economy based on our theoretical analysis. Our study highlights the critical role of platform boundary decisions within the framework of crowd science, as they fundamentally shape how diverse smart entities are coordinated on the platform to impact resource allocation efficiency and market stability.
{"title":"On the Determinants of Platform Boundary: A Study from the Perspective of Transaction Cost Theory","authors":"Hang Liu;Xuan Liu;Baowen Sun;Jiayu Wang","doi":"10.26599/IJCS.2025.9100005","DOIUrl":"https://doi.org/10.26599/IJCS.2025.9100005","url":null,"abstract":"In the digital economy era, the rapid expansion of internet platforms has resulted in highly concentrated market structures in online markets, thereby eliciting intensified scrutiny from regulatory authorities. In this paper, we aim to explore the key factors that shape the boundaries of platform firms by extending the transaction cost theory. We first define the boundaries of platform enterprises and provide specific measurement methods for their boundaries. By analyzing the distinctions between platform enterprises and manufacturing firms, we adapt the classical transaction cost theory to identify the key determinants of platform enterprise boundaries across three dimensions: data assets and digital technology, network effects, and organizational models. Finally, we offer policy recommendations to foster the healthy development of the platform economy based on our theoretical analysis. Our study highlights the critical role of platform boundary decisions within the framework of crowd science, as they fundamentally shape how diverse smart entities are coordinated on the platform to impact resource allocation efficiency and market stability.","PeriodicalId":32381,"journal":{"name":"International Journal of Crowd Science","volume":"9 3","pages":"175-180"},"PeriodicalIF":0.0,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11142636","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904822","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-26DOI: 10.26599/IJCS.2025.9100004
Jin Sun;Leiju Qiu
Human-Machine Collaboration (HMC) is a pivotal manifestation of collective intelligence in the digital age, where the synergistic interaction of humans, machines, and physical systems drives socio-economic evolution. This paper redefines HMC through the lens of co-evolution of human and machine capabilities and distributed decision-making, systematically analyzing its development history, collaboration models, and transformative impact on productivity, innovation, and labor markets. This paper believes that human-machine collaboration is the collaborative participation of people and machines in solving problems. It introduces various models of human-machine collaboration from the perspective of automation and autonomy, and discusses the criteria for selecting appropriate models. In terms of economic and social impact, this article first summarizes the existing quantitative measurement methods at the national, regional, and enterprise levels, and discusses the economic impact and impact path of human-machine collaboration from the micro, market, and macro levels of individuals and enterprises. Finally, this paper proposes future research directions, including the improvement of quantitative data of human-machine collaboration, the clarification of the issue of legal responsibility, the formulation of management-level strategies, and indepth research in the fields of medical care, aviation, banking, etc. This paper aims to deepen the understanding of human-machine collaboration and provide reference for future research.
{"title":"Human-Machine Collaboration: Definitions, Models, and Socio-Economic Impacts","authors":"Jin Sun;Leiju Qiu","doi":"10.26599/IJCS.2025.9100004","DOIUrl":"https://doi.org/10.26599/IJCS.2025.9100004","url":null,"abstract":"Human-Machine Collaboration (HMC) is a pivotal manifestation of collective intelligence in the digital age, where the synergistic interaction of humans, machines, and physical systems drives socio-economic evolution. This paper redefines HMC through the lens of co-evolution of human and machine capabilities and distributed decision-making, systematically analyzing its development history, collaboration models, and transformative impact on productivity, innovation, and labor markets. This paper believes that human-machine collaboration is the collaborative participation of people and machines in solving problems. It introduces various models of human-machine collaboration from the perspective of automation and autonomy, and discusses the criteria for selecting appropriate models. In terms of economic and social impact, this article first summarizes the existing quantitative measurement methods at the national, regional, and enterprise levels, and discusses the economic impact and impact path of human-machine collaboration from the micro, market, and macro levels of individuals and enterprises. Finally, this paper proposes future research directions, including the improvement of quantitative data of human-machine collaboration, the clarification of the issue of legal responsibility, the formulation of management-level strategies, and indepth research in the fields of medical care, aviation, banking, etc. This paper aims to deepen the understanding of human-machine collaboration and provide reference for future research.","PeriodicalId":32381,"journal":{"name":"International Journal of Crowd Science","volume":"9 3","pages":"149-163"},"PeriodicalIF":0.0,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11142638","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904825","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-26DOI: 10.26599/IJCS.2024.9100011
Yan Zhao;Xiaoya Kong;Jing Yang;Piotr Felisiak
In the context of dual-circulation development paradigm and high-quality economic growth in full swing, it is crucial to adjust the structure of fiscal revenue and expenditure to drive the crowd intelligence-driven digital economy's development. From the perspectives of fiscal revenue and expenditure structure and market, this study examines the impact of fiscal and taxation policies on the digital economy in China based on the data from 2007 to 2020 (excluding 2021 and 2022 due to COVID-19). The results show that the digital economy's development is positively correlated with several factors, including the proportion of science and technology, financial supervision, energy conservation, environmental protection expenditure, and income tax revenue. Conversely, general public service expenditure, turnover tax, resource tax, and administrative fees have an unfavorable impact on the digital economy. Furthermore, mainly via their impact on the digital economy, general public services, financial regulatory expenditure, and turnover tax revenues indirectly affect the dual-circulation development paradigm. Among the different markets, the consumer market has the most significant impact. Our research provides policy implications for the government in China. In summary, the Chinese government should reduce the scale of general public service expenditure and turnover tax, increase financial supervision, environmental protection and energy conservation, as well as science and technology expenditure. Additionally, regional differences in fiscal revenue and expenditure structure should be considered, and the inter-regional policy intensity should be adjusted based on general macro measures.
{"title":"Intermediary Effect of Digital Economy on Impact of the Fiscal Revenue and Expenditure Structure on Dual-Circulation Development Paradigm","authors":"Yan Zhao;Xiaoya Kong;Jing Yang;Piotr Felisiak","doi":"10.26599/IJCS.2024.9100011","DOIUrl":"https://doi.org/10.26599/IJCS.2024.9100011","url":null,"abstract":"In the context of dual-circulation development paradigm and high-quality economic growth in full swing, it is crucial to adjust the structure of fiscal revenue and expenditure to drive the crowd intelligence-driven digital economy's development. From the perspectives of fiscal revenue and expenditure structure and market, this study examines the impact of fiscal and taxation policies on the digital economy in China based on the data from 2007 to 2020 (excluding 2021 and 2022 due to COVID-19). The results show that the digital economy's development is positively correlated with several factors, including the proportion of science and technology, financial supervision, energy conservation, environmental protection expenditure, and income tax revenue. Conversely, general public service expenditure, turnover tax, resource tax, and administrative fees have an unfavorable impact on the digital economy. Furthermore, mainly via their impact on the digital economy, general public services, financial regulatory expenditure, and turnover tax revenues indirectly affect the dual-circulation development paradigm. Among the different markets, the consumer market has the most significant impact. Our research provides policy implications for the government in China. In summary, the Chinese government should reduce the scale of general public service expenditure and turnover tax, increase financial supervision, environmental protection and energy conservation, as well as science and technology expenditure. Additionally, regional differences in fiscal revenue and expenditure structure should be considered, and the inter-regional policy intensity should be adjusted based on general macro measures.","PeriodicalId":32381,"journal":{"name":"International Journal of Crowd Science","volume":"9 3","pages":"199-209"},"PeriodicalIF":0.0,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11142637","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904672","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-26DOI: 10.26599/IJCS.2024.9100030
Sihan Zhang;Zuoti Shi;Xiaoran Ni
The application of information technology in tax enforcement represents a paradigmatic case of crowd science within government administration, facilitating the adoption of sound governance principles and the achievement of policy objectives through the enhanced deployment of information technologies. This study empirically examines the impact of advanced information technology in tax enforcement, a critical component of government modernization, on the quality of financial reporting. To establish causality, we leverage the staggered implementation of Stage Three of the Golden Tax Project (GTP-3) as a quasi-natural experiment, which integrates modern information technologies into tax enforcement processes. The difference-in-differences estimation results indicate that GTP-3 significantly curtails corporate earnings manipulation. This effect is particularly pronounced in firms taxed by local taxation bureaus, state-owned enterprises, and those with stronger political connections. Moreover, GTP-3 effectively mitigates the crash risk of stock prices for firms subject to its provisions. Overall, our study contributes to the understanding of how crowd science, such as those deployed in tax enforcement, can improve the information quality of the capital market and intersect with broader societal dynamics.
{"title":"Effects of Advanced Information Technology in Tax Enforcement on Financial Reporting Quality: Evidence from a Quasi-Natural Experiment in China","authors":"Sihan Zhang;Zuoti Shi;Xiaoran Ni","doi":"10.26599/IJCS.2024.9100030","DOIUrl":"https://doi.org/10.26599/IJCS.2024.9100030","url":null,"abstract":"The application of information technology in tax enforcement represents a paradigmatic case of crowd science within government administration, facilitating the adoption of sound governance principles and the achievement of policy objectives through the enhanced deployment of information technologies. This study empirically examines the impact of advanced information technology in tax enforcement, a critical component of government modernization, on the quality of financial reporting. To establish causality, we leverage the staggered implementation of Stage Three of the Golden Tax Project (GTP-3) as a quasi-natural experiment, which integrates modern information technologies into tax enforcement processes. The difference-in-differences estimation results indicate that GTP-3 significantly curtails corporate earnings manipulation. This effect is particularly pronounced in firms taxed by local taxation bureaus, state-owned enterprises, and those with stronger political connections. Moreover, GTP-3 effectively mitigates the crash risk of stock prices for firms subject to its provisions. Overall, our study contributes to the understanding of how crowd science, such as those deployed in tax enforcement, can improve the information quality of the capital market and intersect with broader societal dynamics.","PeriodicalId":32381,"journal":{"name":"International Journal of Crowd Science","volume":"9 3","pages":"164-174"},"PeriodicalIF":0.0,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11142649","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904673","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-26DOI: 10.26599/IJCS.2024.9100010
Wenjun Jing;Linlin Wang;Jun Hu
Competition among platform enterprises is a contest of value creation behaviors by multiple groups and is one of the important manifestations of the crowd science in the field of platform economy. However,the frequent unfair competition among platform enterprises has hindered the healthy and rapid development of the platform economy. Clarifying the source of unfair competition is an important prerequisite to regulate this behavior, and the unfair competition behavior between platform enterprises has a complex generation path. In order to clarify this path, this paper explores the source of unfair competition in the platform economy from the perspective of configuration by qualitative comparative analysis. The results of this paper show that unfair competition among platform enterprises is the result of five factors, such as business difference, enterprise scale difference, innovation ability difference, profitability difference, and regulatory environment. These factors combine with each other to form four configuration paths of unfair competition among platform enterprises. Finally, from the perspective of reducing unfair competition in the industry, this paper puts forward specific ideas to standardize the development of platform economy, including respecting the law of platform development, reducing excessive market intervention, and trying to construct the proactive regulatory model with expected nature.
{"title":"Formation Path of Unfair Competition of Platform Enterprises Based on the Perspective of Configuration","authors":"Wenjun Jing;Linlin Wang;Jun Hu","doi":"10.26599/IJCS.2024.9100010","DOIUrl":"https://doi.org/10.26599/IJCS.2024.9100010","url":null,"abstract":"Competition among platform enterprises is a contest of value creation behaviors by multiple groups and is one of the important manifestations of the crowd science in the field of platform economy. However,the frequent unfair competition among platform enterprises has hindered the healthy and rapid development of the platform economy. Clarifying the source of unfair competition is an important prerequisite to regulate this behavior, and the unfair competition behavior between platform enterprises has a complex generation path. In order to clarify this path, this paper explores the source of unfair competition in the platform economy from the perspective of configuration by qualitative comparative analysis. The results of this paper show that unfair competition among platform enterprises is the result of five factors, such as business difference, enterprise scale difference, innovation ability difference, profitability difference, and regulatory environment. These factors combine with each other to form four configuration paths of unfair competition among platform enterprises. Finally, from the perspective of reducing unfair competition in the industry, this paper puts forward specific ideas to standardize the development of platform economy, including respecting the law of platform development, reducing excessive market intervention, and trying to construct the proactive regulatory model with expected nature.","PeriodicalId":32381,"journal":{"name":"International Journal of Crowd Science","volume":"9 3","pages":"181-189"},"PeriodicalIF":0.0,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11142635","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904824","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-26DOI: 10.26599/IJCS.2025.9100003
Gang Liu;Siyuan Zhao;Muhammad Saleem Sumbal
With the vigorous development of the digital economy and the deepening of enterprise digital transformation in China, cultivating digital talent has become a focus of academic research. This study examines the literature related to digital talent from 2015 to 2025 in the China National Knowledge Infrastructure (CNKI) journal database and employs a bibliometric analysis approach to investigate the research focus and trends in digital talent development. The analysis covers authors, institutions, keywords, research hotspots, and trend evolution. The findings reveal the following: First, with the advancement of digital transformation policies by governments and enterprises, the number of publications related to digital talent has significantly increased in recent years. Second, a stable core group of authors has yet to emerge domestically, and collaborations between research institutions are mostly confined to specific regions with limited collaboration. Third, digital talent research mainly focuses on the digital economy and digitalization, enterprise digitalization and digital transformation, and the cultivation of digital talent. Finally, on the basis of the analysis results of the knowledge graph and the national situation, implementation strategies for digital talent cultivation are proposed. These strategies inherently align with crowd science principles, where human-machine-object intelligence interactions drive collective evolution, collaborative innovation, and decentralized decision-making to enhance socio-economic efficacy.
{"title":"Bibliometric Analysis on Digital Talent Cultivation in China","authors":"Gang Liu;Siyuan Zhao;Muhammad Saleem Sumbal","doi":"10.26599/IJCS.2025.9100003","DOIUrl":"https://doi.org/10.26599/IJCS.2025.9100003","url":null,"abstract":"With the vigorous development of the digital economy and the deepening of enterprise digital transformation in China, cultivating digital talent has become a focus of academic research. This study examines the literature related to digital talent from 2015 to 2025 in the China National Knowledge Infrastructure (CNKI) journal database and employs a bibliometric analysis approach to investigate the research focus and trends in digital talent development. The analysis covers authors, institutions, keywords, research hotspots, and trend evolution. The findings reveal the following: First, with the advancement of digital transformation policies by governments and enterprises, the number of publications related to digital talent has significantly increased in recent years. Second, a stable core group of authors has yet to emerge domestically, and collaborations between research institutions are mostly confined to specific regions with limited collaboration. Third, digital talent research mainly focuses on the digital economy and digitalization, enterprise digitalization and digital transformation, and the cultivation of digital talent. Finally, on the basis of the analysis results of the knowledge graph and the national situation, implementation strategies for digital talent cultivation are proposed. These strategies inherently align with crowd science principles, where human-machine-object intelligence interactions drive collective evolution, collaborative innovation, and decentralized decision-making to enhance socio-economic efficacy.","PeriodicalId":32381,"journal":{"name":"International Journal of Crowd Science","volume":"9 3","pages":"190-198"},"PeriodicalIF":0.0,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11142641","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904734","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-26DOI: 10.26599/IJCS.2025.9100006
Hang Liu
The digital economy has become a transformative force, fundamentally reshaping global economic structures, business models, and governance frameworks. At its core lies crowd science and engineering (CSE), which leverages the interconnectedness of diverse smart entities—composed of individuals, enterprises, and governmental agencies—to enhance the stability of the economic system and the efficiency of resource allocation[1]. This synergy fosters smarter, more adaptive systems, enabling innovation and resilience across industries while optimizing processes for sustainable growth.
{"title":"Editorial: Special Issue on Digital Economy and Platform Governance","authors":"Hang Liu","doi":"10.26599/IJCS.2025.9100006","DOIUrl":"https://doi.org/10.26599/IJCS.2025.9100006","url":null,"abstract":"The digital economy has become a transformative force, fundamentally reshaping global economic structures, business models, and governance frameworks. At its core lies crowd science and engineering (CSE), which leverages the interconnectedness of diverse smart entities—composed of individuals, enterprises, and governmental agencies—to enhance the stability of the economic system and the efficiency of resource allocation<sup>[1]</sup>. This synergy fosters smarter, more adaptive systems, enabling innovation and resilience across industries while optimizing processes for sustainable growth.","PeriodicalId":32381,"journal":{"name":"International Journal of Crowd Science","volume":"9 3","pages":"147-148"},"PeriodicalIF":0.0,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11142648","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904691","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-04-01DOI: 10.26599/IJCS.2024.9100037
Ronghua Xu
This paper discusses the latest progress of generative artificial intelligence in the field of smart finance. With the rapid development of financial technology, generative artificial intelligence has become one of the key technologies to promote the innovation of smart finance. By analyzing the specific applications of generative artificial intelligence in a number of smart financial application scenarios, such as intelligent risk control, credit approval, intelligent investment advice, financial product innovation, and intelligent customer service, this paper reveals its significant advantages in terms of improving the efficiency of financial services, optimizing risk management, enhancing the user experience, and promoting the innovation of financial products. At the same time, this paper also points out the challenges and limitations faced in the application of generative artificial intelligence, such as data quality, model interpretability, technology update speed, and security and privacy, and puts forward corresponding solution strategies. Finally, this paper looks forward to the future development trend of generative artificial intelligence in the field of intelligent finance, and believes that it will continue to promote the innovation and development of the financial industry.
{"title":"Recent Advances in Generative Artificial Intelligence for Smart Finance","authors":"Ronghua Xu","doi":"10.26599/IJCS.2024.9100037","DOIUrl":"https://doi.org/10.26599/IJCS.2024.9100037","url":null,"abstract":"This paper discusses the latest progress of generative artificial intelligence in the field of smart finance. With the rapid development of financial technology, generative artificial intelligence has become one of the key technologies to promote the innovation of smart finance. By analyzing the specific applications of generative artificial intelligence in a number of smart financial application scenarios, such as intelligent risk control, credit approval, intelligent investment advice, financial product innovation, and intelligent customer service, this paper reveals its significant advantages in terms of improving the efficiency of financial services, optimizing risk management, enhancing the user experience, and promoting the innovation of financial products. At the same time, this paper also points out the challenges and limitations faced in the application of generative artificial intelligence, such as data quality, model interpretability, technology update speed, and security and privacy, and puts forward corresponding solution strategies. Finally, this paper looks forward to the future development trend of generative artificial intelligence in the field of intelligent finance, and believes that it will continue to promote the innovation and development of the financial industry.","PeriodicalId":32381,"journal":{"name":"International Journal of Crowd Science","volume":"9 4","pages":"213-220"},"PeriodicalIF":0.0,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11293152","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705926","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-04-01DOI: 10.26599/IJCS.2025.9100014
Ziheng Cheng;Yiming Cao
Large Language Models (LLMs) are increasingly employed in knowledge-intensive tasks but often struggle to effectively apply infused knowledge due to textual-structure mismatches between the infusion and reasoning phases. To address this issue, we propose a prompt-based unification strategy that directly learns from factual triples in knowledge graphs while preserving structural consistency across both phases. This unified design enables seamless transfer of factual knowledge to downstream reasoning tasks without requiring architectural modifications. Extensive experiments on two Knowledge Graph Question Answering (KGQA) benchmarks, WebQSP and MetaQA, demonstrate that our approach consistently outperforms strong baselines. Further ablation and robustness analyses verify that structural unification is the key factor driving the improvements, while its compatibility with adapter-tuning and LoRA highlights practical applicability under parameter-efficient fine-tuning settings. Overall, our results suggest that enforcing textual structural consistency provides a simple yet effective principle for reliable knowledge infusion in LLMs, with broad potential across diverse knowledge-intensive domains.
{"title":"Prompt-Based Learning for Factual Knowledge Infusion in Large Language Models","authors":"Ziheng Cheng;Yiming Cao","doi":"10.26599/IJCS.2025.9100014","DOIUrl":"https://doi.org/10.26599/IJCS.2025.9100014","url":null,"abstract":"Large Language Models (LLMs) are increasingly employed in knowledge-intensive tasks but often struggle to effectively apply infused knowledge due to textual-structure mismatches between the infusion and reasoning phases. To address this issue, we propose a prompt-based unification strategy that directly learns from factual triples in knowledge graphs while preserving structural consistency across both phases. This unified design enables seamless transfer of factual knowledge to downstream reasoning tasks without requiring architectural modifications. Extensive experiments on two Knowledge Graph Question Answering (KGQA) benchmarks, WebQSP and MetaQA, demonstrate that our approach consistently outperforms strong baselines. Further ablation and robustness analyses verify that structural unification is the key factor driving the improvements, while its compatibility with adapter-tuning and LoRA highlights practical applicability under parameter-efficient fine-tuning settings. Overall, our results suggest that enforcing textual structural consistency provides a simple yet effective principle for reliable knowledge infusion in LLMs, with broad potential across diverse knowledge-intensive domains.","PeriodicalId":32381,"journal":{"name":"International Journal of Crowd Science","volume":"9 4","pages":"262-268"},"PeriodicalIF":0.0,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11293174","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705871","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-04-01DOI: 10.26599/IJCS.2024.9100038
Zeyu Jia;Shengling Geng;Yibowen Zhao;Huiguo Zhang
Chain-of-thought prompting has attracted much attention in Artificial Intelligence (AI). Large Language Models (LLMs) can be instructed to imitate human thought processes step by step, and they have demonstrated surprising reasoning capabilities. However, when faced with complex reasoning tasks, LLMs perform poorly and often produce inaccurate results. This may be due to insufficient knowledge and poor real-time performance, resulting in incorrect inference chains. Inspired by knowledge augmented deep learning and retrieval augmented generation, a more feasible approach is knowledge guided chain-of-thought prompting generation, which introduces a large amount of knowledge, including common, logical, and factual information, into the process of generating a chain of reasoning. Although a large amount of research has been conducted in these areas, there is still a gap in the survey literature on knowledge-guided chain-of-thought prompt generation. In this survey, we introduce the concept of knowledge-driven chain-of-thought generation and discuss how knowledge plays an important role in the process of chain-of-thought generation and enhancement, both in terms of knowledge sources and knowledge use. Then, evaluation guidelines for chain-of-thought reasoning are sorted out. Next, a benchmark task and a public dataset for chain-of-thought prompting are presented. Finally, we conducted a comprehensive examination of the current opportunities and challenges and formulated a series of recommendations for future research directions. This survey may be of assistance to researchers in the understanding of the latest research developments in these areas.
{"title":"Comprehensive Survey on Prompts Generating via Knowledge-Guided Chain-of-Thought","authors":"Zeyu Jia;Shengling Geng;Yibowen Zhao;Huiguo Zhang","doi":"10.26599/IJCS.2024.9100038","DOIUrl":"https://doi.org/10.26599/IJCS.2024.9100038","url":null,"abstract":"Chain-of-thought prompting has attracted much attention in Artificial Intelligence (AI). Large Language Models (LLMs) can be instructed to imitate human thought processes step by step, and they have demonstrated surprising reasoning capabilities. However, when faced with complex reasoning tasks, LLMs perform poorly and often produce inaccurate results. This may be due to insufficient knowledge and poor real-time performance, resulting in incorrect inference chains. Inspired by knowledge augmented deep learning and retrieval augmented generation, a more feasible approach is knowledge guided chain-of-thought prompting generation, which introduces a large amount of knowledge, including common, logical, and factual information, into the process of generating a chain of reasoning. Although a large amount of research has been conducted in these areas, there is still a gap in the survey literature on knowledge-guided chain-of-thought prompt generation. In this survey, we introduce the concept of knowledge-driven chain-of-thought generation and discuss how knowledge plays an important role in the process of chain-of-thought generation and enhancement, both in terms of knowledge sources and knowledge use. Then, evaluation guidelines for chain-of-thought reasoning are sorted out. Next, a benchmark task and a public dataset for chain-of-thought prompting are presented. Finally, we conducted a comprehensive examination of the current opportunities and challenges and formulated a series of recommendations for future research directions. This survey may be of assistance to researchers in the understanding of the latest research developments in these areas.","PeriodicalId":32381,"journal":{"name":"International Journal of Crowd Science","volume":"9 4","pages":"251-261"},"PeriodicalIF":0.0,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11293179","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705892","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：