This study analyzes the central government budget execution audit rectification reports for the years 2021 to 2023, from the perspective of collaborative governance. Based on the SFIC model, it systematically investigates the pathways to enhance both audit rectification and budget execution performance. The findings reveal that, while significant progress has been made in establishing a multi-dimensional collaborative supervision mechanism and advancing governance logic, several key challenges persist. These include structural flaws in institutional design, inefficient information and feedback channels, the absence of a shared responsibility mechanism, and the breakdown of vertical collaboration transmission. To address these challenges, the study proposes a solution based on the dynamic cyclical logic of the SFIC model. It outlines a four-pronged approach to improving collaborative governance: activating initial conditions, strengthening multi-dimensional catalytic leadership, innovating institutional design, and streamlining collaborative processes. This framework offers actionable recommendations for enhancing the systemic collaborative governance of budget performance improvement within the context of audit rectification.

Precise segmentation of brain tumors is a crucial research focus in clinical medicine. Currently, magnetic resonance imaging (MRI) is widely used for brain tumor segmentation; however. Leveraging multimodality MRI image information can significantly enhance segmentation performance. This paper aims to provide a comprehensive review of the various methods used for multimodality brain tumor segmentation. It explores the application of deep learning technology in this context, and discusses the unique advantages and limitations of methods based on convolutional neural networks(CNN), fully CNNs, and Transformers. The review also summarizes three commonly employed strategies: modal stacking, network improvement, and modal feature fusion. It also covers additional methodologies, including those utilizing generative adversarial networks, feature, correlation, network improvement,and cross-modality self-supervised learning. Furthermore, the paper addresses the challenges and potential solutions for multimodality brain tumor segmentation in the absence of certain modalities. Finally, it provides insights into the future development direction of multimodality brain tumor segmentation.

The modern environmental governance system integrating “digital-intelligent empowerment + interest coordination” can accurately identify stakeholders in water environment collaborative governance and effectively enhance the modernization level of China's “AI+” water environment governance. Drawing upon Zhejiang's digital-intelligent water governance practices, this paper constructs a tripartite evolutionary game model involving “government-enterprise-public” collaborative governance. It reveals the evolutionary pathway of interest coordination and governance strategies in digital-intelligence-empowered water environment governance. The results show that: (1) Government incentive governance, enterprise contractual governance, public collaborative governance constitutes the optimal and ideal stable point for achieving interest-coordinated governance in digital-intelligence-empowered water environments. (2) Stakeholders' recognition of the effectiveness of digital-intelligence-empowered water governance and the evolutionary pattern of public collaborative participation are key drivers for interest coordination within the digital water governance system. (3) The government's incentive regulatory policies with well-defined rewards and punishments can not only strengthen enterprise investment but also effectively enhance public enthusiasm for participating in water governance. (4) Blockchain and AI technologies, as key nexuses of digital-intelligent empowerment, elevate the level of interest-coordinated governance by building intelligent monitoring platforms, thereby propelling water environment governance strategies toward stable equilibrium. The findings of the study not only expand the scope of collaborative value co-creation theory in the AI era but also provide significant reference value for constructing China's modern environmental governance system.

The creation of a positive educational ecology requires a break with the subject-object dichotomy and a transformation from a single individual perspective centered on teachers or students to a relational perspective of building a dialogue-based, co-subjective relationship between teachers and students. Grounded in humanistic principles, the teacher-student co-presence dialogue is an emerging educational proposition, which examines the teacher-student relationship from the perspective of mutual development and views teachers and students as a unified community. It demonstrates the key characteristics of relationality, collaborativity, and practicality as a matter of priority, all of which together highlight the status of teachers and students as co-subjects, establish the goal of their shared development, and form the concept of joint creation through collaboration. In this sense, constructing a path for the co-presence dialogue entails shifting from one-sided inclination to bidirectional equality in educational roles, transforming external impetus into internal generation as the driving force for growth, and evolving from clear-cut distinctions to co-present openness in the teacher-student relationship. Ultimately, these shifts will contribute to achieving the goal of mutual enhancement in teaching and learning.

The influence of Li Shangyin's marriage on the trajectory of his life has been widely acknowledged; however, the underlying reasons for the markedly different treatment he received from the two rival political factions have not been systematically examined. By analyzing the marital practices of key members of the Niu and Li factions, this study demonstrates that the Niu faction relied heavily on marriage alliances as a central mechanism for factional formation and consolidation, whereas the Li faction, by contrast, did not employ affinal ties as a strategy for building factional networks. This structural divergence in faction-building practices resulted in sharply contrasting responses to Li Shangyin's marriage to the daughter of Wang Maoyuan. The Niu faction interpreted the marriage as an act of betrayal against the Linghu family and consequently subjected Li Shangyin to political suppression, while the Li faction largely disregarded the marriage and accorded it little political significance. The fundamentally different attitudes toward marriage, rooted in the factions' distinct organizational logics, provide a crucial analytical key to understanding the political tragedy that ultimately shaped Li Shangyin's life.

The extraction of buildings from remote sensing images faces multiple challenges such as diverse shapes, scale changes, complex backgrounds, and pixel blurring. To address these issues, this paper proposes a dual path network called DBCTNet, which combines the advantages of CNN and Transformer. This network integrates local detail paths with global context paths, fully utilizing the local feature extraction capabilities of CNN and the global information processing capabilities of Transformer, thereby achieving more efficient building extraction. In order to further improve the performance of the model, a feature refinement fusion module is proposed to reduce the feature differences between different paths and enhance the adaptability of the model in complex scenes. At the same time, a multi-level feature aggregation module is proposed to improve the robustness and accuracy of the model when dealing with buildings of different scales by aggregating multi-scale feature information. Experimental results showthat DBCTNet achieves excellent performanceon the Massachusetts building dataset and the Inria aerial image labeling dataset.

The persistence and resistance to degradation of antibiotic wastewater pose a serious threat to human health and the ecological environment. Electrocatalytic oxidation has emerged as an effective approach to address the challenge of recalcitrant antibiotic wastewater. This study focuses on three representative antibiotics, i.e., fluoroquinolones, amphenicols and tetracyclines,and systematically reviews their degradation mechanisms. Additionally, it delves into the current strategies for modifying electrode materials used in the electrocatalytic treatment of recalcitrant antibiotic wastewater. These strategies primarily include introducing intermediate layers to optimize charge transfer, doping with metal elements to enhance catalytic activity, and incorporating metal oxides and non-metallic materials to improve electrode stability and selectivity. These modifications have significantly improved the catalytic activity and overall efficiency of electrocatalytic reactions.Finally, the paper provides an outlook on the development directions and application prospects of three-dimensional electrocatalytic systems, novel carbon-based composites, metal oxides and nanostructured electrode materials, aiming to offer systematic guidance and reference for both fundamental theoretical research and industrial engineering application of electrocatalytic technology in antibiotic wastewater remediation.

Using China's provincial panel data from 2013 to 2021, this study constructs a comprehensive evaluation index system for rural digital infrastructure in China. The Dagum Gini coefficient, kernel density estimation, and the natural breaks classification method are employed to examine its development level and regional disparities. Furthermore, the DEA window analysis method is applied to assess the efficiency of rural digital infrastructure in narrowing the urban-rural income gap, income disparities across rural regions, and income inequality among farmers. The results indicate that although China's rural digital infrastructure has experienced steady improvement, its overall development level remains relatively low and growth has been modest. A pronounced spatial gradient is evident, with development levels declining from the eastern coastal regions toward inland areas. Interregional disparities contribute approximately 50% of the overall regional inequality. From a dynamic perspective, both overall and interregional disparities in rural digital infrastructure development exhibit a downward trend nationwide. The relative gap between the western region and the eastern and central regions has narrowed, although substantial intraregional disparities persist within the western region. Marked intraprovincial heterogeneity is observed, giving rise to a five-tier structure comprising low, relatively low, medium, relatively high, and high development levels. In terms of overall gap-narrowing efficiency, rural digital infrastructure development in China at the current stage largely conforms to the requirements of moderately forward-looking development. Improvements in comprehensive gap-narrowing efficiency are primarily driven by gains in technical efficiency, while the capacity to translate technological conditions into effective scale and coverage benefits remains limited. On this basis, policy recommendations are proposed from three perspectives: moderately forward-looking investment, comprehensive and coordinated development, and region-specific advancement of rural digital infrastructure.

Adverse childhood experiences (ACEs) exert both a direct influence on negative emotions in college staderts and an indirect effect mediated by social alienation. A greater number of ACEs is associated with heightened social alienation, which in turn leads to increased negative emotions. The perception of involution moderates this pathway by strengthening the positive relationship between ACEs and social alienation, as well as between social alienation and negative emotions. A stronger perception of involution augments the indirect effect of ACEs on negative emotions through social alienation, thereby amplifying the overall impact of ACEs. Introducing the concept of involution perception into research on negative emotions not only deepens theoretical understanding of their development but also offers valuable insights for designing targeted interventions.

Original and disruptive technological innovation serves as the wellspring of scientific and technological progress. It provides essential support for achieving greater self-reliance and strength in science and technology, and for seizing strategic opportunities presented by a new wave of scientific and technological revolution and industrial transformation. Moreover, it plays a decisive role in opening up new domains and pathways and in fostering new-quality productive forces. China has made sustained progress in basic research, capacity building, and the development of application scenarios, thereby laying a solid foundation for advancing original and disruptive innovation and facilitating the emergence and application of a series of significant achievements. Nevertheless, several critical challenges persist, including limited capacity to tackle major original scientific problems, an insufficient pool of top-tier talent, and inadequate mechanisms for translating research outcomes into practical applications. These challenges largely stem from a misalignment between existing support systems and the intrinsic laws of innovation, the absence of a fully developed environment that encourages exploration and tolerates failure, and talent cultivation and utilization mechanisms that remain ill-suited to the demands of breakthrough innovation. Addressing these issues requires a comprehensive approach grounded in top-level strategic design. Priority should be given to strengthening the coordinated allocation of resources and optimizing support mechanisms, while deepening cross-disciplinary collaboration to systematically cultivate high-caliber innovative talent. In addition, it is essential to establish dedicated support institutions and channels, reinforce long-term and stable investment, and foster an innovation ecosystem that encourages bold exploration and breakthrough. Together, these measures will provide a robust institutional foundation for the sustained advancement of original and disruptive technological innovation.

John Rawls' theory of overlapping consensus cannot resolve the issue of the lack of consensus reflected in the frequent social upheavals in the United States. The practical dilemma lies in the disjunction of the overlapping consensus, while the theoretical dilemma arises from the ahistoricity caused by the presuppositions of the theoretical assumptions. The perspective of historical materialism reveals that the root cause of this ahistoricity is Rawls' use of a “historical dualism”—by dividing political and non-political domains, an overlapping consensus is achieved in the latter with respect to the former. However, the growing political demands of the non-political domain challenge the original political framework, and identity politics emerges as a practical consequence. Thus, the ahistorical consensus, represented by the overlapping consensus based on the “historical dualism,” is unlikely to genuinely foster social unity and development; instead, it undermines the existing consensus resources due to its disconnection from history. A historical consensus, which is generated within the totality of history and drives historical progress, possesses the cohesion and leadership to strengthen social unity and promote the sustainable development of social consensus itself.

Electroencephalography (EEG) is a non-invasive technique with high temporal resolution and low cost, making it of significant value in visual neural decoding research. Visual neural decoding technology based on EEG signals aims to reconstruct or recognize perceived visual content from the brain activity of subjects. This paper provides an overview of visual neural decoding technology based on EEG signals, encompassing its theoretical foundations, key steps, datasets, and evaluation metrics. It also summarizes and analyzes research directions in this field, including EEG visual semantic classification, image generation, and 3D object reconstruction. Finally, to address the common challenges currently faced, this paper forecasts future development trends from the perspectives of multimodal fusion and lightweight modeling.

Digital transformation has become a critical pathway for enterprises to enhance their core competitiveness. However, identifying a transformation path that aligns with firms' specific conditions remains a major challenge for manufacturing enterprises. Focusing on the driving mechanisms underlying digital transformation in manufacturing firms, this study integrates six key conditional variables from two dimensions-internal organizational factors and external environmental factors-to construct a performance-driving factor model for digital transformation. Employing the fuzzy-set Qualitative Comparative Analysis (fsQCA) method, the study conducts a configurational analysis using data from listed manufacturing companies in Zhejiang Province. The findings indicate that: (1) no single internal or external driving factor constitutes a necessary condition for achieving high digital transformation performance; rather, successful digital transformation is more likely to result from the configurational alignment of multiple factors; (2) the configurations associated with high digital transformation performance can be classified into three categories: internally driven configurations, including the “digital technology-driven” and “peer pressure-driven” types; externally driven configurations, including the “environment-talent-attention-driven” and “attention-driven” types; and internally-externally synergistic configurations; and (3) compared with the other two pathways leading to high digital transformation performance, the internally-externally synergistic pathway demonstrates greater coverage. In contrast, low digital transformation performance primarily stems from the combined absence of multiple conditions, including digital attention allocation, policy support, and environmental dynamism. These findings provide theoretical insights for manufacturing enterprises seeking to overcome challenges in digital transformation and improve transformation performance.

Promoting and deepening reforms in the mergers and acquisitions (M&A) restructuring market constitutes an important strategic initiative of the State Council, while synergy effects represent the core driving force behind the success of corporate M&A activities. This study aims to construct a comprehensive M&A synergy effect index and empirically examine its effectiveness. Based on an exploration of the formation mechanisms and realization pathways of M&A synergies, this paper develops an indicator system and index model for comprehensive M&A synergy effects from the perspective of value creation, with operational synergy, managerial synergy, and financial synergy serving as the core dimensions. Using A-share listed companies that engaged in M&A activities in 2019 as the research sample, the study further evaluates the reliability and validity of the proposed index. The empirical results indicate that the index maintains a high degree of weight stability over a five-year observation period and effectively explains both long-term changes in corporate value and short-term market expectations, thereby demonstrating strong stability and practical applicability.

“Zhexue” (Zhejiang Studies) represents a distinguished form of traditional Chinese regional culture, rich in ideological and political education resources. Rooted in core values such as moral cultivation, pragmatism, responsibility, erudition, and academic excellence, it carries profound significance for guiding college students to set lofty aspirations, uphold moral integrity, develop their talents, and shoulder national responsibilities. This study advocates for the integration of “Zhexue” into ideological and political education for college students by adhering to principles of systematicity, coherence, and practicality. It explores the inherent value of this integration and proposes effective strategies for embedding “Zhexue” in educational practice. The goal is to advance the inheritance and development of “Zhexue,” enhance the effectiveness of ideological and political education, and better cultivate a new generation capable of undertaking the great mission of national rejuvenation.

With the advancement of modern active phased-array radar systems toward high-density integration, lightweight design, and multifunctionality, conventional microwave circulators, which rely on bulky external permanent magnets for biasing, have become a major obstacle to the further miniaturization and integration of transmit/receive (T/R) modules. Self-biased circulators replace the external magnetic bias field by exploiting either the intrinsic magnetocrystalline anisotropy of magnetic materials or structurally induced shape anisotropy, thereby offering an effective route for the planarization and high-level integration of nonreciprocal microwave devices. This paper systematically reviews recent progress in key magnetic material systems that enable self-biasing mechanisms, including hexagonal ferrite bulk materials, ferrite thick films, and magnetic nanowire arrays. In M-type hexagonal ferrites, magnetocrystalline anisotropy can be tailored via ion substitution and additive doping to reduce microwave losses. Ferrite thick films, fabricated by tape casting and screen printing, demonstrate strong potential for low-temperature co-fired ceramic integration and quasi-self-biased device design. Magnetic nanowire arrays, benefiting from their high aspect ratio, possess pronounced shape anisotropy and exhibit promising nonreciprocal performance in planar transmission lines such as microstrip lines and substrate integrated waveguides. Overall, self-biased circulators still face critical challenges, including relatively high microwave losses in magnetic materials, poor compatibility with planar microwave fabrication processe, and constraints on large-scale integration. Future developments will rely on the synergistic optimization of magnetic anisotropy regulation and low-loss mechanisms, with continued progress toward higher operating frequencies and higher levels of integration.

To overcome the limitations of relying on external internet influencers in rural digital transformation and to explore the endogenous drivers of rural revitalization, this study examines the digital evolution of local rural influencers through the lens of social capital theory. Using three native rural influencers with over one million followers on the Kuaishou platform as case studies, the research adopts content analysis and categorizes the evolution process into three stages: initial, growth, and maturity. An analytical framework integrating network relationships, trust-building, and reciprocity norms is developed, and a fixed-effects model is employed to identify the factors influencing communication effectiveness. The findings reveal a three-stage evolutionary pattern: initial stage (social capital acquisition), growth stage (social capital maintenance), and maturity stage (social capital transformation). The common development path can be summarized as “traffic accumulation→commercial monetization→social empowerment.” Key social capital elements, including co-creation and fan incentives, as well as content features such as editing techniques and video duration, significantly influence communication outcomes. Notably, authenticity and strong local embeddedness emerge as the core competitive advantages of rural influencers. The proposed three-stage evolution model offers both theoretical and practical contributions, providing a robust framework for understanding the development of local rural influencers and actionable guidance for their cultivation in practice.

This paper reviews recent advances in thermoelectric generator (TEG) systems integrated with phase change materials (PCM), along with their application prospects in various structural designs and material deployments. The basic principles of TEG-PCM and the thermoelectric effect are first introduced, followed by a summary of their classification criteria and configuration types. The research on the application of TEG-PCM system in solar PV power generation, including component optimisation and structural optimization, is analyzed and discussed, and the output efficiency of the integrated system is improved by 20%～30% compared to the conventional PV system. The prospects for the development of more efficient TEG-PCM systems and their hybrid configurations are predicted based on the current level of technology.

Investment-financing maturity mismatch has become increasingly prevalent among Chinese enterprises. The associated risks of capital chain disruptions and debt defaults not only constrain the healthy development of the real economy but also pose significant challenges to financial market stability. As one of the most representative frontier technologies, artificial intelligence (AI) not only facilitates the transformation and upgrading of new-quality productive forces but also offers a novel pathway for alleviating corporate investment-financing maturity mismatch and maintaining financial stability. This study examines the impact of AI technology on corporate investment-financing maturity mismatch and explores the underlying mechanisms, with the aim of providing new insights into addressing structural mismatches between investment and financing maturities in the real economy. The results show that AI technology significantly mitigates investment-financing maturity mismatch, with more pronounced effects observed in high-tech enterprises, firms located in regions with higher levels of digitalization, enterprises with weaker human capital, and firms that are more vulnerable to financial distress. Further mechanism analyses indicate that AI technology alleviates maturity mismatch primarily by improving asset allocation efficiency and enhancing firms' access to long-term credit. These findings offer important implications for promoting the adoption of AI technologies by enterprises and for refining regulatory policies related to investment-financing maturity structures.

AI-generated knowledge search is catalyzing a paradigm shift from “retrieving existing information” to “generating knowledge on demand,” thereby emerging as a novel approach to knowledge search that fosters enterprise innovation. From the perspective of promoting the adoption of AI-generated knowledge, this study examines how organizational members choose between internal and external AI-generated knowledge search strategies through two scenario-based experiments. The findings indicate that, compared with external search, internal AI-generated knowledge search significantly enhances employees' perceived credibility of AI-generated knowledge, which in turn increases their willingness to adopt it. Furthermore, in the context of searching for AI-generated technical knowledge, internal search demonstrates a stronger positive effect on perceived credibility than external search. In contrast, when searching for AI-generated market knowledge, no significant difference is observed between internal and external search strategies in terms of their impact on perceived credibility.

While intelligent algorithmic technologies have significantly enhanced efficiency, they have also given rise to a new form of social control conceptualized as the “algorithmic cage.” From the perspective of its formation mechanism, the “algorithmic cage” emerges from the interplay of technological, cognitive, and power logics. Grounded in the value orientation of the free and comprehensive development of human beings, humanistic economics provides a human-centered analytical perspective for addressing the humanistic crises induced by the “algorithmic cage.” From this standpoint, the “algorithmic cage” gives rise to four interrelated value dilemmas: the crisis of human subjectivity, the illusion of free choice, the decline of public rationality, and challenges to fairness and justice. Accordingly, its governance should be advanced through a coordinated framework that strengthens human agency, innovates digital technology paradigms, embeds humanistic value principles at an early stage, and enhances institutional resilience.

In the aftermath of the Sino-Japanese War of 1894—1895, Japan utilized the substantial indemnities extracted from the Qing Dynasty to accelerate its industrialization. Under the support of national fiscal and financial policies and guided by state-led capital, Japan achieved simultaneous development of light and heavy industries, thereby laying the foundation for a modern capitalist economy. Concurrently, extensive government investment in military expansion increasingly infused Japanese capitalism with militaristic characteristics, imbuing industrialization with a pronounced military orientation and a clear focus on supporting external war of aggression. Thus, warfare and expansionism were not only instrumental in facilitating Japan's modernization but also constituted key drivers and pathways in the broader trajectory of its industrial and modern development.

The digital era has markedly increased the demand for electronic evidence in cross-border criminal investigations, exposing significant challenges in China's traditional cross-border criminal e-evidence system, particularly in system design, procedural arrangements, and the protection of rights. To facilitate the system's transformation, it is essential to assess its current status and limitations, analyze the underlying causes of existing problems, and propose reform pathways that effectively support cross-border criminal investigations. China can advance cross-border e-evidence practices by establishing robust international law enforcement cooperation mechanisms and transitioning jurisdictional frameworks toward a sovereignty-adjusted “data controller” model. For the traditional system, creating expedited channels for mutual legal assistance and strengthening routine inter-state cooperation can enhance the efficiency and reliability of criminal judicial assistance. Additionally, implementing tiered electronic evidence collection rules through supporting legislation can safeguard both national sovereignty and citizens' fundamental rights. These measures collectively contribute to the development of a new cross-border criminal e-evidence system with Chinese characteristics, better aligned with the demands of the digital era.

The evolution of civilization follows inherent laws. It is premised on “real individuals and their practical activities”, driven by the movement of fundamental social contradictions and the expansion of social interaction, manifested in the reconfiguration of civilizational forms under technological transformation, and oriented toward the all-round development of human beings. In the contemporary era, digital technologies are generating new-quality productive forces, reshaping social life, transforming cultural landscapes, and enhancing human development, thereby marking humanity's transition into a new stage of digital civilization. However, the evolutionary trajectories of digital civilization have diverged. Under distinct institutional logics, digital civilization has bifurcated into two contrasting paths: digital capitalism, which tends to exacerbate economic inequality, labor alienation, and global polarization; and digital socialism, which emphasizes the shared ownership of the means of production, effective regulation of capital, and the promotion of comprehensive human development. At this critical historical juncture, China should prioritize the cultivation of new-quality productive forces, advance the deep integration of China's fine traditional culture with digital civilization, and uphold the principles of openness, inclusiveness, and mutual learning, so as to contribute to the formation of a new model of digital civilization.

In this paper, the finite-time mixed-performance analysis and filter design of a class of discrete-time interval type-2 Markovian jump systems are studied. Unlike the homogeneous Markovian jump systems, the time-varying transition probability is in a convex bounded domain with all vertices known. A mixed-performance structure is introduced to unify them in a framework. In the case that the transition probability matrix is time-dependent, the conditions of finite time boundedness and satisfying H_∞/L₂-L_∞ performance of the filtering error system are given. By transforming the given parameters, both H_∞ performance and L₂-L_∞ performance can be considered simultaneously, enhancing the flexibility of the design. By introducing two modal-dependent matrix variables, the nonlinear matrix inequality is transformed into a linear matrix inequality, and the filter is designed. Numerical simulations have been executed, revealing the efficacy of the proposed method.

Current brain-computer interface (BCI) systems typically migrate either EEG signal acquisition or decoding functions to embedded platforms. However, these systems often face challenges such as complex architectures and dispersed functional modules. To address these issues, an embedded BCI system based on a compact single-board computer, the Raspberry Pi, has been developed. This system integrates EEG signal acquisition, preprocessing, feature extraction, classification, and application interfaces into the Raspberry Pi platform, in combination with a custom-designed analog front-end module, achieving full-process EEG signal processing in an embedded environment. The system was first validated using a steady-state visual evoked potential (SSVEP) paradigm, achieving an average classification accuracy of 97.1%. To further evaluate its paradigm generalizability, a motor imagery (MI) experiment was conducted, yielding an average classification accuracy of 85.2%. The experimental results demonstrate that the system can not only effectively acquire and decode EEG signals, but also exhibits robust adaptability and generalizability across paradigms, offering a novel technical solution and development path for highly integrated embedded BCI systems.

As the performance of atomic magnetometers improves, magnetic field noise arising from coil current fluctuations has gradually become the dominant factor limiting their sensitivity. To suppress the magnetic field noise introduced by current sources in coil driving and to overcome the limitations of classical technical noise on measurement sensitivity, this paper presents a low-noise current source design tailored for atomic magnetometers. The proposed design employs lithium batteries for power supply, incorporates LC low-pass filtering and supercapacitors for noise reduction, achieves programmable output via a weighted resistor network with feedback control, and utilizes a transistor-based current mirror structure for constant-current output. To evaluate its noise performance, the current source was characterized using two measurement methods: spin noise spectroscopy and dynamic signal analysis. Experimental results show that at an output current of 20 mA, the current noise density is 557 pA/Hz at 1 kHz and 85.7 pA/Hz at 50 kHz, representing an overall noise reduction of approximately one order of magnitude compared to existing commercial low-noise current sources in the laboratory. This design effectively suppresses background magnetic field noise, meets the requirements of atomic magnetometers for low-noise, wide-range coil driving currents, and thereby provides reliable technical support for enhancing the measurement performance of atomic magnetometers.

This paper discusses the classification of block-transitive point-primitive 2-(v,4,λ) designs, and presents the following conclusions: Let $\mathcal{D}$ be a 2-(v,4,λ) design, if G≌M₁₁ acts block-transitively on $\mathcal{D}$, then the parameter (v,λ) of $\mathcal{D}$ can only be one of the following parameters:(11,36),(12,15),(12,30),(55,8),(55,16) and (55,32).

This study empirically examines the impact and transmission mechanism of environmental regulation on green innovation performance using panel data from 30 Chinese provinces between 2010 and 2022. Employing mediation, moderation, and spatial Durbin models, the findings reveal a “U-shaped” relationship between environmental regulation and green innovation performance, with notable regional differences across the eastern, central, and western regions. Environmental regulation indirectly influences green innovation performance through regional competition and human capital optimization. Furthermore, foreign direct investment (FDI) plays a “U-shaped” moderating role in the relationship between environmental regulation and green innovation performance. The analysis also highlights a positive spatial spillover effect of environmental regulation on green innovation performance. Based on these results, the paper offers policy recommendations aimed at enhancing green innovation performance within the context of the new development paradigm.

The integration of smart business, finance, and tax functions (business, finance, and tax) is a pivotal pathway for corporate transformation and upgrading in the digital economy. At its core, this integration seeks to leverage digital technologies and process reengineering to enable collaborative value creation across these domains. With the advancement of the Golden Tax Phase IV project, tax administration is increasingly transitioning to a “data-driven taxation” model, which necessitates a fundamental reform in the training of finance and tax professionals within higher education. This study employs both bibliometric analysis and job demand-based quantitative methods to identify key gaps in current finance and tax talent development, particularly in the areas of intelligent tool usage, business-finance-tax integration, and data analysis skills. Building on these insights, the paper proposes a comprehensive knowledge framework that integrates “business + finance and tax + technology” and introduces a “Three-Classroom Linkage” teaching model. Practical explorations, such as the establishment of an integrated “smart business-finance-tax” curriculum, the expansion of immersive learning environments, and the creation of collaborative school-enterprise smart finance and tax training platforms, aim to facilitate the transformation of students from “tax accountants” to “data governors.” This approach provides a strategic reference for cultivating multi-disciplinary talents capable of adapting to the modernization of finance and tax governance.

Understanding how fiscal support affects agricultural new quality productivity is essential for driving innovation in traditional agriculture and developing a modern agricultural sector. Using panel data from 31 Chinese provinces (2011—2022), this study examines the relationship between fiscal support and agricultural new quality productivity through the lens of resource allocation's dual effects, while also considering the mode-rating roles of scale operation and technological environment. The results indicate a nonlinear U-shaped relationship: fiscal support initially suppresses but later promotes agricultural new quality productivity. Scale operation significantly strengthens this U-shaped relationship, whereas the technological environment weakens it. These findings suggest that policymakers should account for the varying effects of different fiscal support levels, adapt policies flexibly, promote appropriate agricultural scale operation, and improve the technological environment to effectively advance agricultural new quality productivity.

Named Entity Recognition (NER) stands as a pivotal task in the realm of natural language processing. However, the Lattice-LSTM model, incorporating lexical information, exhibits limited accuracy in Chinese NER. To tackle this issue, we propose the Bert-BiLattice-LSTM-MulAt-CRF model. Firstly, this model preprocesses the dataset using the Bert model and subsequently employs BiLattice-LSTM to integrate word-level features for bidirectional semantic encoding of Chinese text, which effectively leverages bidirectional information for each character in a word. And then, the model employs a multi-head attention mechanism for weighted operations, enhancing the extraction of deep semantic information within sentences. Finally, the process concludes with semantic decoding through CRF. Experimental results indicate that the model achieves an F1 score of 95.19% on the MSRA dataset and 95.47% on the Resume dataset. Compared to the baseline model Lattice-LSTM, the F1 scores have improved by 2.93% and 2.88%, respectively.

In contemporary Chinese families, the core values of traditional family virtues have been endowed with new meanings, while elements of red family tradition ideology are increasingly evident. However, under the pressures of an intensely competitive educational environment, Chinese families face a range of challenges, including heightened anxiety, intergenerational communication barriers, and strained parent-child relationships. Based on a systematic review of family and family virtue construction concepts, this study argues that the core of modern family development lies in cultivating family virtues, with red family tradition ideology playing a significant role in nurturing contemporary family values. Red family traditions embody political virtues of loyalty to the family and alignment with the nation, emotional care characterized by disciplined household governance and affectionate family relationships, and a spirit of striving and resilience in overcoming difficulties. These represent valuable ideological resources for fostering family virtue construction in modern households. In practice, cultivating family virtues through red family traditions requires moving beyond a narrow focus on academic “scores” to reconstruct the family foundation, strengthening family community, balancing emotional support and pressure, and identifying resilience stress points to establish healthy family relationships.

The unnatural motion of non-player characters (NPCs) is a common issue in pixel art games. This is especially obvious when an NPC is in motion where it usually moves along a pre-fixed path with a stiff motion posture,which significantly detracts from the game's fun, mystery, and/or vitality. To handle the above issue, i.e., generating a concise, smooth, and unpredictable motion path where an NPC can move along it by using adaptable postures, this study proposes a new motion generation approach based on optimization. Firstly, a novel two-level optimization method is presented, which integrates the RRT-Connect algorithm, particle swarm optimization, and Bresenham algorithm. It can rapidly, effectively, and automatically generate a preliminarily feasible and high-random motion path while also optimizing motion postures. Secondly, a greedy strategy is applied to simplify the generated path, shortening the length and improving the smoothness. Finally, based on the particle swarm optimization, a smoothing corner method is developed to further optimize the above-mentioned path and its related postures as well, resulting in final high-naturalness motion. The effectiveness of the proposed approach is verified through tests on several maps of varying complexities and NPCs with different shapes and sizes. The results demonstrate that the generated motion paths exhibit high randomness and the motion postures show significant adaptability. Meanwhile, compared with the existing related works, the proposed approach demonstrates great potential in enhancing the naturalness of NPC motions in pixel games.

This study conducts a visual analysis of 50 650 research articles on college student entrepreneurship published since 1993 and indexed by CNKI. Drawing on historical institutionalism and bibliometric methods, it maps the evolution of academic attention and identifies key research hotspots in the field. In terms of developmental trajectory, the research can be divided into three phases: the spontaneous exploration period, the policy-driven boom period, and the transformation period characterized by an emphasis on high-quality entrepreneurship. Through data mining and cluster analysis, four major thematic clusters are identified: entrepreneurship as a driver of employment, entrepreneurship education, entrepreneurial practice, and student-led ventures. The study further examines the leading research institutions, most productive universities, and influential scholars in the field, analyzing their contributions and research orientations. Looking ahead, future research is encouraged to shift from qualitative to quantitative approaches, to focus more on college entrepreneurship within the context of the digital economy, and to explore comparative dimensions across regions, development stages, and disciplinary backgrounds. These findings help to clarify current trends and offer scientific guidance for the further development of college student entrepreneurship research in China.

The rising demand for online audio and video technologies like video conferencing has driven considerable interest in speech enhancement. This study focuses on multi-channel speech enhancement with a two-stage processing algorithm. Firstly, the multi-channel data from a microphone array is performed using a beamforming algorithm based on an improved filter and sum network to suppress coherent noise. Then, a convolutional long short-term memory network is employed to remove the residual reverberation and noise components from the beamformer outputs. Experimental results demonstrate that the processingtwo-stage multi-channel speech enhancement algorithmcan effectively improve speech quality and comprehensibility.

The evolution of literature and art is, at its core, a history of media technology. From ancient rock paintings and bronze ware to the invention of printing, and further to digital networks and artificial intelligence, each technological leap has not only transformed how art is disseminated but also fundamentally reshaped its creative mechanisms, aesthetic paradigms, and modes of existence. Printing enabled the popularization of literature and the professionalization of authorship, while digital media has dissolved the traditional “literary field,” giving rise to a new ecology in which multiple sub-fields-elite, popular, and online literature-coexist. A significant shift driven by new media technologies is the transition in aesthetic theory and practice from “intertextuality” to “intermediality,” underscoring the role of media not as a passive conduit but as an ontological force actively constituting meaning. Technologies such as AIGC, VR, and hypertext have destabilized the notion of the “author” while reorganizing narrative structures, modes of reception, and aesthetic experience. Today, AI-generated content further challenges the traditional axiom that “literature is the study of humanity,” and the metaverse turns readers from detached observers into immersed participants. Although technical logic may sometimes override cultural logic, and instrumental rationality can impinge on expressive rationality, practice shows that the muse has not faded. Rather, contemporary online literature and art are gathering renewed energy precisely through the tension between technology and culture. To truly grasp the “digital-intelligent creation” and future direction of Chinese online literature and art, contemporary theory must move beyond conventional perspectives and fully acknowledge the ontological significance of media.