最新刊期
卷 12 ， 期 3 ， 2026
-
- 摘要：Cross-domain data circulation constitutes a core link in the construction of a national integrated data market, however, it is confronted with a series of challenges including complex business forms, intertwined technical pathways, and the absence of a unified framework. Furthermore, the prominent characteristics of megacities-such as decentralized data sources, data heterogeneity, and scenario coupling-impose greater challenges on cross-domain data circulation. To address these challenges, five business forms and seven technical pathways of cross-domain data circulation are systematically collated and summarized. On this basis, drawing on the distinct data characteristics of public and private domains, an innovative "federation-confederation" dual-layer architecture is proposed, which enables complex data circulation within the public domain, within the private domain, and between the two domains via a multi-agent collaboration mechanism. Finally, a "scenario matrix" for cross-domain data circulation is developed, with cross-domain federated training of medical data adopted as a case study for empirical analysis.关键词：megacity;Cross-domain Data Circulation;business form;technical path;multi-agent collaboration141|32|0更新时间：2026-05-15
- 摘要：To address the dilemma of balancing cross-domain circulation and security protection for sensitive educational data within an integrated data market, this paper proposes an intelligent governance framework characterized by a "dual-domain" perspective. First, on an inter-project dimension, an innovative dynamic auditing module based on large language model (LLM) and few-shot prompt engineering is designed. By performing intelligent analysis of technical documentation, the module enables the cross-project reuse of governance experience and compliance logic, ensuring that original privacy documents remain within their respective domains. Second, on an intra-project dimension, a progressive "trust, capability, and effectiveness" (TCE) indicator system is constructed to quantitatively evaluate data allocation efficiency and security levels among multiple stakeholders. Empirical results demonstrate that the framework effectively identifies compliance risks throughout the project lifecycle and significantly enhances governance performance. This framework provides a deployable, intelligent regulatory pathway for the cross-domain circulation of sensitive data, achieving the goal of "reusable experience with data remaining on-premise".关键词：national integrated data market;Cross-domain Data Circulation;Data On-premise;large language model;TCE evaluation model177|49|0更新时间：2026-05-15
- 摘要：The construction of a regulatory system is an important part of building the national integrated data market. The lawful and compliant operation of the data market, as well as the assurance of fairness, openness, and impartiality, all depend on the construction of the regulatory system. Based on an analysis of the morphological standards and production specifications for data products, the data market supervision content in the procurement of raw data materials and the production, circulation, and use of data products was proposed. The framework for data market supervision and management was proposed, and the construction path for the regulatory system was designed, comprising six components: regulatory standards, regulatory rules, regulatory approaches, regulatory participants, regulatory implementation, and regulatory disposal. This provides theoretical support and practical guidance for developing the regulatory system for constructing the national integrated data market.关键词：data as a factor of production;data product;integration;data market;digital economy121|47|0更新时间：2026-05-15
- 摘要：To assess the development status and regional disparities of China's integrated data market, this study focuses on the core characteristics of "integration" and constructs an evaluation index system based on the "Five Unifications and One Opening" policy framework. The system encompasses five dimensions: unified market foundational institutions, unified physical infrastructure, unified government regulation and governance, unified factor resource market, and continuous expansion of opening-up to the outside world. Using the analytic hierarchy process (AHP) for weighting, an empirical measurement is conducted on the development level of integrated data markets in 18 representative cities across nine major strategic regions, including Beijing-Tianjin-Hebei, in 2025. The results show that the market is in the initial transition stage from "administrative-driven" to "market-driven". It presents a three-layer structural characteristic of "leading market foundational institutions, insufficient endogenous market momentum, and differentiated layout of opening-up". Spatially, it exhibits a pattern of "high in the east, low in the west; fast in the south, slow in the north" with Shanghai, Beijing and Shenzhen forming the core leading poles, and several major urban agglomerations developing a synergistic "multi-polar heterogeneous and complementary advantages" model. Based on these findings, policy recommendations are proposed to promote the systematic transformation of the data market from aspects such as scenario-based pricing, categorized policy implementation, and institutionalization of rules and technologies.关键词：Data Factors of Production;national integrated data market;Indicator System;Empirical Study;Five Unifications and One Opening125|48|0更新时间：2026-05-15
- 摘要：The construction of an integrated data market in China holds three strategic values: establishing China's dominant position in the global data value chain, removing market barriers to accelerate the reform of data element allocation, and promoting the emergence of new quality productivity through the intelligent transformation of the entire region. To realize the strategic value of China's integrated data market, Jiangsu province has been promoting the construction of a provincial integrated data market from three dimensions: legislation, platforms rules, and infrastructure - in accordance with the unified requirements of four aspects: basic systems, trading objects, circulation rules, and infrastructure, accumulating practical experience for the construction of China's integrated data market. Aligning with the construction goals and strategic values of mature integrated data market, it is proposed that Jiangsu province should focus on three aspects in the future innovating cross-border data flow mechanisms, advancing the infrastructural development of data exchanges, and unifying data property rights systems to accelerate the establishment of the provincial integrated data market.关键词：data market;integration;data exchange;Jiangsu province;data property right117|39|0更新时间：2026-05-15
-
Regional dilemmas and mechanism innovation in market-oriented allocation reform of data elements AI导读摘要：As an emerging production factor in the digital economy era, data elements face numerous constraints caused by underdeveloped basic institutional frameworks during the advancement of market-oriented allocation reform. Taking the Yangtze River Delta Region and the Guangdong-Hong Kong-Macao Greater Bay Area—the two pioneering pilot zones for data element marketization as the research subjects, this paper systematically compared the development status and policy landscape of data element market-oriented allocation in these two regions. Through comparative study, it identifies the common institutional challenges and distinctive features of their respective basic data element institutions. It is found that leveraging their comparative advantages, both regions have created a series of practical mechanisms to address institutional bottlenecks and unlock the value of data elements, two typical paradigms for regional synergy and cross-border flow in the reform of the market-oriented allocation of data factors have been formed respectively. Furthermore, this paper puts forward a national applicable development approach for the reform of the market-oriented allocation of data elements, which adapts the unified national core basic systems with regionally flexible implementation rules, specific mechanism recommendations are also proposed for each basic system.关键词：data element;market-oriented allocation;the Yangtze River Delta Region;the Guangdong-Hong Kong-Macao Greater Bay Area;comparative study82|23|0更新时间：2026-05-15 - 摘要：The development of a unified national data market has been entrusted with the critical task of supporting and connecting the entire process of data value realization. Drawing on the framework of data value transition from resourceization to productization, assetization, and capitalization, and incorporating the data value network perspective, the key constraints on data value realization were systematically examined from the dimensions of value-creating actors and value activity mechanisms. The results showed that the current value transition of data was primarily constrained by an incomplete data industry chain, underdeveloped multi-modal data circulation forms, and insufficient coordination mechanisms between data and other production factors. To address these challenges, a practical pathway centered on building a full-process data industry chain, improving diversified data circulation mechanisms, and strengthening cross-factor market coordination were proposed. The study provides theoretical support and practical references for accelerating the development of a unified national data market and promoting the continuous value transition of data as a production factor.关键词：integrated data market;data value;transition;logical reasoning;practical approach93|39|0更新时间：2026-05-15
TOPIC
- 摘要：Recent studies have revealed significant “source bias” in retrieval and large language models (LLM), where these models exhibit a preference for model-generated content over authentic human-written data. Humans similarly demonstrate “self-consistency” phenomena, characterized by a tendency to maintain coherent self-concepts and favor existing knowledge over external facts. It was investigated whether LLMs exhibit human-like cognitive biases. It was conducted from two perspectives: an explicit approach (content directly generated by the LLM) and an implicit approach (using prompts to induce the models to perceive the content as self-generated). Confidence was evaluated using both self-assessment and consistency methods. Systematic experiments were carried out on several models, including GPT-4o, DeepSeek-R1, Llama2 (7B70B), and Qwen2 (7B72B). The results demonstrated that under implicit conditions,LLMs exhibit pronounced self-consistency phenomena, displaying significantly higher confidence in self-attributed content. Leveraging this characteristic yields consistent performance gains: the proposed role-based prompting strategy improved accuracy on TriviaQA, NQ, HotpotQA, FEVER, and ZsRE, and maintained a clear advantage under high levels of noisy retrieved documents, indicating strong robustness and generalization. Finally, our mechanistic analysis revealed that instruction fine-tuning and reinforcement learning from human feedback were the primary factors contributing to such biases in models, explaining how self-consistency emerges during training.关键词：large language model;retrieval-augmented generation;knowledge augmentation;source bias79|30|0更新时间：2026-05-15
- 摘要：The link prediction (LP) for social network is an important research direction of complex network mining, which tries to detect the potential relationship between network nodes. The popular LP algorithms for social networks include the static similarity-based indicators, dynamic learning-based predictors and content-based methods. Although the experimental results have reported the good performances in specific application scenarios of these algorithms, some essential defects such as prediction unreliability and detection incapability still exist and cannot be effectively solved up to now. Therefore, a reliability enhancement-based link prediction (RE-LP) algorithm is proposed to make up the above-mentioned shortcomings of existing LP algorithms. There are three main components in RE-LP algorithm, i.e., non-existing link partition, reliable predictor construction, and connection probability calculation. In RE-LP algorithm, the observable non-existing links are sophisticatedly partitioned into highly-reliable non-existing (HRNE) links and possibly-observed existing (POE) links. It then continuously identifies HRNE links from the POE links in an iterative manner, and uses a well-performed Bayesian classifier-based link predictor to calculate the connection probabilities of the POE links, in order to achieve the goal of accurately and reliably predicting unknown links. Through a series of verification experiments, the advantages of RE-LP algorithm in terms of feasibility, rationality and effectiveness are fully demonstrated. The experimental results demonstrate that RE-LP algorithm can obtain the 21.7%~36.1% higher prediction AUC than 5 advanced LP algorithms and meanwhile is able to detect the potential links with high credibility.关键词：social network;link prediction;Bayesian classifier;reliability enhancement;complex network analytics94|51|0更新时间：2026-05-15
- 摘要：The rise of large language model (LLM) has brought new opportunities to recommender systems. However, existing research mainly focuses on the technical frameworks and engineering implementations of LLM-based recommender systems, lacking a systematic review of this interdisciplinary field. In particular, the key research questions that need to be addressed when integrating LLM into recommender systems remain unclear. To this end, the study summarizes the mainstream application patterns of LLM in recommender systems and categorizes the key research issues across various stages of the recommender system lifecycle. Simultaneously, the study investigates how LLM provide innovative solutions to these issues, identifies unresolved problems, and outlines future research directions.关键词：large language model;recommender system;application paradigm;technical pathway290|67|0更新时间：2026-05-15
- 摘要：With the rapid development of artificial intelligence, deepfake technology has become a powerful tool for generating realistic audio, images, and videos. However, its widespread use and decreasing costs pose serious threats to personal privacy and social trust. This paper reviews the generalization issues in the field of video deepfake detection, providing a comprehensive overview of methods, challenges, and advancements. It first traces the development of deepfake detection techniques and compares the performance of neural networks with human capabilities in detection tasks. Regarding the challenges of cross-dataset detection, the paper analyzes how differences in feature distributions between datasets lead to the performance degradation of traditional deep learning models in cross-domain applications. To address this issue, researchers have proposed various solutions, such as neural network architecture design, multi-task learning, new loss functions, data augmentation, innovative training procedures, and the integration of biometric recognition technologies. In recent years, as deepfake detection research has reached a stage of stable development, it has also encountered bottlenecks, calling for breakthroughs. Through a systematic analysis of the latest research achievements in this field, this paper examines the strengths and limitations of current techniques, revealing both the potential and challenges in enhancing generalization and addressing unknown distribution domains. The insights provided offer valuable guidance for future research and applications.关键词：video deepfake detection;face forgery detection;generalization capability103|103|0更新时间：2026-05-15
- 摘要：With the development of large language model technology, the application of retrieval enhancement in the field of education has become one of the hot research directions, with the aim of alleviating the hallucination problem of large language models and improving the accuracy of large language models in answering educational questions. Questions in the field of education are usually more complex and highly personalized. When traditional retrieval methods are applied to educational questions and answers, they often have problems such as inaccurate semantic matching, insufficient context understanding, and difficulty in data processing, resulting in poor answer quality. To address the above challenges, this paper proposes a retrieval enhancement technology based on a neural topic model, which can effectively improve the accuracy of large language models in answering Python programming education questions. This technology reorders the retrieved external knowledge so that information that is more relevant to the question in the educational scenario is used to prompt the large language model to answer the question. Experimental results show that the Python question-answering model built based on the proposed topic enhancement technology generates higher-quality answers than the comparison models.关键词：retrieval-augmented generation;programming education;application;large language model115|349|0更新时间：2026-05-15
- 摘要：Data asset pricing is jointly affected by three factors: the transaction subject, the transaction object and the transaction mechanism. This study comprehensively considers the characteristics of the data transaction subject and object, constructs the entity credit factor (ECF) and data quality factor (DQF) to make a comprehensive evaluation. On this basis, the study designs a comprehensive price sealed-bid auction data asset pricing model based on the (
STUDY
- 摘要：This article desiqns a digital twin pipeline vibration prediction system that integrates physical mechanisms and data-driven approaches to address the problem of limited sensors being unable to achieve real-time monitoring of full-field vibration in complex pipeline systems due to space constraints and structural complexity. A hybrid modeling framework for engineering applications was constructed with the core objective of reconstructing the full field vibration response. The efficient dimensionality reduction of the full field vibration response was achieved by introducing intrinsic orthogonal decomposition. Nonlinear mapping relationships between finite measurement point data and the global vibration modal coefficients were established using a Kriging model and radial basis function surrogate model, respectively. On this basis, the reconstruction system integrates WebGL 3D visualization technology to achieve real-time dynamic display of the global vibration field of the pipeline. The experimental results show that the system can accurately predict the overall vibration distribution of the pipeline under the condition of transmission delay less than 10 ms, verifying the advantages of the proposed method in prediction accuracy, real-time performance, and engineering applicability.关键词：digital twin;vibration reconstruction;pipeline system;sensor placement optimization;surrogate model;proper orthogonal decomposition47|41|0更新时间：2026-05-15
APPLICATION
- Postal code：100079
- Tel：（010）53879208、53859533 Email：bdr@bjxintong.com.cn
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09064830号-19
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
- Total Visits：298834 Visits Today：357
0