SUN Yan , SHI Zhe , ZHANG Shujuan , ZHANG Jiqing , WANG Jiao
2025, 41(19):1-3.
Abstract:Against the backdrop of vocational education that emphasizing the equal importance of“cultivating morality and talent”and skill development,the traditional teaching mode of vocational school“graphic design” course urgently needs to be reformed due to insufficient integration of ideological and political education and weak cultivation of students? comprehensive literacy. This article proposes an innovative teaching model of“ideological and political guidance—case driven—collaborative design”,which deeply integrates curriculum ideological and political education,uses practical cases as teaching drivers,and organizes students to carry out collaborative design projects. By delving into the ideological and political elements in the curriculum,naturally integrating them into case analysis and project practice,guiding students to establish correct values. By analyzing typical cases, stimulate students? interest and innovative thinking,and enhance their ability to solve practical problems. Through group collaboration design,strengthen students? teamwork awareness and communication skills. Teaching practice has shown that this model can effectively achieve resonance between ideological and political education and professional teaching,and significantly enhance students? professional skills,innovation ability,and social responsibility,and provides a feasible path and useful reference for the reform of vocational graphic design teaching.
2025, 41(19):4-6.
Abstract:At present,there is a significant disconnection between the teaching content of the information security major in higher vocational colleges and the actual demands of enterprises,which leads to the knowledge and skills that students master being difficult to meet the operational requirements of real enterprise projects. In response to this issue,this article focuses on the cultivation of core competencies for students majoring in information security in higher vocational colleges and conducts an in-depth discussion on the reform strategies of comprehensive practical teaching. Through a series of measures such as building a“three-dimensional progressive”training system,promoting a dual-education model of schools and enterprises,and establishing an intelligent simulation experiment platform,efforts are made to enhance students? core abilities in technology application,career adaptation,and emergency response. The practical results show that the reform has effectively promoted students? performance in skills competitions and employment quality,providing a valuable implementation path for the systematic reform of practical teaching in information security majors in higher vocational colleges.
2025, 41(19):7-9.
Abstract:This article focuses on the curriculum reform of big data and accounting majors. Based on the new demands for data analysis capabilities in the accounting industry under the background of digital transformation,it systematically analyzes the limitations of the traditional curriculum system in terms of content,methods,and practical links,and proposes a multi-dimensional reform path centered on the reconstruction of the curriculum system,the innovation of teaching methods,and the construction of practical platforms. Based on practical cases from domestic universities and feedback from enterprises,this paper demonstrates the remarkable achievements of the reform in enhancing students? data literacy and employment competitiveness. It further looks forward to the future development direction,providing theoretical references and practical guidance for cultivating high-quality accounting talents who can adapt to the era of intelligent finance.
2025, 41(19):10-12.
Abstract:Carrying out ideological and political education in the curriculum is an important mission entrusted to teachers by the state in the new era,and it is also the core link of“foster virtue through education”. In response to the problems existing in the implementation of ideological and political education in the course of“university computer fundamentals”in pharmaceutical vocational colleges,this study explores teaching from the dimensions of teacher literacy,teaching methods,teaching design,and teaching evaluation,and proposes specific measures to carry out ideological and political education construction in the course,providing reference for pharmaceutical vocational teachers to implement ideological and political education in computer general courses.
FU Sisi , GE Liang , WANG Qian , LI Jia
2025, 41(19):13-15.
Abstract:With the rapid development of new generation AI technologies represented by GPT and DeepSeek,the teaching of computer courses in universities is undergoing unprecedented changes. The new teaching mode and teaching method urgently need a supporting teaching evaluation system to achieve the integrated development of teaching and evaluation. The article summarizes the existing research on teaching evaluation of computer courses from three aspects:teaching evaluation index system,evaluation methods,and application of evaluation results. The aim is to promote the development of teaching evaluation of computer courses under the background of AI and provide reference for cultivating talents who can adapt to the social needs of the intelligent era.
2025, 41(19):16-19.
Abstract:This article focuses on vocational computer majors and constructs a“trinity”data analysis framework, including job demand data collection based on crawler technology,learning behavior tracking based on learning management systems,and employment feedback analysis based on knowledge graphs. In data preprocessing, methods such as MICE multiple interpolation and LOF anomaly detection are applied. In the data analysis stage, Apriori association rule mining,HDBSCAN clustering,and XGBoost prediction modeling are combined to conduct in-depth analysis on the matching degree between course structure and industry demand,the correlation between teaching methods and learning outcomes,and the correlation between evaluation system and employment quality. Based on this,an optimized path for talent cultivation in vocational computer majors is proposed,providing datadriven decision-making basis for teaching design and talent cultivation.
ZHU Hongxia , LI Muxiang , MENG Ran
2025, 41(19):20-22.
Abstract:Under the policy guidance of the National Implementation Plan for Vocational Education Reform and the“Double High Plan”,the integration of industry and education has been elevated to the strategic level of national education development. Achieving this integration hinges on cultivating students? innovative thinking and capabilities,enabling them to creatively apply acquired knowledge to industrial practices,thereby enhancing industrial production efficiency and contributing to the modernization of China ?s industry. In addition to theoretical instruction,vocational colleges urgently need to explore diverse educational approaches. This article proposes using innovation,creativity,and entrepreneurship competitions(the“Three C”competition)as a platform to effectively foster students? innovative thinking and capabilities through practice-oriented contest formats,aiming to provide a reference for innovating talent cultivation models in vocational institutions.
WANG Changhai , HUANG Min , HUANG Wanwei
2025, 41(19):23-25.
Abstract:This article first analyzes the main pain points in the teaching of“deep learning”courses in universities,and proposes targeted principles for project driven curriculum system design. Secondly,a teaching syllabus and class hour allocation plan for the“deep learning”course were constructed,dividing the course content into two parts:basic knowledge and domain applications. Typical cases were provided based on different application fields to balance theoretical teaching and practical operation. Finally,a teaching tracking and evaluation method was designed. The research results provide new ideas for the teaching of“deep learning” courses,which have positive significance for improving students? learning effectiveness and cultivating their innovative thinking.
2025, 41(19):26-28.
Abstract:In the context of“Emerging Engineering Education”,there is a structural contradiction between PLC course teaching and the demand for intelligent manufacturing talents. This article proposes a three in one reform framework of“target reconstruction—platform innovation—project iteration”. By deconstructing the composite capability requirements in industrial scenarios,constructing a fourth-order capability development model,and designing a progressive project system that runs through industrial standards. At the same time,a spiral teaching improvement method was proposed,combined with a dual channel verification mechanism of qualitative analysis and quantitative modeling,providing a methodological reference for engineering practice courses to break through the traditional dilemma of“cognitive disconnection—equipment limitations—single evaluation”.
HU Yating , WANG Wei , LI Zhuoshi , SUN Huiying
2025, 41(19):29-31.
Abstract:Interdisciplinary integration has become an important direction for higher education and scientific research innovation,which puts forward new requirements for the talent cultivation mode of information technology majors. This article explores the construction of curriculum system,integration of teaching staff,establishment of research platforms,and reconstruction of training and evaluation mechanisms,aiming to promote the deep integration of information technology with other disciplines,form a new pattern of interdisciplinary collaborative education,and provide useful references for universities to optimize the training mode of information technology talents.
DUAN Yu , WANG Wanwan , YANG Xun , YAO Jin , MAO Kunpeng
2025, 41(19):32-34.
Abstract:In the context of digitalization,universities urgently need to clarify the fundamental issues of“what kind of people to cultivate,how to cultivate people,and for whom to cultivate people”,and make“cultivating virtue and nurturing people”the core task. Course ideological and political education can not only cultivate students? innovative spirit and creativity,but also enable them to become determined artificial intelligence professionals. Taking the course“introduction to artificial intelligence”as an example,this article explores the entry point of ideological and political education in the course and the implementation path of integrating it into the classroom. The practical results show that the constructed teaching design model and method are feasible and effective,and can effectively achieve the educational goals of ideological and political education in the curriculum.
SUN Yingchun , LIANG Yeling , WANG Junfang
2025, 41(19):35-37.
Abstract:With the rapid development of information technology,educational informatization has become an important means to promote the high-quality development of education. In response to the challenges faced by private universities in sharing high-quality educational resources under the background of big data,this study adopts literature research and case analysis methods to deeply analyze the current situation and challenges faced by private universities in sharing educational resources. Based on big data technology,a shared platform model with a four-layer architecture of“data collection data processing intelligent application user interaction”is constructed, and optimization strategies are proposed from two levels:policy system and technology platform.
2025, 41(19):38-40.
Abstract:In response to the shortcomings of traditional teaching models in secondary vocational computer education,this article explores the application of project-based teaching models in depth. By analyzing the connotation and characteristics of project-based teaching mode,combined with the current situation of computer teaching in vocational schools,a complete application system including project design,implementation, evaluation and other aspects is constructed. Practice shows that this model can effectively enhance students? computer skills,problem-solving abilities,and teamwork spirit,providing useful references for the reform of computer teaching in secondary vocational schools.
2025, 41(19):41-43.
Abstract:Taking medical colleges in Hebei province as the research object,this article explores the impact mechanism of digital intelligence technology on the employment structure of medical students. By analyzing the contradiction between the expansion of primary healthcare demand,the rise of emerging medical fields,and the intensification of competition among tertiary hospitals,combined with employment data from universities such as Hebei Medical University,a“three-dimensional driving”transformation path is proposed:constructing a digitalized employment guidance system,deepening the cultivation of“medicine+X”composite talents,and building a government-school-enterprise collaborative employment platform. Research finds that digital intelligence technology provides innovative solutions to the dual dilemma of“grassroots sinking difficulty”and“high-end breakthrough difficulty”for medical students in Hebei province by reshaping the employment information ecosystem and optimizing the efficiency of job matching.
CHEN Huan , LIU Tao , ZHANG Ping
2025, 41(19):44-46.
Abstract:In the context of the new era,IoT technology has become an important driving force for promoting economic and social development. Based on the perspective of great craftsmen,this article explores the integration path of ideological and political education in the course of“Internet of Things technology and application”,and constructs an educational system centered on“craftsmanship,craftsmanship soul,craftsmanship emotion,and craftsmanship”. Through project driven and“four stage scenario based”teaching mode,optimize course content and evaluation mechanism,integrate the craftsmanship spirit of“dedication,lean,focus,and innovation”into IoT engineering practice,and form a curriculum evaluation standard with craftsmanship characteristics,aiming to guide students to learn and reflect based on practice,internalize and externalize the craftsmanship spirit,and become high-quality craftsmanship talents that meet the needs of the new era.
2025, 41(19):47-51.
Abstract:In the context of the deep integration of digital humanities technology and ancient book research,the teaching of digital humanities courses still faces many challenges,such as the fragmentation of ancient book digital materials,the fragmentation of platforms and systems,the insufficient ability of lower grade students in interdisciplinary knowledge integration,and the general lack of computer technology foundation. We explored and analyzed the new roles and problems of the artificial intelligence model DeepSeek in the research of ancient books for humanities and social science students. DeepSeek has the characteristic of conversational teaching,and its core mechanism is to integrate heterogeneous resources through semantic retrieval and knowledge association technology. In teaching practice,it can assist students in text analysis,data visualization,and computer programming,and is valuable for in improving the efficiency of ancient literature processing. However,due to the linguistic features of ancient Chinese,DeepSeek may still have deviations in semantic parsing,and may also lead to cognitive dependence among students,accompanied by risks such as AI hallucinations.
2025, 41(19):52-55.
Abstract:By comparing and analyzing the two teaching philosophies of“student-centered”and“teachercentered”,the theoretical basis of the“student-centered”teaching philosophy is summarized. Further exploration is conducted on the principles and ideas of instructional design under this concept, and the specific implementation process of the“student-centered”concept in practical teaching is explained using the course “Python programming fundamentals”as an example.
YE Jing , CHEN Yanyun , DUAN Zhimin
2025, 41(19):56-58.
Abstract:With the rapid development of artificial intelligence technology,its application in higher education is becoming increasingly widespread,especially demonstrating significant value in the practical teaching of computer science. Based on the current curriculum system of practical teaching in computer science and the main existing problems,this article systematically analyzes the typical application methods of artificial intelligence technology in teaching, including the construction of intelligent experimental platforms, personalized learning path recommendations, virtual simulation-assisted teaching, and the design and implementation of intelligent assessment systems. Research shows that the introduction of artificial intelligence technology can effectively enhance teaching efficiency,enrich teaching methods,and significantly improve students? comprehensive abilities,which has a positive promoting effect on optimizing the practical teaching system of computer science.
CAO Qiang , LIU Chang , TAO Jiayuan , WU Deming
2025, 41(19):59-64.
Abstract:The level of automation is an important indicator to measure the progressiveness of a country ?s productivity. The close integration of advanced control theory and control technology not only stimulates the development of new quality productivity,but also promotes the implementation of new application scenarios. Based on the OBE concept,a construction mode diagram for the course“principles of automatic control”is designed, and course modules,content,and ability matrices are constructed. The construction concept of vocational undergraduate textbooks is refined to promote the construction of high-quality professional course resources. By drawing on the case model of classic textbooks,it strengthens the knowledge carrying function of textbooks. By combining point and surface approaches,it effectively promotes the ideological and political construction of professional courses,and deeply explores the construction connotations of the“gold majors”in automation,the “gold courses”in automatic control theory,and the“gold textbooks”in vocational undergraduate education. Exploring the implementation mode and practical path of ideological and political education in courses to assist in the cultivation of high-end skilled talents for vocational undergraduate programs.
2025, 41(19):65-67.
Abstract:The rapid development of AI technology has raised new demands on vocational education,emphasizing the cultivation of compound,innovative,and practice-oriented talents. As a core policy of higher vocational education reform,the“Double High Plan”promotes the transformation of talent training from single-skill training to comprehensive ability development through deep industry-education integration and interdisciplinary collaboration. Based on the interaction between industrial demand and educational supply,this article explores practical pathways for curriculum restructuring,teaching innovation,and evaluation mechanism optimization in higher vocational institutions under the AI era. The results indicate that the joint construction of training bases by universities and enterprises,interdisciplinary curriculum integration,and project-based teaching practices are key measures for advancing the transformation of talent training models,providing both theoretical support and practical paradigms for the high-quality development of vocational education.
2025, 41(19):68-70.
Abstract:This article proposes a personalized teaching model that integrates intelligent content recommendation, adaptive learning management,and intelligent Q&A evaluation to address the problem of difficulty in achieving personalized education in the“computer network”course. The teaching experiment results show that compared with traditional methods,this model effectively enhances students? learning interest and participation,with an average score increase of 18.7% and a learning satisfaction increase of 26.3%. This verifies the application value of artificial intelligence technology in enhancing the teaching effectiveness of“computer network”courses and achieving personalized training.
FENG Qijing , ZHANG Jinxue , FAN Tairan , WANG Shiqi , DU Wenxuan
2025, 41(19):71-73.
Abstract:The experimental course of“analog electronic technology”is an important practical foundation for electronic information majors. In response to the problems of passive learning,insufficient practical ability,and single assessment methods in traditional experimental teaching,this article optimizes the teaching design of the “analog electronic technology”experimental course based on the Outcome Based Education(OBE)concept and blended learning mode,aiming to strengthen students? self-learning ability. Based on specific teaching cases,the application effect and challenges faced by this model in curriculum implementation were analyzed,and corresponding improvement strategies were proposed. The results indicate that blended learning based on OBE can effectively enhance students? learning initiative and experimental skills,which has a positive significance for improving teaching quality.
2025, 41(19):74-77.
Abstract:In the context of teaching innovation and reform,this article explores the implementation of the “precision teaching”strategy in courses supported by knowledge graphs from the perspective of student learning profiles. Firstly,starting from the current teaching situation,the significance of implementing“precision teaching” in meeting students? personalized needs,improving teaching quality,and promoting professional development is elaborated. Secondly,the specific path of empowering teaching reform with knowledge graphs was systematically analyzed in terms of analyzing learning situations,reconstructing teaching content,innovating teaching methods, and evaluating and providing feedback on teaching. Combined with the teaching case of the industrial Internet data collection project,the teaching implementation process is described in detail. Finally,it summarizes the problems existing in the current teaching and points out the direction of future improvement,aiming to provide theoretical and practical reference for the implementation of“precision teaching”in the industrial Internet specialty of higher vocational education.
2025, 41(19):78-80.
Abstract:A tomato leaf disease detection method based on an improved YOLOv11 model is proposed to address the problems of low accuracy and efficiency in manual recognition of tomato leaf diseases in the current agricultural field. Firstly,a correlation analysis is conducted on a dataset containing images of 9 common diseases,including tomato leaf spot disease,early blight disease,and leaf mold disease. On this basis,ASF(Attention Scale Fusion) technology is integrated into the original YOLOv11 structure to construct the ASF-YOLO model. ASF-YOLO,as a novel model based on YOLO,is initially used for cell instance segmentation. Its core lies in introducing the ASF mechanism into the YOLO framework,thereby improving image detection and segmentation performance. The model structure includes the Scale Sequence Feature Fusion(SSFF)module,Triple Feature Encoder(TFE) module,and Channel and Position Attention Mechanism(CPAM). The collaborative effect of multiple modules effectively improves detection accuracy. The experimental results show that the improved YOLOv11 model has better performance in processing diverse tomato leaf disease images,with recall,mAP50,and mAP50-95 reaching 96.0%,98.9%,and 91.2%,respectively. This method not only provides a feasible technical path for the automated detection of tomato leaf diseases,but also provides a reference for the detection and research of other crop diseases,which is of great significance for promoting sustainable agricultural development.
YAO Yingzhe , YANG Pengcheng , CHANG Xuechuan
2025, 41(19):81-83.
Abstract:Aiming at the problems such as single data source,insufficient feature extraction,and inadequate dynamic adaptability of the model in the construction of user portraits,this article proposes a user portrait modeling framework based on multi-source data fusion. By integrating users? basic information,behavioral data, social network data and third-party credit data,a tag system covering dimensions such as population attributes, credit characteristics,consumption preferences and social influence is constructed. The improved random forest algorithm is adopted for feature selection,combined with the time decay factor to optimize the calculation of label weights,and a dynamic weight update mechanism is proposed. The experimental results show that this model significantly improving in indicators such as the accuracy of user credit assessment and the recall rate of high-risk user identification compared with traditional models,providing strong support for precise marketing and risk control in consumer finance.
2025, 41(19):84-86.
Abstract:A framework for automatic classification and retrieval of employee health records based on intelligent algorithms is proposed,which integrates natural language processing,deep learning,and knowledge graph technology to achieve structured processing and semantic retrieval of multimodal health data. Experimental results shows that proposed method achieves a disease classification accuracy of 92.7%,reduces retrieval response time to 0.3 seconds,and improves efficiency by 80% compared to traditional methods.
ZHANG You , WANG Xianling , SUN Xiaoli , YANG Jun , TENG Shuai
2025, 41(19):87-89.
Abstract:Artificial neural networks can directly learn feature information from raw data. When a structure is damaged,the response data often contains characteristic information related to the damage. This article proposes a structural damage identification method based on artificial neural networks. Taking simply supported beams as the research object,the free vibration response is simulated by finite element method,and the first-order modal strain energy is extracted as the characteristic index. Batch automatic calculation is used to generate sample data covering multiple damage scenarios,thereby training the neural network. The verification results show that this method can not only accurately identify the location of structural damage,but also make high-precision judgments on the degree of damage,verifying the potential application of artificial neural networks in structural health monitoring.
2025, 41(19):90-94.
Abstract:Ant colony algorithm,as an optimization method based on swarm intelligence,is widely used in path planning problems due to its strong anti-interference ability,distributed computing capability,and wide applicability. However,there are also phenomena such as being easily trapped in local optima,low search efficiency,and the possibility of deadlock or stagnation. In response to these issues,the article proposes a goal guided improved ant colony algorithm from the perspectives of heuristic function design, anti deadlock mechanism,state transition probability calculation,and pheromone update. Firstly,improving the distance heuristic function of traditional ant colony algorithm through adaptive target guidance mechanism. Secondly,in order to improve the search probability of effective solutions and reduce the generation of invalid solutions,a dynamically adjusted deadlock avoidance mechanism is introduced. Furthermore,considering the impact of obstacles and potential threats,a state transition probability model based on security awareness was designed,and a dual path pheromone update strategy was introduced in the pheromone update process to reduce the excessive accumulation of pheromones on specific nodes or paths. Finally,the algorithm is applied to the path planning problem of unmanned aerial vehicles in a three-dimensional environment. The experimental results show that compared with traditional ant colony algorithm (ACO),ant colony A* algorithm (ACO-A*),and genetic algorithm(GA),this method reduces the total length of optimized paths by more than 7%,and improves path planning efficiency by more than 60%,demonstrating significant advantages.
2025, 41(19):95-97.
Abstract:A welding defect detection method based on improved RT-DETR is proposed to improve the accuracy and efficiency of the welding defect detection process,reduce missed and false detections. This method has been optimized based on the RT-DETR model:on the one hand,the ADown downsampling module is introduced to compress the feature map size while preserving key information;On the other hand,the WTConv convolution module is introduced to integrate multi-scale frequency domain and spatial features,thereby enhancing the models feature extraction ability for multi-scale targets and complex backgrounds. The experimental results show that after adding these two modules,the model mAP@50 indicator has increased from 67.6% of the original RT-DETR to 70.1%,while the parameter count has decreased from 19 885 212 to 18 623 772. While reducing computational complexity,the detection accuracy has been improved,significantly enhancing the efficiency of defect detection in welded components.
ZHAO Liang , LU Wenhao , PEI Zhenghua , ZHAO Longpeng
2025, 41(19):98-100.
Abstract:Based on the advantages of the transient electromagnetic method in mineral resource exploration,the application characteristics and data response patterns of this method in different types of ore bodies were analyzed. The adaptability and practical effects of the transient electromagnetic method in the exploration of metallic minerals,coal resources,as well as oil,gas and salt minerals were studied. The application results show that this method has significant guiding significance and broad application potential in aspects such as resource potential prediction,hidden ore body location,and mine disaster prevention and control.
2025, 41(19):101-104.
Abstract:The traditional neural network architecture search methods based on latent space usually have the architecture representation distribution in the latent space fixed after the pre-training is completed. AG-Net proposes a sustainably updated search method for the architecture of latent space neural networks,which alleviates this problem to a certain extent. However,the single-implicit space structure adopted by AG-Net is prone to causing the search process to fall into local optimal solutions. To address this limitation,the article proposes a novel neural network architecture search method based on multi-subhidden Spaces and dual predictors. This method effectively enhances the exploration ability of the search process by introducing multiple sub-latent space structures and avoids getting trapped in local optima. Meanwhile,by leveraging the dual predictor mechanism to integrate local and global information,the comprehensive performance of the search algorithm is further enhanced.
2025, 41(19):105-107.
Abstract:The article proposes a multi-source structured information extraction method for physical examination reports that integrates optical character recognition(OCR)and natural language processing(NLP)technologies. This method first uses PaddleOCR to complete the layout analysis,text detection and recognition of the report image,and then uses the PP-ChatOCRv3-doc model to handle complex table structures. By integrating named entity recognition,relation extraction and LayoutLM model in NLP,the alignment and association of cross-line text fields are optimized,effectively solving the misalignment and fragmentation problems that are prone to occur in traditional OCR in cross-line attribute recognition. The experimental results show that the accuracy rate of this method in extracting key medical indicators from physical examination reports reaches 96.7%,and its performance is improved by 21.3% compared with the single OCR scheme,providing reliable technical support for the automatic processing,standardization and intelligent analysis of medical data.
2025, 41(19):108-110.
Abstract:Unmanned aerial vehicle(UAV)remote sensing technology has significant application value in fields such as environmental monitoring,urban planning and agricultural management. The article proposes a real-time analysis method for UAV remote sensing data based on point cloud data processing,with a focus on key technologies such as UAV data acquisition,point cloud preprocessing,ground object classification,and 3D modeling. Through a case analysis of the vegetation coverage in a certain mountainous area,the feasibility and effectiveness of this method were verified. The experimental results show that the application of UAV LiDAR technology can efficiently obtain high-precision three-dimensional data. Combined with point cloud processing algorithms,it can achieve precise monitoring of the spatial distribution and height features of vegetation,providing effective technical support for the further application and development of UAV remote sensing technology.
2025, 41(19):111-114.
Abstract:Image blur set is an important tool for dealing with uncertainty and fuzziness. It not only considers positive membership,neutral membership,and non membership,but also introduces rejection membership. In recent years,how to conduct research on decision-making methods in image blur environments has always been a hot topic in the academic community. Evidence theory,as a decision-making model that combines multiple sources of evidence information,can more accurately express incomplete information and its subjective uncertainty, making it a relatively reliable decision-making tool. Based on this,a method is proposed to convert image blur information into evidence representation,and its effectiveness is verified through numerical examples,providing necessary prerequisites for applying evidence theory to decision-making in image blur environments.
2025, 41(19):115-117.
Abstract:Accurately extracting key information from massive data has become an important research topic in the field of intelligent information processing. A text feature extraction method based on deep learning and natural language processing techniques has been researched and implemented to meet the specific needs of Chinese text processing in a particular field. The main research contents include Chinese word segmentation technology based on BiLSTM-CRF model,named entity recognition method combined with dictionary supervision,entity set expansion scheme based on collaborative Bootstrap algorithm,and key element extraction mechanism integrating statistics,rules and dictionaries.
2025, 41(19):118-120.
Abstract:The spatial collision problem between gas pipelines and other municipal pipelines in urban comprehensive pipe galleries directly affects urban public safety and construction efficiency. Traditional 2D design and collision detection methods have shortcomings such as low efficiency and easy omission of hidden conflicts, while existing BIM algorithms also face challenges of insufficient computational efficiency and dynamic adaptability when dealing with complex pipe gallery environments. A hybrid algorithm model of“hierarchical bounding box dynamic spatial decomposition”is proposed for the special characteristics of gas pipelines:noncollision pipelines are quickly screened through the initial bounding box,and potential conflicts are detected through dynamic grid refinement. The construction error parameters and multi-disciplinary collaborative rule library are integrated to achieve precise and dynamic collision detection.
JIANG Fengyue , ZHAO Chenxi , ZHANG Yintong , SHEN Zhen , WANG Yue
2025, 41(19):121-125.
Abstract:A single traditional time series model often has problems such as insufficient accuracy,short prediction period,and weak generalization ability when predicting pollutant concentrations(such as PM2.5). To this end, the idea of multi model fusion is introduced,combining two traditional models-RBF model and ConvLSTM model, and two optimization strategies are adopted to improve them. Firstly,the grey wolf swarm algorithm(GWO)is used to optimize the central parameters and weights of the RBF model. Secondly,the sliding window method(window_ 2) is used to construct the dataset,and the mean square error (MSE) is optimized through seed-values. Meanwhile,AdaBoost is utilized to enhance the trend prediction ability of the ConvLSTM model. On this basis,the optimized model is weighted and fused with the SARIMA model to construct the SARIMA-RBF-GWO ConvLSTM AdaBoost fusion model. Taking various pollutant data from Nanyang City as the experimental object,after training and testing with different models,it was found that the fusion model performed well in PM2.5 time series prediction,with an MSE of 8.60,MAE of 2.09,RMSE of 3.05,and R2 of 0.97. Compared with the SARIMA-RBFGWO model,MSE decreased by 5.07,MAE decreased by 0.88,RMSE decreased by 0.64,and R2 increased by 41%. The experimental results show that the fusion model is significantly better than single or partial fusion models in predicting PM2.5 concentration,with higher accuracy and stability.
2025, 41(19):126-129.
Abstract:Aiming at the problem of inaccurate location of concrete damage by existing methods,this article proposes a damage detection method for reinforced concrete structures based on the IWOA-LSTM algorithm. This method optimizes the hyperparameters of the Long Short-Term Memory network(LSTM)by improving the Whale Optimization Algorithm(IWOA)to enhance the feature extraction ability for complex damage signals. Firstly, separate the independent source signal from the mixed observation signal,and then calculate the arrival time difference using the unmixed signal to achieve the preliminary location of the damage position. By further comparing the frequency change ratio before and after the injury,the damaged area can be accurately identified to complete the entire damage detection process. The experimental results show that this method can effectively identify the location of concrete damage caused by steel bar corrosion due to chloride ion erosion,and the positioning results are highly consistent with the actual damage situation. The simulation positioning accuracy is relatively high,and the signal fluctuation is small,achieving precise detection of the damaged area and providing a reliable means for comprehensively obtaining damage information of reinforced concrete structures.
2025, 41(19):130-131.
Abstract:A Python based FFmpeg video batch synthesis method is introduced,aiming to efficiently solve the problem of merging a large number of segmented video files. This method utilizes the concat function of FFmpeg and combines it with Python scripts to achieve automated operations,which can quickly complete video synthesis while preserving the original image quality. The experimental results show that this method exhibits high efficiency and stability in processing massive video files,providing a feasible and effective solution for the field of video processing.
GUO Zhiyang , LU Jun , LI Ruifeng , ZHOU Chao
2025, 41(19):132-134.
Abstract:Accurate identification of mature fruits is crucial for automated harvesting systems. The article aims to explore a high-precision and real-time method for identifying mature fruits and vegetables. The existing fruit and vegetable recognition methods have insufficient accuracy in determining the location of fruits and vegetables,and their ability to handle occlusion problems is limited. A real-time fruit and vegetable detection network based on boundary feature learning is proposed to address this issue. Firstly,designing a unique octagonal boundary based on the shape of fruits and vegetables for initial boundary extraction. Secondly,deep snake convolution is used to extract boundary features,and a multi feature boundary refinement module is combined to prioritize key boundary feature channels. At the same time,the aggregation of multi-layer fruit and vegetable boundary features is used to enhance the expression of boundary features. Finally,the experimental results validate the significant improvement of the model in fruit and vegetable recognition performance.
2025, 41(19):135-137.
Abstract:The key capability of adaptive artificial intelligence systems lies in their ability to continuously learn in dynamic environments,including incremental acquisition,updating,and accumulating knowledge. The article focuses on the task of continuous object detection,aiming to solve the problem of model adaptability in dynamic new category scenarios. In terms of methodology,a framework based on pre trained large models is adopted, utilizing its powerful generalization ability as the foundation,and combining LoRA(Low Rank Adaptation)to efficiently fine tune newly added tasks to reduce computational overhead and avoid catastrophic forgetting. To address the issue of inconsistent labels between new and old tasks during incremental learning,a method is proposed to generate pseudo labels from the old model,which are then used as knowledge for distillation in the new model. Consistency constraints are implemented to stabilize the training process and improve detection performance. The experimental results show that this method can effectively adapt to new categories while maintaining the ability to recognize basic categories,and its performance is significantly better than traditional incremental learning strategies.
2025, 41(19):138-140.
Abstract:In response to the problem that traditional control methods in the sedimentation process of mine water treatment are difficult to adapt to complex water quality fluctuations and dynamic working conditions,an intelligent control strategy based on deep reinforcement learning is proposed. By constructing a deep Q-network (DQN)and a multi-agent reinforcement learning framework,combined with key parameters such as suspended solids concentration,flow rate,and flocculant dosage in mine water,dynamic optimization of sedimentation tank sludge discharge cycle and flocculant dosage rate is achieved.
2025, 41(19):141-143.
Abstract:The natural circulation steam generator is a key piece of equipment in a pressurized water reactor nuclear power plant,connecting the primary circuit and the secondary circuit. It is responsible for transferring the heat generated by the primary circuit of the nuclear reactor coolant to the secondary circuit and simultaneously generating steam to drive the steam turbine to do work. The steam generator serves to isolate the medium in the primary and secondary circuits and is an important barrier to ensure the safe operation of nuclear power plants.
2025, 41(19):144-146.
Abstract:In response to the problems of low accuracy and poor timeliness in leak detection of urban gas pipelines,an intelligent detection system based on the improved YOLOv8 algorithm is proposed. By integrating CBAM attention mechanism and deformable large kernel attention(D-LKA Attention),the small target detection capability is optimized,and combined with geographic information system(GIS)to achieve precise leak point localization and multi-level alarm. Experiments have shown that the improved model is effective in gas pipeline scenarios mAP@0.5 Reached 82.3%,an increase of 3.1 percentage points compared to the original YOLOv8,with a detection speed of 38 FPS,meeting real-time requirements.
2025, 41(19):147-149.
Abstract:With the rapid development of Internet technology,the medical industry is gradually achieving digital transformation. A medical information interaction platform based on RESTful API has emerged,aiming to promote the sharing and exchange of medical data and improve the efficiency and quality of medical services. This article discusses the development and application of the platform,with a focus on analyzing aspects such as the platform architecture,API interface design,security policies,and practical applications. The aim is to provide theoretical basis and practical experience for the development of medical information interaction platforms and promote the digitalization process of the medical industry.
XIAO Lie , TANG Yulin , ZENG Xiaowei
2025, 41(19):150-152.
Abstract:Gauss projection is an important foundation for geospatial data conversion,and its accuracy and efficiency in forward and backward calculations directly affect the quality of surveying and mapping results. This article designs and implements a Gauss projection forward and backward calculation program based on the Visual C++platform. By deriving the forward and inverse formulas for coordinates under different ellipsoid models, optimizing the meridian arc length iteration algorithm and adjacent band conversion process,the problem of redundant calculations in traditional methods has been effectively solved. The test results show that the deviation of the programs calculation accuracy from professional software is less than 1 mm,and it can achieve automatic conversion of cross band coordinates and improve the human-computer interaction experience through a graphical interface. The research results can provide high-precision and practical coordinate transformation tools for GIS development and engineering surveying,with certain theoretical value and engineering application prospects.
ZHANG Ziliang , WANG Fengling , GONG Shumei , SUN Jun , CAO Feilong
2025, 41(19):153-155.
Abstract:Against the backdrop of the rapid development of the industrial Internet,mechanical manufacturing enterprises are facing an urgent need for intelligent transformation in their safety production management. To this end,the article designs a security monitoring platform based on industrial Internet. By integrating multi-source data collection,edge computing and deep learning technologies,it realizes real-time monitoring and intelligent early warning of the production process. The platform uses the CNN-LSTM hybrid model to analyze the vibration data of the equipment,and the accuracy rate of fault prediction reaches 93.5%. At the same time,visual recognition technology was adopted to monitor the operation behavior,and the recognition rate of violation operations reached 92.8%. Practical application shows that this platform has effectively helped reduce the enterprise safety accident rate by 67.1%,increased the timely rate of hidden danger handling to 97.6%,and the payback period of investment is only 15 months.
2025, 41(19):156-159.
Abstract:Based on an analysis of the pain points in traditional on-site interview processes and the weaknesses of paperless interview systems with a highly integrated approach,this article proposes the design concept and advantages of a loosely coupled architecture by summarizing the characteristics of personnel examination interviews. It comprehensively constructs the overall design goals and architecture of the system,covering the presentation layer,business logic layer,and data access layer. The design of each functional module is elaborated in detail,including subsystems such as candidate management,interview scheduling,paperless scoring,and interview results. Additionally,introducing the design ideas of the system database and the structure of key tables, and analyzes key technologies in system implementation,such as data interface design and the application of message queues. Finally,the system is tested and evaluated,and the results show that the system can effectively improve interview efficiency,reduce costs,and realize paperless and flexible management of the interview process.
2025, 41(19):160-162.
Abstract:This article investigates the application of Python-based intelligent recommendation systems in ecommerce,drawing on practical cases from platforms such as Meituan Youxuan,Taobao,and JD. com,etc. It analyzes the implementation path and performance optimization strategies of collaborative filtering,matrix factorization,deep learning and other algorithms. By comparing the performance of different algorithms in scenarios such as user cold start,data sparsity,and real-time recommendation,it reveals the key role of tools such as Pandas,NumPy,and TensorFlow in improving recommendation accuracy and system throughput in the Python ecosystem. Experimental results demonstrate that hybrid recommendation models combined with real-time feature updates can increase click through rates by 12%~17%,while matrix factorization optimization based on Cython can reduce computation time by 90%.
2025, 41(19):163-165.
Abstract:In response to the problems of“storage wall”and“power wall”faced by neural network processors under the traditional von Neumann architecture,this article proposes a low-power neural network processor design method based on a storage computing integrated architecture. By using a memristor array to achieve deep integration of storage and computation,combined with key technologies such as dynamic voltage frequency adjustment,near threshold calculation,and mixed precision training,an energy efficiency ratio of 5.3 TOPS/W was achieved in the 28 nm FD-SOI process. Experiments show that the average power consumption of this architecture is only 1.2 W when dealing with ResNet-50 networks, 78% lower than that of traditional architectures,and it supports millisecond level dynamic power management,providing energy efficient solutions for edge computing and IoT devices.
CHEN Dongxin , XU Yangjun , ZENG Qiaodi
2025, 41(19):166-168.
Abstract:By deeply analyzing the characteristics of the main transformer load data and combining the advantages of traditional time series models and deep learning models,a fusion model is constructed. Through experimental verification,the model shows excellent performance in prediction accuracy,real-time performance,and stability. It can provide scientific basis for main transformer load management and assist in the stable operation and optimized scheduling of power systems.
2025, 41(19):169-171.
Abstract:This article proposes an Unmanned Aerial Vehicle(UAV)fault diagnosis and maintenance system based on AI. By integrating deep learning and machine learning algorithms with multi-sensor data fusion technology,the system can monitor UAV faults in real-time,diagnose them intelligently,and repair them autonomously. Experimental results show that the system can significantly improve the accuracy of UAV fault diagnosis and maintenance efficiency,effectively reduce maintenance costs,and enhance the overall reliability and safety of UAV.
2025, 41(19):172-174.
Abstract:Aiming at the problems such as insufficient communication real-time performance,complex access of heterogeneous devices and prominent security risks existing in the remote operation and maintenance of substation equipment,this article proposes a communication architecture design scheme for the power Internet of Things. This architecture is based on a hierarchical model,integrating the perception layer,network layer and application layer. It adopts low-latency transmission protocols and edge computing technology to enhance data processing efficiency,designs multi-protocol gateways to achieve ubiquitous device access,and builds a full-chain security protection system covering device authentication and dynamic encryption. The experimental results show that this architecture provides effective technical support for the intelligent operation and maintenance of the power system.
2025, 41(19):175-177.
Abstract:In response to the problems of high equipment loss,single fault scenarios,and limited practical opportunities for students in the traditional mode of vocational electrician training teaching,a development plan for a fault injection training system based on digital twin technology is proposed. The system constructs high-precision digital twins of power equipment,integrates fault injection algorithms and real-time interactive interfaces,and achieves dynamic generation of virtual fault scenarios and safe and controllable training operations. Taking the pilot application of electrical engineering major in a certain vocational school as an example,the system reduced equipment maintenance costs by 62%,improved student fault handling efficiency by 45%,and achieved a training coverage rate of 100%,verifying the feasibility and effectiveness of digital twin technology in vocational electrical training teaching.
2025, 41(19):178-180.
Abstract:Against the backdrop of the continuous growth in the demand for heterogeneous platform integration and sensitive data collaboration,traditional data sharing mechanisms generally encounter problems such as ambiguous permission boundaries,low encryption execution efficiency,and insufficient security of cross-domain links. To this end,this article constructs a cross-platform data sharing security system based on ciphertext compution. By integrating functions such as dynamic key policy scheduling,data security encapsulation and isolation,crossdomain storage and processing,precise mapping of computing tasks,and communication integrity verification,a closed-loop and controllable security sharing system is formed. Actual deployment and testing have shown that this system significantly outperforms the traditional shared model in key indicators such as scheduling delay, throughput load,cross-domain response efficiency,and abnormal behavior identification,demonstrating excellent engineering portability and the stability and reliability of the dense processing chain.
WANG Yanan , XIE Hongzhao , WANG Yaru
2025, 41(19):181-183.
Abstract:Aiming at the problems such as lagging progress management information and low multi-disciplinary collaboration efficiency existing in building water supply and drainage and electrical installation projects,this article proposes a progress management system solution based on visual modeling. This system integrates BIM technology,4D construction simulation and the coupled optimization algorithm of progress and resources to build a dynamic progress management model,achieving real-time visual monitoring of installation project progress, intelligent early warning of construction conflicts and dynamic allocation of resources.
2025, 41(19):184-185.
Abstract:With the advent of the big data era,optimizing production processes has become a key means to enhance the competitiveness of enterprises. The article studies the production process optimization model based on big data and its application. Firstly,the data collection and preprocessing process was analyzed in detail,an optimization model for the big data environment was constructed,and the selection methods of key parameters and decision variables were proposed. Subsequently,combined with actual cases,the application effect of this model in production data analysis and optimization was verified. The results show that the proposed model can effectively utilize big data technology to enhance production efficiency,reduce operating costs,and strengthen the flexibility and self-adaptability of the production system,thereby providing a scientific basis for enterprise decision-making.
2025, 41(19):186-188.
Abstract:This article proposes an improved solution based on HTML5 to address the challenges faced by mobile web front-end in cross platform adaptation,real-time interaction,and performance optimization. This solution adopts modular design,combined with new generation computing acceleration technology and intelligent caching mechanism,to achieve collaborative optimization of front-end rendering,resource loading,and layout response. Meanwhile,through compilation optimization,hardware acceleration,and adaptive layout algorithms,the execution efficiency of complex tasks and the user experience of multiple terminals have been effectively improved. The experimental results show that this scheme can significantly shorten response time and reduce system resource consumption,providing theoretical support and practical reference for the intelligent development and efficient iteration of mobile web applications.
ZAO Minghua , KONG Xiangxu , GUAN Xiongming , LIU Rong , GAO Yulei
2025, 41(19):189-191.
Abstract:In the context of digital transformation,security operation centers have gradually become an important support for enterprise network security management. However,traditional security operation centers focus on security operation and equipment monitoring,which is difficult to meet the increasingly growing security management needs of enterprises. The article proposes a security management oriented operation center system design,which takes enterprise security management business as the main line,organically connects security capabilities such as threat discovery,asset management,vulnerability management,etc.,and strengthens supervision and rectification through analysis and evaluation of execution results to achieve a management loop, thereby significantly improving the efficiency of enterprise network security management.
2025, 41(19):192-195.
Abstract:In view of the threats of adversarial attacks faced by artificial intelligence network intrusion detection systems and the deficiencies of existing defense schemes in terms of protocol compliance,multimodal fusion and processing efficiency,this article proposes the RoboNIDS robustness enhancement framework. This framework generates perturbation samples that conform to the RFC standard through protocol-aware adversarial training, optimizes the adaptive weighting of spatio-temporal features by using a gated multimodal fusion mechanism,and introduces a defense strategy based on uncertainty to achieve hierarchical risk response. Experiments on the CICIDS2017 dataset show that this system can reduce the average misjudgment rate of four types of mainstream adversarial attacks to 7.07%. While maintaining a real-time detection delay of 42.3 ms,it achieves an attack success rate/delay ratio of 0.17,and the unit protection effectiveness is increased by 11.6 times compared with the baseline method. It provides an efficient and reliable security paradigm for the protection of critical information infrastructure.
YANG Bin , YE Ling , ZHANG Haitao , YU Junjie
2025, 41(19):196-198.
Abstract:With the rapid development of smart parks,network and data security have become the core issues to ensure their efficient operation and sustained stable development. This article proposes a network and data security model based on multi-layer protection and intelligent management to address the problem of traditional security architecture being unable to effectively respond to diverse security threats. This model consists of four layers:basic security layer,network security layer,data security layer,and secure application layer. Each layer is equipped with corresponding functional modules to achieve comprehensive protection of the park network and data. The innovation lies in the introduction of a zero trust architecture and the integration of AI driven threat perception and data leakage prevention mechanisms,thereby building a more intelligent,dynamic,and efficient security system.
2025, 41(19):199-201.
Abstract:Based on the CiteSpace knowledge graph analysis tool,this article conducted a visual analysis of research literature related to network attack techniques,methods,and detection in the China National Knowledge Infrastructure(CNKI)database. Between 2014 and 2024,the overall number of publications in this field showed an upward trend,and peaked in 2022 before experiencing a decline. Analysis shows that the application of artificial intelligence technology in attack detection and defense has become the dominant direction,while overall research presents the characteristics of scattered distribution and weak cooperation.
2025, 41(19):202-204.
Abstract:With the development of the Internet era,the phenomenon of online copyright infringement has shown a continuous upward trend,and the traditional copyright liability system is gradually becoming inadequate in controlling online infringement activities. The current Civil Code of the People ?s Republic of China has supplemented and improved the relevant content in the tort liability section,but there are still unclear issues in the application of rules and the order of legal enforcement,which brings certain operational difficulties in the specific implementation process. The article draws on relevant theories from an international perspective to define the secondary tort liability of network service providers. By reviewing the current legal provisions in China,this article systematically summarizes the criteria for determining indirect infringement liability and its current implementation status,while pointing out the problems existing in the current system.
2025, 41(19):205-206.
Abstract:Quantum network technology is an interdisciplinary field of quantum information science and network communication technology,relying on core mechanisms such as quantum bits,entangled states,and quantum gates to construct information processing systems that surpass classical networks. This article systematically analyzes the technical architecture and protocol design of quantum networks,and deeply explores their unique advantages in applications such as quantum communication,computing,and sensing. The aim is to provide theoretical basis for the research and development of quantum network technology and decision-making reference for industrial application planning.
2025, 41(19):207-209.
Abstract:Regarding the application value of terminal log data in internal network behavior auditing,this article constructs an audit framework based on terminal logs and analyzes its actual effectiveness in scenarios such as abnormal behavior detection,security event tracing,and compliance verification. The experiment shows that this scheme can improve audit coverage to 98.7%,identify abnormal behavior with an accuracy of 92.4%,and shorten event response time by 65%,providing quantifiable technical support for enterprise network security protection and operational optimization.
2025, 41(19):210-212.
Abstract:The construction of cloud management systems in public places faces problems such as scattered architecture,“data silos”,and low management efficiency,which seriously restrict resource allocation and service quality. To enhance the level of system intelligence,the article proposes the construction of a unified data interaction framework,which optimizes multi departmental collaboration mechanisms,relies on standardized API, real-time data synchronization,and multi-source data fusion to achieve efficient data sharing. At the same time,by combining identity authentication,permission grading,and data encryption to ensure system security,intelligent scheduling algorithms are used to optimize the allocation of battlefield resources,and a data-driven decisionmaking loop is formed through IoT monitoring and automated operation and maintenance to achieve long-term stable operation of the system.
2025, 41(19):213-215.
Abstract:With the continuous increase in the penetration rate of new energy sources such as distributed photovoltaic and wind power in distribution networks,traditional radial distribution networks are gradually transforming into active distribution networks with multi-source interaction. Their operational reliability is facing new challenges such as intensified voltage fluctuations,bidirectional power flow,and complex protection coordination. To this end,the article proposes a reliability improvement framework for distribution networks based on the collaborative optimization of“source—grid—load—storage”,constructs a dynamic carrying capacity assessment model with a high proportion of distributed resources,and presents a configuration strategy for energy storage systems based on intelligent optimization algorithms. The simulation verification through the IEEE 33-node system shows that in the scenario where the penetration rate of distributed power sources reaches 45%,the proposed method can increase the system voltage qualification rate to 99.1%,shorten the fault recovery time to within 3 minutes,and effectively improve the power supply reliability and operational resilience of the system.
2025, 41(19):216-218.
Abstract:This article conducts a systematic analysis of the dissemination power of government wechat official accounts in the transportation field based on the WCI. Research has found that,regardless of whether they are provincial or municipal-level public accounts,mid-range influential accounts occupy a dominant position,and in terms of content types,“public services”and“government information”are the most concerned by users. Taking “Shenzhen Transportation”as a typical case,this article deeply analyzes the theme distribution and dissemination characteristics of its high-frequency hit content,and thereby proposes that improving content quality and innovating interactive forms are the key strategies to enhance the dissemination power of government wechat.
2025, 41(19):219-221.
Abstract:In fields such as industrial control and intelligent devices,embedded systems have put forward higher requirements for real-time data processing capabilities. To improve the efficiency of data processing,the application of parallel computing technology in embedded systems was studied. The task division and binding model in a multi-core environment was constructed. The on-chip cache consistency maintenance mechanism was designed,and the data channel structure and access path were optimized. The experimental results show that this strategy can effectively reduce the task response delay,improve the utilization rate of system resources,and at the same time has good adaptability and promotion value,which is of positive significance for enhancing the real-time processing performance of embedded systems.
LI Jun , LI Yuan , ZHAO Shuanbao , YU Shaorui , LIU Wei , BAN Bintang , ZHANG Shuoyi
2025, 41(19):222-224.
Abstract:This article focuses on the application of Phytium processors in the login process of desktop terminals, deeply analyzes its architectural characteristics and actual performance in the existing login mechanism,identifies the current performance and security bottlenecks,and proposes a systematic optimization plan. Specific measures include reconfiguring the login process,enhancing the identity verification mechanism,optimizing the process startup strategy,improving network communication and protocol design,as well as enhancing the compatibility of software and hardware. Research shows that the above optimization can significantly enhance the efficiency and security of desktop terminal login,providing beneficial technical support and reference basis for the design and practice of terminal systems based on domestic processors.
2025, 41(19):225-227.
Abstract:The renovation of water supply and drainage systems in old residential areas is a crucial aspect of urban renewal. However,traditional renovations often face challenges such as aging pipe networks,secondary pollution, and inefficient operation and maintenance. This article proposes a three-in-one renovation solution based on the technology of smart water supply and drainage systems,incorporating cutting-edge technologies such as the Internet of Things,big data,and artificial intelligence. Research indicates that smart renovation can increase the water supply reliability in old residential areas to 99.5%,reduce leakage rates to below 8%,and decrease operation and maintenance costs by 30%. This provides a replicable technical path for urban renewal.
2025, 41(19):228-230.
Abstract:Focusing on the informatization transformation of the maintenance process for large-scale equipment in hospitals,this article analyzes the issues present in the traditional maintenance process,elaborates on the specific content and implementation path of the informatization transformation,and conducts a comparative analysis from multiple dimensions,including direct costs,indirect costs,and benefits. The research indicates that the informatization transformation can significantly enhance maintenance efficiency, reduce failure rates of equipment,and prolong service life of equipment,thereby achieving a win-win situation for both economic and social benefits,and providing a scientific basis for decision-making in hospital equipment management.
2025, 41(19):231-233.
Abstract:Driven by the dual factors of uneven distribution of medical resources and worsening population aging, telemedicine is transitioning from auxiliary services to core medical models. As a key technology carrier, intelligent medical devices build a telemedicine infrastructure covering the whole chain of “perception transmission decision”by integrating multimodal sensors,edge computing and low latency communication modules. With the breakthrough of flexible electronics and digital twin technology,intelligent medical devices will promote the evolution of remote healthcare towards the 3.0 stage of“non-invasive monitoring,precise intervention, and global collaboration”,providing a technological foundation for building a universal medical and health service system.
2025, 41(19):234-236.
Abstract:Artificial Intelligence(AI)technology is deeply reconstructing the career planning and employment entrepreneurship ecosystem. This article analyzes 12 typical application scenarios, including intelligent recruitment,personalized learning,and virtual internships,to reveal the empowering mechanism of AI in career positioning,skill matching,entrepreneurial incubation,and other aspects. Research shows that AI driven career planning systems increase job matching efficiency by 47% and increase entrepreneurial project survival rates by 32%. Based on practical cases in industries such as education,healthcare,and finance,a three-dimensional integrated AI career planning framework of“technology scenario capability”is proposed,providing theoretical support and practical paths for talent development in the digital age.
2025, 41(19):237-239.
Abstract:Under the background of continuous promotion of medical informatization policies and constant technological innovation,hospital information systems (HIS) are gradually evolving towards intelligence and integration. However,the current systems still generally face challenges such as“data silos”,insufficient compatibility and information security risks. Based on the practical experience of medical informatization at home and abroad,the system proposed multi-dimensional optimization paths,and focused on discussing the deep integration and application of technologies such as cloud computing,artificial intelligence and the Internet of Things in HIS. Centering on core modules such as the upgrade of electronic medical records,clinical decision support,and data security management,this paper analyzes their intelligent development trends and innovative strategies,such as intelligent diagnosis and treatment assistance,cross-institutional data sharing,and blockchainbased security protection mechanisms. Research shows that the intelligent transformation of HIS not only helps improve the operational efficiency of hospitals,but also significantly optimizes the quality of medical services, providing theoretical references and practical lessons for the construction of digital hospitals.
2025, 41(19):240-242.
Abstract:In emergency scenarios such as natural disasters and sudden events,traditional communication networks are vulnerable to damage,leading to information transmission disruptions. Edge computing provides a solution with low latency and high reliability for emergency communication by sinking computing and storage resources to the edge of the network.
2025, 41(19):243-245.
Abstract:This article aims to investigate the knowledge level and application behavior status of college students regarding Generative Artificial Intelligence(GenAI),and analyze the relationship between the two. By using the convenience sampling method,823 college students from a certain medical college were selected as the research subjects. A generative artificial intelligence“knowledge,attitude and practice”survey was conducted through a self-made questionnaire to collect demographic information,knowledge level,attitude and application behavior data. Research shows that college students? application behavior of generative artificial intelligence is relatively common and positive,but their knowledge mastery is relatively low. Therefore,it is necessary to incorporate the artificial intelligence education system into the higher education curriculum system to comprehensively enhance students? digital literacy and artificial intelligence cognitive abilities.
2025, 41(19):246-249.
Abstract:With the continuous breakthroughs and deepening applications of artificial intelligence technology, intelligent manufacturing has become the core driving force for promoting the transformation and upgrading of the global manufacturing industry. Currently,the intelligent robot industry maintains a significant growth trend and plays an irreplaceable role in the intelligent manufacturing system,gradually becoming an important engine for the high-quality development of the manufacturing industry. This article focuses on key fields such as industrial robots,service robots,and special robots,systematically analyzes their current industrial development status and competitive landscape,and further looks forward to emerging development trends in this field.
2025, 41(19):250-252.
Abstract:Based on the analysis of the main problems faced by the impact of coating damage on aircraft RCS,this article explores simulation strategies for the influence of coating damage on aircraft RCS from the aspects of component selection and model construction,damage setting,and scheme formulation. It has certain guiding significance for specific simulation analysis.
2025, 41(19):253-255.
Abstract:In recent years,neural networks have been widely applied in various tasks. In supervised learning, neural networks often rely on given training data to optimize overall classification accuracy. However,existing methods struggle to specifically reduce the classification performance of particular categories. To address this issue,this article proposes a Threshold-based Inhibitory Synaptic Join method. This approach adjusts the original network structure by introducing additional cross-layer inhibitory synaptic connections between hidden-layer neurons and target output-layer neurons,with appropriately assigned weights. Experimental results demonstrate that the TISJ method can effectively suppress the classification accuracy of target categories while maintaining the stability of the networks overall performance.
2025, 41(19):256-258.
Abstract:Soil organic matter(SOM)can improve soil structure,enhance soil fertility,and retain moisture. Its content is a core indicator for measuring soil fertility and is of great significance for improving farmland productivity and ensuring food security. In recent years,hyperspectral technology has provided a new approach for rapidly and effectively estimating soil organic matter content. To further improve the estimation accuracy,the article proposes an innovative spectral processing method and constructs a high-precision SOM estimation model.
2025, 41(19):259-261.
Abstract:Due to the complex process,diverse resources,and susceptibility to environmental interference, traditional schedule management in landscape engineering relies on two-dimensional drawings and static plans, which leads to problems such as information lag and unreasonable resource allocation. To explore the application of 4D construction progress simulation technology in landscape engineering,an integrated dynamic model of“space time resources”is constructed to visualize and dynamically manage the construction process,achieving the goals of optimizing resource allocation,shortening construction period,and reducing costs.
LI Xinwei , LIU Xiejun , HU Xinwen , YANG Yu
2025, 41(19):262-264.
Abstract:Due to the coupling effect of multiple factors,predicting the prices of traditional Chinese medicinal materials is highly difficult. To accurately predict the prices of traditional Chinese medicine,this article integrates multi-source data and sentiment analysis methods to construct an ARIMA-LSTM combination model. Firstly, collect and clean multi-source data,extract text features using word embeddings and BERT models,and calculate sentiment scores. Subsequently,based on the ARIMA model,long-term trend prediction is carried out and relevant parameters are determined. For short-term fluctuations,the residuals generated by the ARIMA model are modeled using the LSTM model,and relevant data is introduced to enhance the prediction ability,ultimately obtaining an integrated prediction result. The model was evaluated by MAPE,RMSE and other evaluation indicators,and the results showed that the combined model had better prediction performance than the single model,which could provide a reliable basis for the decision-making of Chinese herbal medicine related industries.
2025, 41(19):265-267.
Abstract:Drawing on the advanced experience of CNAS laboratory management system, this article systematically summarizes the practical experience of software engineering cost services,guided by standards and specifications such as GB/T 36964—2018 and ISO/IEC 24570,combined with industry regulations,relevant theories,and existing project experience databases. On this basis,a comprehensive software cost evaluation methodology and practical operation mode have been constructed. This methodology can effectively improve service quality and efficiency,provide more accurate and reliable software cost assessment services for enterprises,and has strong application value and promotion significance.
2025, 41(19):268-270.
Abstract:In order to promote the mutual recognition of inspection and testing results among medical institutions, assist in the efficient integration of hospital information construction and medical resources, this paper systematically reviews the implementation status of mutual recognition work,explores its theoretical significance and core values,points out the practical difficulties and management barriers faced in the process,and proposes a targeted mechanism construction plan supported by policy guidance,technical empowerment,and institutional guarantees.
Address:重庆市渝北区洪湖西路 18 号西南信息楼Zip Code:401121
Phone:023-63507395 Email:jsjyywzbjb@163.com
® 2026 All Rights Reserved
渝公网安备50019002505578号
京ICP备09084417号-24
Supported by:Beijing E-Tiller Technology Development Co., Ltd.