Top Read Articles

    Published in last 1 year |  In last 2 years |  In last 3 years |  All
    Please wait a minute...
    For Selected: Toggle Thumbnails
    Analysis of the status, hotspots, and trends of open-source innovation: A bibliometric study based on CNKI literature from 2005 to 2024
    Rui WANG, Qiuyue LYU, Jia LIAO
    J* E* C* N* U* N* S*    2025, 2025 (5): 125-139.   DOI: 10.3969/j.issn.1000-5641.2025.05.012
    Abstract1201)   HTML5)    PDF(pc) (1646KB)(440)       Save

    This study systematically analyzes the evolutionary characteristics and research hotspots of open- source innovation in China. A dataset comprising 732 valid journal articles, with “open-source” in the title, was retrieved from the China National Knowledge Infrastructure (CNKI) for the period 2005–2024. A bibliometric approach was employed to examine such dimensions as annual publication volume, disciplinary distribution, keyword co-occurrence and clustering, burst keywords, and timeline evolution. The results indicate that research in this field has progressed through three stages, initial exploration, steady development, and rapid growth, with a significant surge in publications over the past five years. Disciplinary distribution analysis reveals a multidisciplinary landscape centered on library and information science, computer science, and industrial technology, which extends to fields such as education, management, and law. Keyword clustering analysis identifies nine core research areas, accompanied by a review of the representative literature within each cluster. Timeline evolution analysis suggests that future research will likely focus on the deep integration of artificial intelligence (AI) and open-source ecosystems, the evolution of collaboration and governance models in open-source communities, open-source software security and supply chain risk identification, and open-source law and intellectual property protection. On the basis of these findings, we propose several recommendations to foster the sustainable development of open-source innovation in China, including strengthening the synergistic governance of AI and open-source ecosystems, enhancing supply chain security systems, advancing innovations in legal and licensing frameworks, and constructing a digital open-source infrastructure oriented toward industrial and public services.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Research on the GitHub developer geographic location prediction method based on multi-dimensional feature fusion
    Sijia ZHAO, Fanyu HAN, Wei WANG
    J* E* C* N* U* N* S*    2025, 2025 (5): 1-13.   DOI: 10.3969/j.issn.1000-5641.2025.05.001
    Abstract1069)   HTML25)    PDF(pc) (1118KB)(107)       Save

    The geographic location information of developers is important for understanding the global distribution of open source activities and formulating regional policies. However, a substantial number of developer accounts on the GitHub platform lack geographic location information, limiting the comprehensive analysis of the geographic distribution of the global open source ecosystem. This study proposed a hierarchical geographic location prediction framework based on multidimensional feature fusion. By integrating three major categories of multidimensional features—temporal behavior, linguistic culture, and network characteristics—the framework established a four-tier progressive prediction mechanism consisting of rule-driven rapid positioning, name cultural inference, time zone cross-validation, and a deep learning ensemble. Experiments conducted on a large-scale dataset built from 50000 globally active developers demonstrated that this method successfully predicted the geographic locations of 82.52% of the developers. Among these, the name cultural inference layer covered most users with an accuracy of 0.7629, whereas the deep learning ensemble layer handled the most complex cases with an accuracy of 0.7557. A comparative analysis with the prediction results from the Moonshot large language model validated the superiority of the proposed method in complex geographic inference tasks.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Interactive data structure and algorithm visualization based on AI agents
    Ruiyang PANG, Xuesong LU
    J* E* C* N* U* N* S*    2025, 2025 (5): 32-42.   DOI: 10.3969/j.issn.1000-5641.2025.05.004
    Abstract824)   HTML16)    PDF(pc) (1065KB)(490)       Save

    Data structures and algorithms (DSA), as a core course in computer science education, play a key role in cultivating programming skills and algorithmic thinking of students. Visualization can significantly enhance teaching effectiveness and deepen student understanding in DSA education. However, existing DSA visualization tools often rely on manually written visualization codes that lead to limitations such as limited coverage, high maintenance costs, and lack of interactivity; hence, the needs of dynamic demonstrations and personalized teaching are difficult to meet. With the outstanding performance of large language models (LLMs) in code generation, automated DSA visualization has become a promising possibility. Therefore, this study proposed an interactive visualization code generation method based on the reasoning and acting (ReAct) AI agent framework, aiming to address the low automation and insufficient interactivity of traditional visualization tools. By leveraging the code generation capabilities of LLMs and integrating with the data structure visualization (DSV) platform interface, the proposed method transformed Python-based DSA code into interactive, executable, and dynamically visualized code, thereby enhancing teaching clarity and learning experience. To systematically evaluate the effectiveness of the method, we constructed a dataset of 150 pairs of DSA code and corresponding DSV visualization code and compared three approaches—direct prompting, chain-of-thought prompting, and the ReAct AI agent approach—across several mainstream LLMs. The experimental results showed that the proposed ReAct AI agent-based method significantly outperformed the other approaches in terms of the compilation rate, execution rate, and usability rate, with the best performance observed in the DeepSeek-R1 model. This demonstrated notable improvements in the accuracy and interactivity of generated visualization code. This research confirms the feasibility and advantages of integrating LLMs with agent frameworks in DSA visualization teaching, offering a novel path toward building efficient, personalized, and automated tools for computer programming education.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Synergy between large language models and open source ecosystems in AI education
    Lijun XU, Li YANG, Ziyi HUANG
    J* E* C* N* U* N* S*    2025, 2025 (5): 66-75.   DOI: 10.3969/j.issn.1000-5641.2025.05.007
    Abstract786)   HTML11)    PDF(pc) (1061KB)(105)       Save

    To address the challenges of outdated teaching resources, insufficient practical skills, and a lack of value-oriented guidance in education, this study constructs an innovative pedagogical model driven by the dual-engine of large language model (LLM) and open source ecosystem. The model is designed to bridge the gap between theoretical knowledge and real-world engineering practice by integrating open-source tools, dynamic code repositories, and authentic project scenarios into the curriculum. Meanwhile, LLMs are employed as intelligent teaching assistants to enable personalized learning paths, generate automated feedback, and support immersive ideological and ethical modules. This research was implemented in the course “Artificial intelligence and its applications”, where a mixed-method evaluation was conducted. Quantitative metrics such as attendance, interaction frequency, repository contributions, and assignment performance were used to measure student engagement and learning effectiveness. Additionally, a set of custom-designed assessment formulas was used to evaluate cross-platform transferability and community participation. Experimental results from 90 undergraduate students showed that learners engaged in open-source collaboration and LLM-assisted learning achieved significantly higher scores in both technical proficiency and value cognition than those in the control group. The study demonstrates that the integration of LLMs and open-source collaboration can effectively enhance student autonomy, promote engineering skills, and reinforce ethical awareness. This dual-driven model not only offers a feasible approach for modernizing AI education but also contributes to the broader goal of cultivating socially responsible and technically competent AI talents.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Application and evaluation of large language models in open source project topic annotation
    Dexin HE, Fanyu HAN, Wei WANG
    J* E* C* N* U* N* S*    2025, 2025 (5): 14-24.   DOI: 10.3969/j.issn.1000-5641.2025.05.002
    Abstract763)   HTML13)    PDF(pc) (795KB)(82)       Save

    With the rapid development of open source communities, the number of GitHub projects has increased exponentially. However, a considerable portion of these projects lack explicit topic labels, creating challenges for developers in technology selection and project retrieval processes. Existing topic generation methods rely primarily on supervised learning paradigms that suffer from strong dependencies on high-quality annotated data and other limitations. This study addresses the accuracy and efficiency issues in open source community project topic annotation by conducting the first comprehensive study on the application effectiveness of large language models in GitHub project topic prediction tasks. We constructed a dataset containing 3000 popular GitHub projects that were selected based on a quantitative metric specifically designed to evaluate the activity and influence of open source projects, encompassing multidimensional features including repository names, README documents, and description information. Comparative experiments were conducted using several mainstream large language models from domestic and international sources including Claude 3.7 Sonnet, DeepSeek-V3, Gemini 2.0 Flash, GPT-4o, and Qwen-Plus. The results demonstrated that Claude 3.7 Sonnet achieved optimal performance across most evaluation metrics, and as the dataset scale expanded, the performances of all models tended to stabilize. The experiments proved that large language models exhibited excellent applicability in project topic annotation tasks, although significant performance differences existed among different models. These findings provide an important reference foundation for open source community project management and intelligent annotation system design.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Open source evaluatology: A framework and methodology for evaluating open source ecosystems based on evaluatology
    Shengyu ZHAO, Wei WANG, Fanyu HAN, Jiaheng PENG, Lan YOU
    J* E* C* N* U* N* S*    2025, 2025 (5): 151-161.   DOI: 10.3969/j.issn.1000-5641.2025.05.014
    Abstract746)   HTML8)    PDF(pc) (670KB)(488)       Save

    The open source ecosystem, as a critical component of the modern software industry, has garnered increasing attention from both academia and industry regarding its evaluation challenges. However, existing evaluation methods face issues such as inconsistent evaluation standards, lack of theoretical grounding, and poor comparability of evaluation results. Guided by foundational theories of evaluatology, this study introduced a novel interdisciplinary research domain, open source evaluatology, for the first time. It established a theoretical framework and methodological system for evaluating the open source ecosystem. The primary contributions of this paper include the following. Developing the theoretical foundation of open source evaluatology based on the five axioms of evaluatology and defining fundamental concepts, evaluation dimensions, and standards for open source ecosystem evaluation. Designing an evaluation conditions framework comprising five levels: problem definition, task instances, algorithm mechanisms, implementation examples, and supporting systems. A hybrid evaluation model combining statistical and network metrics was proposed. Based on the experiments conducted using the GitHub dataset, this study validated the proposed method from three dimensions: open source repositories, developers, and communities. The results demonstrated that the proposed evaluation model exhibited strong applicability and explanatory power in open source scenarios.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Research on challenges and optimization of large multimodal model applications in treefall scenarios
    Lei FENG, Chaonan LI, Chunjie SHENG, Yuxing SHI, Yicheng HUANG, Jianhong JIN, Yun XU, Yuzhou DU, Nina ZHOU, Sihao MIAO
    J* E* C* N* U* N* S*    2025, 2025 (5): 53-65.   DOI: 10.3969/j.issn.1000-5641.2025.05.006
    Abstract675)   HTML12)    PDF(pc) (1399KB)(680)       Save

    To address the limited robustness of large multimodal models (LMMs) in complex visual scenarios, such as identifying responsibility for fallen trees, which emanates from their reliance on single-path reasoning. This study proposes a novel reasoning optimization method based on Beam Search Chain-of-Thought (BS-CoT). Conventional models often fall into a “first-impression” trap, in which an initial incorrect inference leads to an irreversible analytical failure. The proposed BS-CoT method counteracts this by exploring and evaluating multiple potential inference paths in parallel. It maintains a diverse set of hypotheses about the scene, continuously pruning less likely hypotheses, which effectively overcomes the tendency to commit to a single, fallacious line of reasoning. This significantly enhances visual decision-making capabilities in complex and noisy environments. To validate its efficacy, we constructed a specialized dataset capturing a wide array of treefall incidents in urban governance. Experimental results demonstrated that the proposed method achieved substantial improvements in both event recall and key information capture rates compared with baseline models. This research not only provides a reliable technical solution for visual decision-making challenges in urban public safety but also introduces a new, more robust paradigm for improving the reasoning reliability of large models in critical applications.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    OSS Insight: A platform for open source ecosystem spatiotemporal data analysis and insights
    Xiaowei CHEN, Wei WANG, Fanyu HAN, Guanglei BAO, Fei DONG, Hao HUO, Chen LIU
    J* E* C* N* U* N* S*    2025, 2025 (5): 170-182.   DOI: 10.3969/j.issn.1000-5641.2025.05.016
    Abstract641)   HTML4)    PDF(pc) (1447KB)(605)       Save

    An open source ecosystem abounds with valuable data, yet extracting insights requires innovative data infrastructure and analytical methods. To address this, OSS Insight was developed that innovatively used the hybrid transactional analytical processing(HTAP) database for efficient storage and query of billions of GitHub event data and offered real-time exploration via a visual interface. It delved into spatiotemporal data analysis, modeling developer behaviors and ecosystem evolution, such as visualizing global contribution patterns. Integrated with large language models(LLMs), it enabled natural language to structured query language(SQL) conversion for intelligent querying. A case study of Kubernetes showcased its capabilities in analyzing developers, project evolution, and organizational collaboration. Experiments proved that OSS Insight efficiently analyzed large-scale open source data, and its LLM-driven interaction simplified data analysis and provided automated insights.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    A DTA based activity evaluation method for high star GitHub repositories
    Mingdong YOU, Jiaheng PENG, Fanyu HAN, Wei WANG
    J* E* C* N* U* N* S*    2025, 2025 (5): 140-150.   DOI: 10.3969/j.issn.1000-5641.2025.05.013
    Abstract640)   HTML5)    PDF(pc) (1378KB)(909)       Save

    In the context of identifying GitHub’s long-term active, high-star repositories—critical for assisting the development of robust open-source communities and vital digital infrastructure—we propose a novel method for evaluating the long-term activity of these repositories. This method is firmly based on a time series prediction model, which excels in forecasting repository activity metrics rather than being specifically designed for this purpose. A key innovation of our method is the first-time use of the developer activity cycle as a pivotal feature. This improves the accuracy of predictions for repository development trends and provides a more nuanced understanding of project evolution. After meticulously modeling and mining the time series data of various activity indicators, we developed a new activity calculation formula: development trend-based activity (DTA). This formula allows a precise quantitative evaluation of a repository's true activity level. To rigorously validate our methodology, we designed and curated a comprehensive benchmark dataset with fine time granularity and broad coverage. Subsequently, we systematically evaluated the performance of multiple prediction models against this dataset, eventually identifying the best model for forecasting open-source repository activity. The experimental results conclusively demonstrate the effectiveness of our proposed method in accurately predicting the long-term activity of repositories. Consequently, using DTA to evaluate repository activity can enable open-source participants to effectively identify repositories poised for long-term engagement, strategically determine their participation focus, and thereby significantly promote the sustained development of open-source communities and critical digital infrastructure.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Dynamics model in two-layer networks and case study on generative artificial intelligence bias cognition propagation
    Hongmiao ZHU, Xiaodong ZHAO, Huimin ZHOU, Jiayin QI
    J* E* C* N* U* N* S*    2025, 2025 (5): 191-201.   DOI: 10.3969/j.issn.1000-5641.2025.05.018
    Abstract637)   HTML11)    PDF(pc) (1263KB)(1783)       Save

    This study developed a model to understand the communication dynamics of generative artificial intelligence (GAI) bias cognition within a two-layer network comprising enterprise managers and ordinary employees. The model integrated the effects of communication between different levels and the impact of cognitive training. Using the next-generation matrix method, the study accurately calculated the propagation threshold, R0, that served as a crucial quantitative foundation for effective governance. Specifically, when R0<1, deviant cognition tended to disappear spontaneously, whereas when R0>1, a risk of biased cognition spreading existed. Additionally, the study compared and evaluated two intervention strategies through numerical simulations, providing a comprehensive analysis of the mechanisms that drove the generation and dissemination of deviant cognition within enterprises, supported by case studies.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Research on video question answer for the development of theory of mind
    Yuanyuan MAO, Xin LIN, Qin NI, Ciping DENG, Yiming MA
    J* E* C* N* U* N* S*    2025, 2025 (6): 46-52.   DOI: 10.3969/j.issn.1000-5641.2025.06.006
    Abstract637)   HTML5)    PDF(pc) (664KB)(146)       Save

    In recent years, with the continuous development of machine theory of mind (ToM), research has found that the development of machine ToM differs significantly from the triangular model of children’s ToM development. Consequently, we propose a machine-oriented theory of mind triangular model. This model elucidates the relationships among various tools in the process of developing machine ToM. Additionally, we introduce an evaluation dataset suitable for the dynamic assessment of machine ToM. Finally, this paper designs a VideoQA(video question answer) model, named FOMemNet (fact and observer memory network), specifically tailored for cognitive reasoning—a model addressing belief, desire, and intention reasoning. Considering that models in cognitive reasoning tasks need to infer from the observer’s perspective, we incorporate the FOEM (vision fact and observer perception encoder module) module in FOMemNet for the fusion of multimodal features, thereby obtaining visual factual features and observer features. Subsequently, the model utilizes the FOF (fact and observer fusion) module and two memory modules to integrate features from both perspectives for obtaining a global representation. FOMemNet results in a 2.27% improvement of BDIQA. Our experiments demonstrate the effectiveness of the concept of fact and observer perception in enhancing cognitive reasoning abilities in VideoQA.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Research progress on the interaction mechanisms and combined toxicity of microplastics and antibiotics in aquatic environments
    Yongle ZHAO, Yan ZHANG
    J* E* C* N* U* N* S*    2026, 2026 (1): 25-42.   DOI: 10.3969/j.issn.1000-5641.2026.01.003
    Abstract635)   HTML16)    PDF(pc) (1347KB)(1430)       Save

    Microplastics and antibiotics are emerging contaminants that have become a major focus of current research. Data from Chinese and English databases were compiled to assess their distribution, adsorption mechanisms, and biological hazards in aquatic environments across China. Microplastic levels were highest in reservoirs (4.70~27.5 items/L, average 12.08 items/L), followed by rivers (average 8.83 items/L), and lowest in lakes (average 6.19 items/L). Average antibiotic concentrations were greatest in rivers (102.93 ng/L), exceeding those in lakes (34.37 ng/L) and reservoirs (43.91 ng/L). Van der Waals forces consistently occur between microplastics and antibiotics, while higher polarity MPs readily form hydrogen bonds. Functional groups facilitated π-π interactions and aging microplastics expose more oxygen-containing functional groups that significantly enhance adsorption. Once ingested, microplastic-antibiotic complexes can accumulate in organs, inhibit growth, alter biological structures under prolonged exposure, and promote the spread of antibiotic resistance genes. Effective mitigation requires strengthening control of emission sources, improved recycling systems, adoption of biodegradable plastics, and strengthening research on adsorption processes and combined toxicity under real-world conditions.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Carbon emissions accounting and carbon emissions reduction benefits of sponge city construction based on life cycle assessment
    Xingyan BAO, Ruihui CHENG, Sheng XIE, Zhaokang WU, Haiyan KUAI, Boxiao ZHANG, Bowen LYU, Kai YANG
    J* E* C* N* U* N* S*    2025, 2025 (6): 94-105.   DOI: 10.3969/j.issn.1000-5641.2025.06.011
    Abstract624)   HTML7)    PDF(pc) (1503KB)(560)       Save

    In response to the new requirements for systematic sponge city development under China’s “carbon peaking and carbon neutrality” strategy, scientifically evaluating the carbon emissions and carbon emission reduction benefits of sponge city construction holds significant theoretical and practical value. This study took Wuhu, a national sponge demonstration city, as an example. Taking on the perspective of the whole life cycle, and combining the emission factor method and the Technical Guidelines for Carbon Emission Accounting for Sponge City Construction in Anhui Province, this study utilized the carbon emission accounting method for sponge cities that is applicable to engineering in practice, and evaluated the carbon emission and emissions reduction results of four types of typical sponge projects in 2022. Based on this, the study took a residential community as a representative example and employed the NSGA-Ⅱ algorithm to explore strategies for achieving synergistic carbon emissions reduction by optimizing the configuration of multiple types of sponge facilities. The results indicate the following. (1) The carbon emissions from sponge city projects in Wuhu are primarily concentrated in the construction phase, with total emissions of 11438.6 t. Among these, material production and transportation account for 53% and 36%, respectively, indicating considerable potential for emissions reduction. (2) During the operational phase, sponge cities largely rely on sustained natural processes to function, with the carbon reduction effects being relatively concentrated in this stage. The total carbon emissions during this phase are approximately –242.3 t. Assuming that existing operational conditions are maintained without any new facilities, sponge facilities are expected to achieve a cumulative carbon reduction of 7269.7 t in 30 years. Although the annual carbon reduction is relatively limited, long-term operation can gradually offset carbon emissions from the construction phase, demonstrating strong carbon neutrality potential. (3) In terms of specific facility types, green stormwater infrastructure such as grassed swales (4.95 kg/m2) and sunken green spaces (11.35 kg/m2) exhibit relatively low carbon emissions intensities during the construction phase. The carbon reduction benefits of sponge facilities in the operation stage are significantly influenced by their functional characteristics and the scale of implementation. (4) Taking a residential community as an example, and based on the annual total runoff control rate requirement, the coordinated carbon reduction capacity of sponge facilities can be enhanced by reasonably adjusting the scales of sunken green spaces, permeable pavements, rain gardens, and grassed swales. This study provides a quantitative evaluation of multi-facility sponge city systems from a holistic perspective, offering methodological support and a theoretical reference for the development of low-carbon urban drainage systems.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Effects of soil conditioners on soil properties and vegetable growth in an open-pit coal mine dump
    Xiaoliang JIAO, Xinxin YU, Heyi GONG, Kun SONG, Liangjun DA
    J* E* C* N* U* N* S*    2025, 2025 (6): 106-115.   DOI: 10.3969/j.issn.1000-5641.2025.06.012
    Abstract610)   HTML6)    PDF(pc) (760KB)(546)       Save

    This study addresses the challenge of water and nutrient deficiencies in the planted soil of spoil dumps in open-pit coal mines by investigating the effectiveness of soil amendments containing a water-retaining agent, organic fertilizer, and microbial inoculant in enhancing soil properties and promoting plant growth. A pot experiment was conducted using Brassica rapa var. chinensis as the model plant to evaluate the effects of different amendment dosages. Key soil parameters were analyzed, including pH, electrical conductivity, alkali-hydrolyzable nitrogen, available phosphorus, available potassium, and organic matter content. Plant growth indicators such as plant height, leaf number, and dry biomass were measured. The findings indicated that applying the water-retaining agent and organic fertilizer significantly improved soil properties and promoted plant growth, whereas the microbial inoculant did not produce statistically significant effects. The optimal combined effect of the amendments on soil properties and plant growth was achieved at a water-retaining agent dosage of 5‰ and an organic fertilizer dosage of 30%.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Research and analysis of the development of the open source ecosystem in the field of geographic information system
    Yuang ZHANG, Zhong XIE, Qinjun QIU, Liufeng TAO
    J* E* C* N* U* N* S*    2025, 2025 (5): 99-108.   DOI: 10.3969/j.issn.1000-5641.2025.05.010
    Abstract605)   HTML28)    PDF(pc) (1330KB)(394)       Save

    With the rapid advancement of information technology, the open source paradigm has become popular in multiple domains, including geographic information system (GIS). Developing an open, collaborative, and sustainable open source GIS ecosystem can promote GIS technology innovation, lower implementation costs, and foster development within the field. This study systematically investigates methods for developing an open source GIS ecosystem and its future trends, addressing four main aspects: ① reviewing the development history of open source GIS and the current technological landscape to refine a four-stage evolutionary framework; ② from a GIS perspective and based on the existing open source foundation, proposing a multi-layered ecosystem construction model specifically tailored to GIS; ③ introducing HyperCRX to perform quantitative analysis and visualization of four metrics—OpenRank, Activity, Contributors, and Participants—for eight representative open source GIS projects, thereby revealing differences in their influence, activity levels, and community engagement to reflect the current state of the ecosystem; and ④ summarizing the challenges faced by the open source GIS ecosystem in terms of public perception, talent cultivation, governance mechanisms, data–software coordination, and sustainable business models, as well as outlining future development directions and research hotspots in the era of large-scale models. It is hoped that this study will provide useful references for future research and practical applications.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    ATBench: Benchmark for evaluating analysis trajectories in end-to-end data analysis
    Xufei WANG, Huarong XU, Panfeng CHEN, Mei CHEN, Dan MA, Zhengxi CHEN, Xu TIAN, Hui LI
    J* E* C* N* U* N* S*    2025, 2025 (5): 43-52.   DOI: 10.3969/j.issn.1000-5641.2025.05.005
    Abstract599)   HTML8)    PDF(pc) (809KB)(130)       Save

    This paper introduces ATBench, a benchmark designed for evaluating analysis trajectories in end-to-end data analysis tasks, to address the limitations in granularity and domain coverage present in current benchmarks. Analysis trajectories represent the process in which an agent iteratively poses questions, derives insights, and formulates conclusions around a specific analysis goal via iterative interactions. Leveraging both existing benchmarks and real Kaggle task data, we constructed 151 evaluation datasets spanning eight distinct domains by employing an annotation strategy that balances goal-driven and exploratory approaches. Additionally, we propose a fine-grained evaluation metric, the analysis trajectory score, to assess an agent's coherent analytical capabilities during end-to-end data analysis tasks. Experimental results demonstrate that ATBench exhibits strong stability and discriminative power, effectively distinguishing performance differences among models in analytical tasks. The results also reveal the limitations in agents’ abilities for coherent analysis and insight discovery, thereby providing data-driven support for future improvements.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    A study on the control effectiveness of Spartina alterniflora based on combined cutting and herbicide application measures
    Zongyue LIU, Yunyi CHI, Qiang WANG
    J* E* C* N* U* N* S*    2025, 2025 (6): 116-127.   DOI: 10.3969/j.issn.1000-5641.2025.06.013
    Abstract575)   HTML4)    PDF(pc) (2219KB)(552)       Save

    Spartina alterniflora is a globally invasive plant that has caused significant damage to coastal ecosystems, including biodiversity loss, decline of native vegetation, alteration of benthic communities, and degradation of bird habitats. Effective control of S. alterniflora is therefore critical for coastal ecological conservation. Current control methods—physical, chemical, biological, and integrated—have limitations: physical removal alone is often inefficient, and chemical control poses risks of ecological pollution. This study evaluates an integrated approach combining mechanical cutting with low-dose herbicide application to improve control efficacy and assesses its impact on soil microbial diversity. In August 2022, two sites with uniformly growing S. alterniflora were selected, each divided into 21 quadrats of 1 m × 1 m. One site was mowed at the end of August, while the other remained uncut. In September, three concentrations of Haloxyfop-P-methyl (HP) and Glyphosate (GP) were applied. Plant responses were monitored during the flowering stage (September), seed-setting stage (October), the end of the growing season (December), and the following April. Four treatments were compared: HP alone, GP alone, cutting followed by HP, and cutting followed by GP. Soil samples were collected from herbicide-treated and control quadrats for physicochemical analysis. High-throughput sequencing was used to characterize bacterial and fungal community structure and diversity. Data were analyzed using one-way ANOVA with LSD post-hoc tests following validation of homogeneity of variances (Levene’s test). Results showed that: (1) Cutting followed by HP application during the flowering stage effectively controlled S. alterniflora, with a dose of 150 mg/m2 achieving 98.93% control efficiency by the end of the growing season, while doses of 300 mg/m2 and 600 mg/m2 achieved complete plant mortality; all doses significantly inhibited regrowth the following year, with suppression rates of 93.49%, 92.68%, and 96.48%, respectively. (2) Cutting during flowering significantly reduced the sexual reproduction capacity of regenerated plants. (3) Although the combined treatment reduced the relative abundance of Ascomycota in soil fungi, it did not significantly affect overall bacterial or fungal diversity. The combined treatment of cutting followed by Haloxyfop-P-methyl application at 150 mg/m2 effectively controls S. alterniflora and strongly suppresses its regrowth without causing significant adverse effects on soil microbial diversity, supporting the development of efficient and environmentally sustainable management strategies for this invasive species.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Brief discussion on fair use for distribution of open-source large model datasets
    Yunhu ZHAO, Yuzhou YANG, Lin QIN
    J* E* C* N* U* N* S*    2025, 2025 (5): 183-190.   DOI: 10.3969/j.issn.1000-5641.2025.05.017
    Abstract565)   HTML9)    PDF(pc) (741KB)(852)       Save

    The openness of large models requires not only sharing conventional computer software elements such as model architectures and training codes but also disclosing model parameters and datasets. Applying the analytical frameworks of the “four-factor test” and “three-step test” while considering the transformative nature and purpose of dataset distribution under open licenses as well as the public interest in technological development and application, one may conclude that distributing datasets for open-source large models constitutes fair use, thus obviating the necessity for obtaining copyright licenses from upstream right holders. Such an approach satisfies governance requirements regarding artificial-intelligence transparency and actively contributes to promoting knowledge sharing.

    Reference | Related Articles | Metrics | Comments0
    Open-source collaboration structure modeling and multilayer-network link-prediction methods
    Pu ZHAO, Qingxi PENG, Yuang ZHANG, Xiejie JIN, Dezhou ZHAO
    J* E* C* N* U* N* S*    2025, 2025 (5): 109-124.   DOI: 10.3969/j.issn.1000-5641.2025.05.011
    Abstract562)   HTML7)    PDF(pc) (1082KB)(818)       Save

    Collaborative relationships among open-source projects are becoming increasingly complex, involving multiple reuse mechanisms such as dependency co-usage, language consistency, and contributor overlap. Traditional graph models struggle to represent these heterogeneous structures in a unified manner, limiting their ability to identify potential collaboration links. This paper proposes an analytical framework that integrates multilayer graph modeling with structure-aware link prediction, tailored to open-source ecosystems. A three-layer unweighted graph is constructed to capture different types of collaborations, and two structural enhancements—layer overlap modulation and community-aware scoring—are introduced to improve structural perception and semantic interpretability. Experimental results on multiple real-world datasets show that the proposed method consistently outperforms mainstream link prediction algorithms, particularly in networks with high structural heterogeneity. Further analysis reveals that the predicted links exhibit strong community consistency and semantic recoverability. Overall, the proposed approach effectively uncovers latent collaboration paths among open-source projects and provides structural support for reuse modeling and community evolution analysis.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Effects of typhoons landfalling on the north bank of Hangzhou Bay, Shanghai, on saltwater intrusion in the Changjiang River Estuary: A case study of typhoons “Bebinca” and “Pulasan” in September 2024
    Cheng QIU
    J* E* C* N* U* N* S*    2025, 2025 (6): 141-151.   DOI: 10.3969/j.issn.1000-5641.2025.06.015
    Abstract551)   HTML7)    PDF(pc) (4155KB)(102)       Save

    Saltwater intrusion in the Changjiang River Estuary is generally considered a key threat to urban water supply security, and the risks may be exacerbated under the weather conditions of concurrent typhoons and low river discharge. Taking the successive landfalls of typhoons “Bebinca” and “Pulasan” along the north bank of Hangzhou Bay, Shanghai, in September 2024, as examples, we employed a high-resolution 3D estuarine saltwater intrusion numerical model to simulate the differential impacts of the typhoons on saltwater intrusion in the Changjiang River Estuary. The results indicated that whereas the onshore wind-driven circulation promoted a significant enhancement of the frontal intrusion upstream of saltwater in the North Channel during the landfall of typhoon “Bebinca,” the circulation weakened the saltwater intrusion in the South Passage. In contrast, given its less intense strength and deflected track, typhoon “Pulasan” had a comparatively limited impact on the frontal intrusion in the North Channel. Given their limited duration, the short-term storm surges induced by the typhoons, which coincided with spring tides, caused comparatively little saltwater spillover from the North Branch into the South Branch. The findings of this study reveal the response mechanisms of saltwater intrusion to the landfall of typhoons on the north bank of Hangzhou Bay. This will provide a scientific basis for defense against compound natural disasters involving the combined effects of typhoons and saltwater intrusion on urban water supplies under extreme climate conditions.

    Table and Figures | Reference | Related Articles | Metrics | Comments0