Abstract To enhance the paper's integrity and emphasize an engineering perspective, this study systematically integrates elements of Mechanical Design, Manufacturing, and Automation (MD&A) into the existing framework for regional development disparity research. Based on city-level panel data, we first construct the Urban-Urban Development Disparity Index (UUDI) using the TOPSIS-Entropy Weighting Method for multi-indicator comprehensive measurement, followed by PCA for robustness testing. Second, within a panel fixed-effects framework, we identify the impact of digital finance (DFI) on regional disparities. We incorporate MD&A indicators—including prefabricated construction, BIM-MES-SCADA integration, robot/CNC equipment density, and first-pass yield—to map the transmission chain: "capital accessibility → manufacturing automation investment → engineering supply efficiency/quality → urban performance → gap convergence."Findings reveal that digital finance exhibits significant overall equilibrium effects, particularly pronounced in cities with higher automation foundations. Mechanistically, equipment renewal financing and data integration enhance component manufacturing and assembly efficiency while reducing rework rates, thereby improving public works supply and residential accessibility, ultimately promoting coordinated regional development. This study provides a unified metric and engineering implementation framework for subsequent causal identification and policy simulation.

Abstract This study constructs a comprehensive index of regional development disparities based on city-level panel data and employs a fixed-effects model to examine the impact of digital finance development on regional gaps. Benchmark regression reveals that digital finance significantly narrows regional development disparities (with a coefficient of approximately −0.127 across the entire sample), exhibiting a stronger effect in less developed regions (approximately −0.231) and demonstrating pronounced inclusive characteristics. Robustness checks—Including dependent variable substitution (index reconstruction via PCA/TOPSIS), instrumental variables two-stage least squares (Kleibergen–Paap F≈15), and placebo tests with 500 random permutations—Consistently confirm the direction and significance of these findings. Mechanism identification reveals that optimized financing environments serve as the key mediating channel (accounting for approximately 38.6% of the mediating effect), while factors such as entrepreneurial activity, corporate innovation investment, and industrial structure upgrading significantly moderate the equilibrium effects of digital finance. This study provides data support and policy implications for narrowing regional disparities and enhancing factor allocation efficiency.

Abstract This paper comprehensively discusses the fault detection and diagnosis techniques of inverter circuits. It begins with an introduction to the fundamental principles of inverter circuits and their applications in solar photovoltaic systems, electric vehicles, and uninterruptible power supplies (UPS). Inverter circuits achieve efficient transmission and utilization of electrical energy by converting direct current (DC) into alternating current (AC). However, due to the non-ideal characteristics of switching devices and environmental factors, the output waveform of inverter circuits may become distorted and contain harmonics, impacting system performance. Therefore, a deep understanding of the basic principles of inverter circuits is critical for effective fault detection and diagnosis. Following this, the paper analyzes common fault types and electrical characteristics of inverter circuits, including overheating, short circuits, open circuits, and parameter drift of switching devices. To extract fault features, signal processing methods such as Fourier transform and wavelet transform are widely applied. Meanwhile, machine learning techniques, especially support vector machines (SVMs) and neural network algorithms, are used for accurate identification and early warning of complex failure modes. Regarding fault detection methods, the paper delves into fault detection technologies based on signal processing and machine learning. Signal processing technology analyzes voltage and current signals and employs Fourier and wavelet transforms to extract fault characteristics. Machine learning algorithms then enable precise fault recognition and advance warnings. Finally, this paper proposes a design framework for an inverter circuit fault diagnosis system that includes key technologies such as data acquisition, signal processing, and fault detection and diagnosis algorithms. The research results indicate that the integration of advanced signal processing technology and machine learning algorithms can significantly enhance the accuracy and efficiency of fault detection and diagnosis for inverter circuits, providing a robust guarantee for the stable operation of power electronic systems.

Abstract This study constructs the "Regional Development Gap Index (UUDI)"based on prefecture-level city data from Guangdong Province spanning 2013–2023. Employing the TOPSIS-entropy weighting method and validating robustness through PCA, it examines the impact of digital financial development (DFI) on regional disparities. Construction and mechanical design, manufacturing, and automation (MD&A) are embedded as pivotal sectors for in-depth analysis. The indicator system incorporates green and digital elements from construction (prefabrication share, green building area, energy/carbon intensity per unit floor area) and MD&A linkage metrics (prefabricated component standardization rate, BIM-MES-SCADA integration, CNC/robotic equipment density, on-site automated sensor coverage) to reflect the transmission chain: "capital accessibility—Manufacturing capacity—Engineering supply—Urban performance."Descriptive statistics reveal: UUDI mean 0.52, standard deviation 0.18, with significant regional variation; DFI mean 0.68, standard deviation 0.23, showing high concentration in the Pearl River Delta; 67% of cities have high-speed rail access, with financing convenience significantly better in the Pearl River Delta. Mechanism analysis reveals that digital finance alleviates financing constraints for equipment renewal, promotes coordination within the prefabricated construction supply chain, and stimulates automation investment. This enhances engineering supply efficiency and quality, ultimately narrowing regional disparities. This study contributes by: proposing a unified measurement framework linking "digital finance—MD&A—Construction industry—Urban performance”; developing reusable UUDI metrics and embedded MD&A indicators; and establishing a data and methodological foundation for future causal identification and policy simulation.

Abstract Through literature review, research surveys, and case studies, it systematically examines the practical basis and value implications of such integration. The paper actively investigates the logical mechanisms, practical challenges, and implementation strategies of digital technology in promoting high-quality development of physical education in private universities. With the rapid advancement of technologies such as cloud computing, big data, and artificial intelligence, digital empowerment has become a crucial approach to fostering innovation in university sports.

Abstract Driven by the combination of high-resolution remote sensing and deep learning, road extraction research has advanced along three main lines (2023-2025): semi- and weakly supervised learning, CNN-Transformer hybrid architectures, and graph/vector modeling. Semi-supervision has proven effective in robustly improving road connectivity in annotation-scarce scenarios. Hybrid architectures that fuse local texture and global context are gradually replacing single CNN/Transformer architectures in feature representation. Direct prediction of road network graphs (nodes and edges) has significantly improved topological consistency and achieved breakthroughs in large-scale inference efficiency. This paper reviews representative methods and datasets/evaluation systems, and provides a horizontal comparison based on publicly available comparison results from the past two years. It summarizes current bottlenecks and trends: Learning paradigms focused on connectivity, cross-domain generalization, and zero-shot geographic transfer, along with more economical supervisory signals, will be the primary areas of focus.

Abstract This paper dives into how artificial intelligence (AI) is actually being used in smart city transportation systems and breaks down the step-by-step ways it goes from predicting traffic flow to boosting the efficiency of real-world social operations. It analyzes the principles and advantages of AI technology in traffic flow prediction and elaborates on how AI enhances social operation efficiency through optmizing traffic management, improving travel experiences, and promoting resource allocation. Research shows that AI empowering smart city transportation systems has significant practical significance and broad development prospects, but it also faces challenges such as data security, technical standards, and talent shortages. This paper proposes countermeasures including strengthening data protection, formulating unified standards, and cultivating interdisciplinary talents, providing references for promoting the application of AI in smart city transportation.

Abstract Information processing in the big data era is often likened to ‘drinking from a fire hose'—A metaphor that underscores the inherent inadequacy of traditional approaches. The synergy between AI and Electronic Information Processing Technology (EIPT) emerges as the essential ‘barrel' to contain this deluge. Over an 18-month interdisciplinary study, our team found that the integration of AI and EIPT transcends mere technical superposition; it acts as a core engine redefining the very logic of data processing.This paper systematically examines their theoretical foundations and strategic value, evaluates application models in key sectors like healthcare, transportation, and smart homes, and proposes an empirical development framework addressing core challenges including data sovereignty, standardization gaps, and interdisciplinary talent shortages. We hope this research will provide practical references for advancing the "real-world implementation" of AI-EIPT synergy.Through our collaboration with engineers at Intel and clinicians at Johns Hopkins, we've developed a framework that doesn't just sound good on paper—It's already being tested in smart factories and rural hospitals. Our goal? To build systems that aren't just efficient,but human-centered—Tools that work for people, not against them.

Abstract This paper addresses the common issues of "emphasizing skills over thinking and lacking standards"in current digital media professional courses, proposing a new approach to constructing a three-dimensional curriculum system comprising "AI tool operation + creative logic + ethical standards."This system is based on mastering AI tools, emphasizing the enhancement of students' innovative abilities through the guidance of creative thinking logic, and using ethical standards as a value-driven framework to form a progressive curriculum content that integrates skills, thinking, and values. In terms of tool operation, the curriculum not only teaches the basic functions of mainstream AI tools but also focuses on advanced techniques and the exploration of cutting-edge industry technologies. In terms of creative logic, the curriculum utilizes prompt engineering, creative methodologies, and interdisciplinary integration to help students enhance their generative creation and innovative design capabilities. In terms of ethical norms, the curriculum employs case-based teaching, scenario simulations, and specialized discussions to guide students in developing proper professional ethics and a sense of social responsibility. This paper also combines teaching practice to introduce the specific implementation pathways of the three-dimensional curriculum system in terms of teaching methods, evaluation mechanisms, and faculty collaboration, demonstrating its significant in enhancing students' comprehensive literacy, innovative thinking, and professional literacy. Research indicates that this curriculum system not only addresses the shortcomings of existing digital media education but also provides a practical reference model for professional education reform and industry talent cultivation.

Abstract As virtual reality (VR) technology continues to evolve, its applications in the field of education are increasingly gaining attention. This paper aims to explore the application of VR technology in the reform of practical teaching in the Digital Media Technology program. By analyzing the current state and needs of practical teaching in the Digital Media Technology program, this paper elucidates the application scenarios of VR technology in practical teaching within this program. Through case studies, the feasibility and effectiveness of VR technology in teaching are validated. The research findings indicate that VR technology can effectively address issues such as limited traditional practical teaching resources and low student engagement, providing new insights and pathways for teaching reform in the Digital Media Technology program. This approach helps enhance students' practical skills and innovative capabilities, fostering the development of high-quality talent capable of adapting to the evolving needs of the digital media industry.

Abstract With the rapid development of digital technology, network information security has become a core issue affecting social operation and enterprise development. Traditional network security protection methods have shortcomings such as poor flexibility and high deployment costs, making it difficult for them to adapt to the dynamic and complex modern network environment. As a new type of network architecture technology, Virtual Network Technology (VNT) possesses characteristics including resource virtualization, flexible deployment, and isolated operation, providing a new solution for network information security protection. This paper expounds on the core technologies of VNT and focuses on exploring its application scenarios in security protection, such as remote access security, network isolation, and attack defense. Research shows that VNT can effectively improve the flexibility, scalability, and security of network systems, and has broad application prospects in the field of network information security.

Abstract As a core course for graduate students in fields such as aerospace and energy power, Computational Fluid Dynamics (CFD) presents challenges for traditional teaching models due to its abstract theory, complex practice, and delayed feedback. These models struggle to meet the requirements of personalized parenting and industry demands. To address these issue, this paper systematically elaborates the implementation path of AI in theoretical teaching, practical training, and application expansion. It boosts students' efficiency in mastering core CFD knowledge, enhances programmatic and emulation practice competency, and provides a reference for teaching reform in graduate courses in science and engineering.

Abstract As an important resource bearing the development history and core information of the university, the digital transformation of university archives is a key link in the construction of smart campus, which is of great significance for improving management efficiency, promoting resource sharing and empowering teaching and scientific research. By expounding the connotation and significance of the digital transformation of university archives, this paper deeply analyzes the problems existing in the current digital transformation of university archives, and puts forward the path to promote the digital transformation of university archives from the dimensions of perfecting the top management mechanism, building the safety protection barrier of the whole process, and improving the efficiency of archives utilization, so as to provide reference for promoting the leap from "entity management" to "data management" of university archives management and help fully release the value of archives in school construction and social development.

Abstract To deal with the variety of appliances, complex usage patterns, and hidden abnormal events in dormitory electricity use, we propose an anomaly detection method based on multi-source high-frequency feature fusion. We collect data from multiple power circuits, environmental sensors, and high-frequency voltage/current waveforms, then combine them into a multi-dimensional time-series feature set that mixes both low-frequency statistical data and high-frequency dynamic details. In the modeling stage, we use an LSTM autoencoder to learn the normal usage patterns over time, and detect anomalies by checking the reconstruction errors. Tests show that this approach works well for pinpointing sudden power changes, describing appliance behaviors, and spotting abnormal loads. Compared with single-variable detection methods, our multi-source fusion greatly boosts detection accuracy, helps find safety risks earlier, and supports smarter dormitory energy management and protection.

Abstract With the acceleration of urbanization and the frequent occurrence of extreme weather events, traffic safety hazards and urban waterlogging caused by road waterlogging are becoming increasingly prominent. Traditional detection methods mainly rely on manual inspections or physical sensors, which have problems such as low efficiency, high cost, and poor real-time performance. Based on the YOLOv8 target detection framework, this paper proposes a systematic optimization solution for the challenges of small target missed detection, complex background interference, and multi-scale morphological changes that are common in road waterlogging scenes. In terms of model construction, this paper improves the CSPDarknet53 backbone network, and enhances the model's ability to extract multi-scale features by optimizing the cross-stage partial connection (CSP) structure and the spatial pyramid pooling (SPPF) module; in the neck network, the PANet bidirectional feature pyramid structure is introduced and the dynamic convolution mechanism is integrated, which effectively improves the fusion efficiency of shallow details and deep semantic information; the detection head adopts the Anchor-free design, directly regresses the target center point and width and height parameters, simplifies the model structure, and improves the positioning accuracy of irregular water accumulation areas. To verify the performance of the model, this paper constructs a road waterlogging dataset covering different lighting, rainfall intensities and waterlogging forms, and uses enhancement strategies such as random rotation and reflection simulation to improve generalization. The experimental results show that the optimized YOLOv8 model can achieve real-time detection speed and exhibit high segmentation accuracy and robustness.

Abstract The stability and reliability of electrical power systems are critically dependent on effective protection schemes. The transmission system, in particular, is the most frequent site of electrical incidents due to the high voltages and complex environmental conditions it endures. Surges on transmission lines often lead to severe consequences, including equipment failure, corona discharge, and insulation breakdown. This report aims to analyze two primary adverse phenomena on transmission lines: internal switching transients and external lightning surges. By simulating their practical occurrences, we analyze their characteristics and impacts. Furthermore, we test the performance of a core protective device—The surge arrester—And propose optimization strategies, providing a reference for future solutions.

Abstract Traditional production scheduling algorithms for mechanical assembly workshops often struggle with slow responses to dynamic changes and inaccurate decision-making. To tackle this, we designed a dynamic scheduling algorithm that integrates machine vision inspection data—enabling real-time perception of on-site status (e.g., workpiece assembly deviations, equipment abnormalities) and swift adaptation to unexpected events (e.g., equipment failures, non-conforming products entering the line).This paper provides a detailed exposition of the implementation process of the algorithm and validates its effectiveness through experiments: compared with traditional methods, the production cycle is shortened by 15%, equipment utilization is improved by 10%, and product quality is significantly enhanced. The research results offer a practical and feasible path for intelligent scheduling in mechanical assembly workshops.

Abstract As digitalization and intelligence gain increasing prominence, numerous studies have investigated their influence on green innovation and related domains. This paper conducts a comprehensive literature review along horizontal and vertical dimensions. The vertical analysis distinguishes between early and recent research, while the horizontal examination focuses on four key aspects: grounded theories, measurement methodologies for digitalization and intelligence, underlying mechanisms, and innovative approaches. Several critical gaps are identified: the lack of comprehensive theoretical foundations leading to uncertainty in theoretical frameworks (whether linear or U-shaped relationships), oversimplified quantitative methodologies (primarily relying on keyword screening in strategy data), and limited innovative perspectives (predominantly mechanism-focused). Through the novel lens of signal transmission theory, this study proposes an 'Intelligence-ESG-green innovation' theoretical framework. The research contributes new evidence toward resolving theoretical framework uncertainties regarding the relationship between digitalization (intelligence) and green innovation.

Abstract Human emotions, reflected through facial expressions, provide valuable insight into an individual's state of mind. Automatic facial emotion recognition (FER) has made significant progress. However, the recognizing emotions on occluded faces remains an area that requires further exploration. We propose a method to address the challenges of partial occlusion, that is, to recognize emotions on faces where the lower part is obscured by a mask. We tackle the classical FER task by employing a transfer learning approach: first, training a teacher model on a dataset of unmasked faces, and then fine-tuning a student model with the learned weights on a dataset of masked faces. Our model integrates a combination of convolutional layers, Inception blocks, Residual blocks, and Squeeze-and-Excitation (SE) networks. We evaluate the proposed approach on three datasets: JAFFE, KDEF, and Oulu-CASIA, achieving accuracy rates of 80.00%, 78.57%, and 89.38%, respectively.

Abstract With the advancement of the "Belt and Road"initiative, the internationalization degree of Yunnan's tourism is continuously increasing, with 5.8 million inbound tourists received in 2019. This study takes the public display language of Yunnan's tourist attractions as the research object and systematically compares the performance of neural machine translation (NMT) and human translation through a combination of quantitative analysis and qualitative assessment. The study reveals the limitations of machine translation in processing culturally sensitive texts and proposes an "intelligent + human"collaborative translation model along with algorithm optimization suggestions to provide practical references for the field of cultural tourism translation.

Abstract The rapid integration of artificial intelligence (AI) in media production has significantly transformed content creation and consumption patterns. In the context of media convergence, AI-generated content (AIGC) plays an increasingly influential role in shaping user cognition, attitudes, and behaviors. This paper explores the cognitive impact mechanisms of AIGC on users, focusing on perceived authenticity, source credibility, cognitive overload, and selective exposure. By adopting a mixed-methods approach that includes a quantitative survey and qualitative interviews, we examine how different types of AIGC (e.g., news, marketing, entertainment) affect user trust, memory retention, and information processing. The findings reveal that while AIGC enhances engagement through personalization and relevance, it also raises challenges related to misinformation, echo chambers, and cognitive fatigue. This research provides insights for media practitioners, AI developers, and policymakers on fostering ethical AI-mediated communication and improving digital literacy in the age of converged media.

Abstract To address the urgent need for high-level compound talents in water conservancy industry's tech innovation and industrial upgrading under new engineering, and solve issues in master's training (unstable interdisciplinary concepts, supervisors' single research directions,curriculum-social demand disconnection), this study develops a "system-platform-faculty-method" four-in-one multi-disciplinary integration talent cultivation model.The model integrates resources and deepens industry-education integration by improving compound talent-oriented multi-disciplinary training systems(increasing cross-disciplinary courses), optimizing collaborative education mechanisms, strengthening recruitment/development of multi-disciplinary teachers and grassroots teaching organizations,innovating teaching methods, and enhancing industry-education integration.Aiming to break disciplinary barriers, it boosts students' cross-disciplinary knowledge integration, innovative practical ability and industry adaptability, meeting national strategic needs and industry transformation.Conclusion:The model has significant theoretical and practical value in cultivating new water conservancy talents (tackling complex challenges, leading tech innovation), improving training quality and service capacity, and advancing new engineering in water conservancy. Future efforts require continuous optimization and strengthened policy-resource support.

Abstract Aiming at the problems of high model computation cost and easy feature loss during lightweighting in small target detection task, this paper proposes a compact feature migration method based on YOLOv7, which reduces the number of model parameters by combining with knowledge distillation technique and can maintain relatively good detection effect. The method reduces the number of model parameters by designing a lightweight student network through channel pruning; then dynamically assigns spatial weights to enhance the feature alignment in small target regions through the Adaptive Feature Distillation (AFD) module; and finally, further compresses the model by combining structured pruning with quantized perceptual training. The experimental results show that this paper's method decreases the model mAP by only 1.8% with 35% reduction in the amount of parameters and 53% reduction in the amount of computation. It can be seen that the method provides an effective lightweight solution for small target detection, which has certain practical significance.

Abstract Palmprint recognition is an authentication technology based on human biometrics, which has gained wide attention in the biometrics field due to its uniqueness, stability and non-invasiveness. And now it is widely used in the fields of financial payment, security monitoring, and intelligent medical treatment. Furthermore, this technology has already been widely adopted in China's daily life, particularly in mobile payment systems. This study proposes an innovative palmprint recognition method that combines data preprocessing, attention mechanism, and migration learning to optimize feature extraction and classification performance. The results show that the method in this paper outperforms traditional methods on multiple datasets, exhibiting high robustness and adaptability. It provides some ideas for future palmprint research.