With the rapid evolution of technology and the increasing complexity of software systems,there is a growing demand for effective educational approaches that empower learners to acquire and apply software engineering s...With the rapid evolution of technology and the increasing complexity of software systems,there is a growing demand for effective educational approaches that empower learners to acquire and apply software engineering skills in practical contexts.This paper presents an intelligent and interactive learning(Meta-SEE)framework for software engineering education that combines the immersive capabilities of the metaverse with the cognitive processes of metacognition,to create an interactive and engaging learning environment.In the Meta-SEE framework,learners are immersed in a virtual world where they can collaboratively engage with concepts and practices of software engineering.Through the integration of metacognitive strategies,learners are empowered to monitor,regulate,and adapt their learning processes.By incorporating metacognition within the metaverse,learners gain a deeper understanding of their own thinking processes and become self-directed learners.In addition,MetaSEE has the potential to revolutionize software engineering education by offering a dynamic,immersive,and personalized learning experience.It allows learners to engage in realistic software development scenarios,explore complex systems,and collaborate with peers and instructors in virtual spaces.展开更多
As a highly advanced conversational AI chatbot trained on extensive datasets,ChatGPT has garnered significant attention across various domains,including academia,industry,and education.In the field of education,existi...As a highly advanced conversational AI chatbot trained on extensive datasets,ChatGPT has garnered significant attention across various domains,including academia,industry,and education.In the field of education,existing studies primarily focus on 2 areas:Assessing the potential utility of ChatGPT in education by examining its capabilities and limitations;exploring the educational scenarios that could benefit from the integration of ChatGPT.In contrast to these studies,we conduct a user survey targeting undergraduate students specializing in Software Engineering,aiming to gain insights into their perceptions,challenges,and expectations regarding the utilization of ChatGPT.Based on the results of the survey,we provide valuable guidance on the effective incorporation of ChatGPT in the realm of software engineering education.展开更多
This paper focuses on the problems,opportunities,and challenges faced by software engineering education in the new era.We have studied the core ideas of the new model and reform,the specific measures implemented,and t...This paper focuses on the problems,opportunities,and challenges faced by software engineering education in the new era.We have studied the core ideas of the new model and reform,the specific measures implemented,and the challenges and solutions faced.The new model and reform must focus on cultivating practical abilities,introducing interdisciplinary knowledge,and strengthening innovation awareness and entrepreneurial spirit.The process of reform and innovation is carried out from the aspects of teaching methods,teaching means,and course performance evaluation in the teaching practice of software engineering courses.We adopt a method of“question guiding,simple and easy to understand,flexible and diverse,and emphasizing practical results”,optimizing the curriculum design,providing diverse learning opportunities,and establishing a platform for the industry-university-research cooperation.Our teaching philosophy is to adhere to the viewpoint of innovative teaching ideas,optimizing teaching methods and teaching means,and comprehensively improving the teaching quality and level of software engineering education.展开更多
In order to respond to the new engineering construction of the Ministry of Education,and explore the innovative talent training model of collaborative education and multidisciplinary integration,this paper relies on t...In order to respond to the new engineering construction of the Ministry of Education,and explore the innovative talent training model of collaborative education and multidisciplinary integration,this paper relies on the software engineering teaching team of the School of Software Engineering,Beijing University of Posts and Telecommunications,through the implementation of the collaborative education project of the Ministry of Education,and proposes the multi-course collaborative practice teaching system,through the reasonable cross-fusion of the practical links of the 5 software engineering courses in the college,realizes the multi-course collaborative education and reasonable cross-fusion of courses,shares practical project resources,introduces new enterprise technologies,and guides students’innovation and entrepreneurship provide a meaningful reference for the collaborative arrangement of teaching content and cross-disciplinary integration in the current university education system.展开更多
Financial time series prediction,whether for classification or regression,has been a heated research topic over the last decade.While traditional machine learning algorithms have experienced mediocre results,deep lear...Financial time series prediction,whether for classification or regression,has been a heated research topic over the last decade.While traditional machine learning algorithms have experienced mediocre results,deep learning has largely contributed to the elevation of the prediction performance.Currently,the most up-to-date review of advanced machine learning techniques for financial time series prediction is still lacking,making it challenging for finance domain experts and relevant practitioners to determine which model potentially performs better,what techniques and components are involved,and how themodel can be designed and implemented.This review article provides an overview of techniques,components and frameworks for financial time series prediction,with an emphasis on state-of-the-art deep learning models in the literature from2015 to 2023,including standalonemodels like convolutional neural networks(CNN)that are capable of extracting spatial dependencies within data,and long short-term memory(LSTM)that is designed for handling temporal dependencies;and hybrid models integrating CNN,LSTM,attention mechanism(AM)and other techniques.For illustration and comparison purposes,models proposed in recent studies are mapped to relevant elements of a generalized framework comprised of input,output,feature extraction,prediction,and related processes.Among the state-of-the-artmodels,hybrid models like CNNLSTMand CNN-LSTM-AM in general have been reported superior in performance to stand-alone models like the CNN-only model.Some remaining challenges have been discussed,including non-friendliness for finance domain experts,delayed prediction,domain knowledge negligence,lack of standards,and inability of real-time and highfrequency predictions.The principal contributions of this paper are to provide a one-stop guide for both academia and industry to review,compare and summarize technologies and recent advances in this area,to facilitate smooth and informed implementation,and to highlight future research directions.展开更多
The critical rainfall of runoff-initiated debris flows is utmost importance for local early hazard forecasting.This paper presents research on the critical rainfall of runoff-initiated debris flows through comparisons...The critical rainfall of runoff-initiated debris flows is utmost importance for local early hazard forecasting.This paper presents research on the critical rainfall of runoff-initiated debris flows through comparisons between slope gradients and three key factors,including topographic contributing area,dimensionless discharge,and Shields stress.The rainfall amount was estimated by utilizing in-situ rainfall records and a slope-dependent Shields stress model was created.The created model can predict critical Shields stress more accurately than the other two models.Furthermore,a new dimensionless discharge equation was proposed based on the corresponding discharge-gradient datasets.The new equation,along with factors such as contributing area above bed failure sites,channel width,and mean diameter of debris flow deposits,predicts a smaller rainfall amount than the in-situ measured records.Although the slope-dependent Shields stress model performs well and the estimated rainfall amount is lower than the in-situ records,the sediment initiation in the experiments falls within sheet flow regime due to a large Shields stress.Therefore,further sediment initiation experiments at a steeper slope range are expected in the future to ensure that the sediment transport belongs to mass failure regime characterized by a low level of Shields stress.Finally,a more accurate hazard forecast on the runoff-initiated debris flow holds promise when the corresponding critical slope-dependent dimensionless discharge of no motion,fluvial sediment transport,mass flow regime,and sheet flow regime are considered.展开更多
Aim:This study aims to establish an artificial intelligence model,ThyroidNet,to diagnose thyroid nodules using deep learning techniques accurately.Methods:A novel method,ThyroidNet,is introduced and evaluated based on...Aim:This study aims to establish an artificial intelligence model,ThyroidNet,to diagnose thyroid nodules using deep learning techniques accurately.Methods:A novel method,ThyroidNet,is introduced and evaluated based on deep learning for the localization and classification of thyroid nodules.First,we propose the multitask TransUnet,which combines the TransUnet encoder and decoder with multitask learning.Second,we propose the DualLoss function,tailored to the thyroid nodule localization and classification tasks.It balances the learning of the localization and classification tasks to help improve the model’s generalization ability.Third,we introduce strategies for augmenting the data.Finally,we submit a novel deep learning model,ThyroidNet,to accurately detect thyroid nodules.Results:ThyroidNet was evaluated on private datasets and was comparable to other existing methods,including U-Net and TransUnet.Experimental results show that ThyroidNet outperformed these methods in localizing and classifying thyroid nodules.It achieved improved accuracy of 3.9%and 1.5%,respectively.Conclusion:ThyroidNet significantly improves the clinical diagnosis of thyroid nodules and supports medical image analysis tasks.Future research directions include optimization of the model structure,expansion of the dataset size,reduction of computational complexity and memory requirements,and exploration of additional applications of ThyroidNet in medical image analysis.展开更多
The concept of“New Engineering”has put forward new challenges to the talents cultivation of universities.Due to some problems of the traditional Software Engineering curriculum,e.g.separated design at undergraduate-...The concept of“New Engineering”has put forward new challenges to the talents cultivation of universities.Due to some problems of the traditional Software Engineering curriculum,e.g.separated design at undergraduate-level and graduate-level courses,poor curriculum structure,lacking of industry characteristics.This paper proposes an integrated undergraduate-graduate education curriculum for Software Engineering Major,which is based on Software Engineering specialty knowledge system(C-SWEBOK)and focuses on the current national strategic demands.Additionally,the curriculum combines with the University’s transportation characteristics,and fuses the discipline of Software Engineering and Intelligent Transportation.The multi-level curriculum designed in this paper is with reasonable structure,complete system,progressive content,and salient feature,which provides the strong support for cultivating high-qualified software talents in line with national strategies and industry needs.展开更多
"Semester Training"has been adopted as an important part of the personnel training in software engineering majors since it was first put forward.The ultimate goal of semester training is to improve the profe..."Semester Training"has been adopted as an important part of the personnel training in software engineering majors since it was first put forward.The ultimate goal of semester training is to improve the professional quality of students in an all-round way,then eventually achieve the goal of satisfactory employment for both students and enterprises.However,in order to achieve the above purpose,the design of traditional training project still has the following problems:the topic selection of traditional training is designed by teachers in college,which lacks the training of engineering ability aiming at practical problems;the content and technology of traditional project training are out of date,ignoring the urgent demand of software industry development for advanced technology application;the traditional project training inspects the mastery of knowledge in each semester Degree,ignores the incremental of a progressive training system.In view of the above problems,this study proposes an Application-Oriented Software Engineering Semester Training System.Practice has proved that the construction of the training system can effectively improve the quality of teaching,so as to further improve the comprehensive quality of students.展开更多
In view of the increasingly rapid development of global economic integration and combined with the existing modes of training international software engineering talents in China,this paper deeply analyzes and obtains ...In view of the increasingly rapid development of global economic integration and combined with the existing modes of training international software engineering talents in China,this paper deeply analyzes and obtains the existing problems in the current teaching process,and proposes various teaching reform measures under the guidance of CDIO higher engineering education thought.Through many years of teaching practice experience,we can find that our reform has achieved remarkable results.展开更多
In recent years,with the great success of pre-trained language models,the pre-trained BERT model has been gradually applied to the field of source code understanding.However,the time cost of training a language model ...In recent years,with the great success of pre-trained language models,the pre-trained BERT model has been gradually applied to the field of source code understanding.However,the time cost of training a language model from zero is very high,and how to transfer the pre-trained language model to the field of smart contract vulnerability detection is a hot research direction at present.In this paper,we propose a hybrid model to detect common vulnerabilities in smart contracts based on a lightweight pre-trained languagemodel BERT and connected to a bidirectional gate recurrent unitmodel.The downstream neural network adopts the bidirectional gate recurrent unit neural network model with a hierarchical attention mechanism to mine more semantic features contained in the source code of smart contracts by using their characteristics.Our experiments show that our proposed hybrid neural network model SolBERT-BiGRU-Attention is fitted by a large number of data samples with smart contract vulnerabilities,and it is found that compared with the existing methods,the accuracy of our model can reach 93.85%,and the Micro-F1 Score is 94.02%.展开更多
With the extensive integration of the Internet,social networks and the internet of things,the social internet of things has increasingly become a significant research issue.In the social internet of things application...With the extensive integration of the Internet,social networks and the internet of things,the social internet of things has increasingly become a significant research issue.In the social internet of things application scenario,one of the greatest challenges is how to accurately recommend or match smart objects for users with massive resources.Although a variety of recommendation algorithms have been employed in this field,they ignore the massive text resources in the social internet of things,which can effectively improve the effect of recommendation.In this paper,a smart object recommendation approach named object recommendation based on topic learning and joint features is proposed.The proposed approach extracts and calculates topics and service relevant features of texts related to smart objects and introduces the“thing-thing”relationship information in the internet of things to improve the effect of recommendation.Experiments show that the proposed approach enables higher accuracy compared to the existing recommendation methods.展开更多
With the expansion of network services,large-scale networks have progressively become common.The network status changes rapidly in response to customer needs and configuration changes,so network configuration changes ...With the expansion of network services,large-scale networks have progressively become common.The network status changes rapidly in response to customer needs and configuration changes,so network configuration changes are also very frequent.However,no matter what changes,the network must ensure the correct conditions,such as isolating tenants from each other or guaranteeing essential services.Once changes occur,it is necessary to verify the after-changed network.Whereas,for the verification of large-scale network configuration changes,many current verifiers show poor efficiency.In order to solve the problem ofmultiple global verifications caused by frequent updates of local configurations in large networks,we present a fast configuration updates verification tool,FastCUV,for distributed control planes.FastCUV aims to enhance the efficiency of distributed control plane verification for medium and large networks while ensuring correctness.This paper presents a method to determine the network range affected by the configuration change.We present a flow model and graph structure to facilitate the design of verification algorithms and speed up verification.Our scheme verifies the network area affected by obtaining the change of the Forwarding Information Base(FIB)before and after.FastCUV supports rich network attributes,meanwhile,has high efficiency and correctness performance.After experimental verification and result analysis,our method outperforms the state-of-the-art method to a certain extent.展开更多
We present a novel approach for the prediction of crystal material properties that is distinct from the computationally complex and expensive density functional theory(DFT)-based calculations.Instead,we utilize an att...We present a novel approach for the prediction of crystal material properties that is distinct from the computationally complex and expensive density functional theory(DFT)-based calculations.Instead,we utilize an attention-based graph neural network that yields high-accuracy predictions.Our approach employs two attention mechanisms that allow for message passing on the crystal graphs,which in turn enable the model to selectively attend to pertinent atoms and their local environments,thereby improving performance.We conduct comprehensive experiments to validate our approach,which demonstrates that our method surpasses existing methods in terms of predictive accuracy.Our results suggest that deep learning,particularly attention-based networks,holds significant promise for predicting crystal material properties,with implications for material discovery and the refined intelligent systems.展开更多
To capitalize on the primary role of major course teaching and to facilitate students’understanding of abstract concepts in the data structure course,it is essential to increase their interest in learning and develop...To capitalize on the primary role of major course teaching and to facilitate students’understanding of abstract concepts in the data structure course,it is essential to increase their interest in learning and develop case studies that highlight fine traditional culture.By incorporating these culture-rich case studies into classroom instruction,we employ a project-driven teaching approach.This not only allows students to master professional knowledge,but also enhances their abilities to solve specific engineering problems,ultimately fostering cultural confidence.Over the past few years,during which educational reforms have been conducted for trial runs,the feasibility and effectiveness of these reform schemes have been demonstrated.展开更多
Automatic crack detection of cement pavement chiefly benefits from the rapid development of deep learning,with convolutional neural networks(CNN)playing an important role in this field.However,as the performance of cr...Automatic crack detection of cement pavement chiefly benefits from the rapid development of deep learning,with convolutional neural networks(CNN)playing an important role in this field.However,as the performance of crack detection in cement pavement improves,the depth and width of the network structure are significantly increased,which necessitates more computing power and storage space.This limitation hampers the practical implementation of crack detection models on various platforms,particularly portable devices like small mobile devices.To solve these problems,we propose a dual-encoder-based network architecture that focuses on extracting more comprehensive fracture feature information and combines cross-fusion modules and coordinated attention mechanisms formore efficient feature fusion.Firstly,we use small channel convolution to construct shallow feature extractionmodule(SFEM)to extract low-level feature information of cracks in cement pavement images,in order to obtainmore information about cracks in the shallowfeatures of images.In addition,we construct large kernel atrous convolution(LKAC)to enhance crack information,which incorporates coordination attention mechanism for non-crack information filtering,and large kernel atrous convolution with different cores,using different receptive fields to extract more detailed edge and context information.Finally,the three-stage feature map outputs from the shallow feature extraction module is cross-fused with the two-stage feature map outputs from the large kernel atrous convolution module,and the shallow feature and detailed edge feature are fully fused to obtain the final crack prediction map.We evaluate our method on three public crack datasets:DeepCrack,CFD,and Crack500.Experimental results on theDeepCrack dataset demonstrate the effectiveness of our proposed method compared to state-of-the-art crack detection methods,which achieves Precision(P)87.2%,Recall(R)87.7%,and F-score(F1)87.4%.Thanks to our lightweight crack detectionmodel,the parameter count of the model in real-world detection scenarios has been significantly reduced to less than 2M.This advancement also facilitates technical support for portable scene detection.展开更多
Accurate automatic segmentation of gliomas in various sub-regions,including peritumoral edema,necrotic core,and enhancing and non-enhancing tumor core from 3D multimodal MRI images,is challenging because of its highly...Accurate automatic segmentation of gliomas in various sub-regions,including peritumoral edema,necrotic core,and enhancing and non-enhancing tumor core from 3D multimodal MRI images,is challenging because of its highly heterogeneous appearance and shape.Deep convolution neural networks(CNNs)have recently improved glioma segmentation performance.However,extensive down-sampling such as pooling or stridden convolution in CNNs significantly decreases the initial image resolution,resulting in the loss of accurate spatial and object parts information,especially information on the small sub-region tumors,affecting segmentation performance.Hence,this paper proposes a novel multi-level parallel network comprising three different level parallel subnetworks to fully use low-level,mid-level,and high-level information and improve the performance of brain tumor segmentation.We also introduce the Combo loss function to address input class imbalance and false positives and negatives imbalance in deep learning.The proposed method is trained and validated on the BraTS 2020 training and validation dataset.On the validation dataset,ourmethod achieved a mean Dice score of 0.907,0.830,and 0.787 for the whole tumor,tumor core,and enhancing tumor core,respectively.Compared with state-of-the-art methods,the multi-level parallel network has achieved competitive results on the validation dataset.展开更多
Graphics Processing Units(GPUs)are used to accelerate computing-intensive tasks,such as neural networks,data analysis,high-performance computing,etc.In the past decade or so,researchers have done a lot of work on GPU ...Graphics Processing Units(GPUs)are used to accelerate computing-intensive tasks,such as neural networks,data analysis,high-performance computing,etc.In the past decade or so,researchers have done a lot of work on GPU architecture and proposed a variety of theories and methods to study the microarchitectural characteristics of various GPUs.In this study,the GPU serves as a co-processor and works together with the CPU in an embedded real-time system to handle computationally intensive tasks.It models the architecture of the GPU and further considers it based on some excellent work.The SIMT mechanism and Cache-miss situation provide a more detailed analysis of the GPU architecture.In order to verify the GPU architecture model proposed in this article,10 GPU kernel_task and an Nvidia GPU device were used to perform experiments.The experimental results showed that the minimum error between the kernel task execution time predicted by the GPU architecture model proposed in this article and the actual measured kernel task execution time was 3.80%,and the maximum error was 8.30%.展开更多
Cascading failure can cause great damage to complex networks, so it is of great significance to improve the network robustness against cascading failure. Many previous existing works on load-redistribution strategies ...Cascading failure can cause great damage to complex networks, so it is of great significance to improve the network robustness against cascading failure. Many previous existing works on load-redistribution strategies require global information, which is not suitable for large scale networks, and some strategies based on local information assume that the load of a node is always its initial load before the network is attacked, and the load of the failure node is redistributed to its neighbors according to their initial load or initial residual capacity. This paper proposes a new load-redistribution strategy based on local information considering an ever-changing load. It redistributes the loads of the failure node to its nearest neighbors according to their current residual capacity, which makes full use of the residual capacity of the network. Experiments are conducted on two typical networks and two real networks, and the experimental results show that the new load-redistribution strategy can reduce the size of cascading failure efficiently.展开更多
基金partially funded by the 2023 Teaching Quality Engineering Construction Project of Sun Yat-sen University(No.76250-12230036)the 2023 Project of Computer Education Research Association of Chinese Universities(No.CERACU2023R02)。
文摘With the rapid evolution of technology and the increasing complexity of software systems,there is a growing demand for effective educational approaches that empower learners to acquire and apply software engineering skills in practical contexts.This paper presents an intelligent and interactive learning(Meta-SEE)framework for software engineering education that combines the immersive capabilities of the metaverse with the cognitive processes of metacognition,to create an interactive and engaging learning environment.In the Meta-SEE framework,learners are immersed in a virtual world where they can collaboratively engage with concepts and practices of software engineering.Through the integration of metacognitive strategies,learners are empowered to monitor,regulate,and adapt their learning processes.By incorporating metacognition within the metaverse,learners gain a deeper understanding of their own thinking processes and become self-directed learners.In addition,MetaSEE has the potential to revolutionize software engineering education by offering a dynamic,immersive,and personalized learning experience.It allows learners to engage in realistic software development scenarios,explore complex systems,and collaborate with peers and instructors in virtual spaces.
基金supported in part by the Guangdong Basic and Applied Basic Research Foundation(Grant No.2023A1515012292)the 2023 Teaching Quality Engineering Construction Project of Sun Yat-sen University(Grant No.76250-12230036)the 2023 Project of Computer Education Research Association ofChinese Universities(Grant No.CERACU2023R02)。
文摘As a highly advanced conversational AI chatbot trained on extensive datasets,ChatGPT has garnered significant attention across various domains,including academia,industry,and education.In the field of education,existing studies primarily focus on 2 areas:Assessing the potential utility of ChatGPT in education by examining its capabilities and limitations;exploring the educational scenarios that could benefit from the integration of ChatGPT.In contrast to these studies,we conduct a user survey targeting undergraduate students specializing in Software Engineering,aiming to gain insights into their perceptions,challenges,and expectations regarding the utilization of ChatGPT.Based on the results of the survey,we provide valuable guidance on the effective incorporation of ChatGPT in the realm of software engineering education.
基金supported in part by the postgraduate demonstration course of Guangdong Province Department of Education Programmed Trading(No.2023SFKC_022)the Computer Architecture First Class Course Project,South China Normal University-Baidu Pineapple Talent Training Practice Basethe 2023 Project of Computer Education Research Association of Chinese Universities(No.CERACU2023R02)。
文摘This paper focuses on the problems,opportunities,and challenges faced by software engineering education in the new era.We have studied the core ideas of the new model and reform,the specific measures implemented,and the challenges and solutions faced.The new model and reform must focus on cultivating practical abilities,introducing interdisciplinary knowledge,and strengthening innovation awareness and entrepreneurial spirit.The process of reform and innovation is carried out from the aspects of teaching methods,teaching means,and course performance evaluation in the teaching practice of software engineering courses.We adopt a method of“question guiding,simple and easy to understand,flexible and diverse,and emphasizing practical results”,optimizing the curriculum design,providing diverse learning opportunities,and establishing a platform for the industry-university-research cooperation.Our teaching philosophy is to adhere to the viewpoint of innovative teaching ideas,optimizing teaching methods and teaching means,and comprehensively improving the teaching quality and level of software engineering education.
基金supported in part by Educational Reform Projects of BUPT.
文摘In order to respond to the new engineering construction of the Ministry of Education,and explore the innovative talent training model of collaborative education and multidisciplinary integration,this paper relies on the software engineering teaching team of the School of Software Engineering,Beijing University of Posts and Telecommunications,through the implementation of the collaborative education project of the Ministry of Education,and proposes the multi-course collaborative practice teaching system,through the reasonable cross-fusion of the practical links of the 5 software engineering courses in the college,realizes the multi-course collaborative education and reasonable cross-fusion of courses,shares practical project resources,introduces new enterprise technologies,and guides students’innovation and entrepreneurship provide a meaningful reference for the collaborative arrangement of teaching content and cross-disciplinary integration in the current university education system.
基金funded by the Natural Science Foundation of Fujian Province,China (Grant No.2022J05291)Xiamen Scientific Research Funding for Overseas Chinese Scholars.
文摘Financial time series prediction,whether for classification or regression,has been a heated research topic over the last decade.While traditional machine learning algorithms have experienced mediocre results,deep learning has largely contributed to the elevation of the prediction performance.Currently,the most up-to-date review of advanced machine learning techniques for financial time series prediction is still lacking,making it challenging for finance domain experts and relevant practitioners to determine which model potentially performs better,what techniques and components are involved,and how themodel can be designed and implemented.This review article provides an overview of techniques,components and frameworks for financial time series prediction,with an emphasis on state-of-the-art deep learning models in the literature from2015 to 2023,including standalonemodels like convolutional neural networks(CNN)that are capable of extracting spatial dependencies within data,and long short-term memory(LSTM)that is designed for handling temporal dependencies;and hybrid models integrating CNN,LSTM,attention mechanism(AM)and other techniques.For illustration and comparison purposes,models proposed in recent studies are mapped to relevant elements of a generalized framework comprised of input,output,feature extraction,prediction,and related processes.Among the state-of-the-artmodels,hybrid models like CNNLSTMand CNN-LSTM-AM in general have been reported superior in performance to stand-alone models like the CNN-only model.Some remaining challenges have been discussed,including non-friendliness for finance domain experts,delayed prediction,domain knowledge negligence,lack of standards,and inability of real-time and highfrequency predictions.The principal contributions of this paper are to provide a one-stop guide for both academia and industry to review,compare and summarize technologies and recent advances in this area,to facilitate smooth and informed implementation,and to highlight future research directions.
基金supported by the by the Second Tibetan Plateau Scientific Expedition and Research Program(Grant No.2019QZKK0902)Beijing Municipal Science and Technology Project(Z191100001419015).The。
文摘The critical rainfall of runoff-initiated debris flows is utmost importance for local early hazard forecasting.This paper presents research on the critical rainfall of runoff-initiated debris flows through comparisons between slope gradients and three key factors,including topographic contributing area,dimensionless discharge,and Shields stress.The rainfall amount was estimated by utilizing in-situ rainfall records and a slope-dependent Shields stress model was created.The created model can predict critical Shields stress more accurately than the other two models.Furthermore,a new dimensionless discharge equation was proposed based on the corresponding discharge-gradient datasets.The new equation,along with factors such as contributing area above bed failure sites,channel width,and mean diameter of debris flow deposits,predicts a smaller rainfall amount than the in-situ measured records.Although the slope-dependent Shields stress model performs well and the estimated rainfall amount is lower than the in-situ records,the sediment initiation in the experiments falls within sheet flow regime due to a large Shields stress.Therefore,further sediment initiation experiments at a steeper slope range are expected in the future to ensure that the sediment transport belongs to mass failure regime characterized by a low level of Shields stress.Finally,a more accurate hazard forecast on the runoff-initiated debris flow holds promise when the corresponding critical slope-dependent dimensionless discharge of no motion,fluvial sediment transport,mass flow regime,and sheet flow regime are considered.
基金supported by MRC,UK (MC_PC_17171)Royal Society,UK (RP202G0230)+8 种基金BHF,UK (AA/18/3/34220)Hope Foundation for Cancer Research,UK (RM60G0680)GCRF,UK (P202PF11)Sino-UK Industrial Fund,UK (RP202G0289)LIAS,UK (P202ED10,P202RE969)Data Science Enhancement Fund,UK (P202RE237)Fight for Sight,UK (24NN201)Sino-UK Education Fund,UK (OP202006)BBSRC,UK (RM32G0178B8).
文摘Aim:This study aims to establish an artificial intelligence model,ThyroidNet,to diagnose thyroid nodules using deep learning techniques accurately.Methods:A novel method,ThyroidNet,is introduced and evaluated based on deep learning for the localization and classification of thyroid nodules.First,we propose the multitask TransUnet,which combines the TransUnet encoder and decoder with multitask learning.Second,we propose the DualLoss function,tailored to the thyroid nodule localization and classification tasks.It balances the learning of the localization and classification tasks to help improve the model’s generalization ability.Third,we introduce strategies for augmenting the data.Finally,we submit a novel deep learning model,ThyroidNet,to accurately detect thyroid nodules.Results:ThyroidNet was evaluated on private datasets and was comparable to other existing methods,including U-Net and TransUnet.Experimental results show that ThyroidNet outperformed these methods in localizing and classifying thyroid nodules.It achieved improved accuracy of 3.9%and 1.5%,respectively.Conclusion:ThyroidNet significantly improves the clinical diagnosis of thyroid nodules and supports medical image analysis tasks.Future research directions include optimization of the model structure,expansion of the dataset size,reduction of computational complexity and memory requirements,and exploration of additional applications of ThyroidNet in medical image analysis.
文摘The concept of“New Engineering”has put forward new challenges to the talents cultivation of universities.Due to some problems of the traditional Software Engineering curriculum,e.g.separated design at undergraduate-level and graduate-level courses,poor curriculum structure,lacking of industry characteristics.This paper proposes an integrated undergraduate-graduate education curriculum for Software Engineering Major,which is based on Software Engineering specialty knowledge system(C-SWEBOK)and focuses on the current national strategic demands.Additionally,the curriculum combines with the University’s transportation characteristics,and fuses the discipline of Software Engineering and Intelligent Transportation.The multi-level curriculum designed in this paper is with reasonable structure,complete system,progressive content,and salient feature,which provides the strong support for cultivating high-qualified software talents in line with national strategies and industry needs.
基金supported by the Fundamental Research Funds for the Central Universities under Grant 2020RC011.
文摘"Semester Training"has been adopted as an important part of the personnel training in software engineering majors since it was first put forward.The ultimate goal of semester training is to improve the professional quality of students in an all-round way,then eventually achieve the goal of satisfactory employment for both students and enterprises.However,in order to achieve the above purpose,the design of traditional training project still has the following problems:the topic selection of traditional training is designed by teachers in college,which lacks the training of engineering ability aiming at practical problems;the content and technology of traditional project training are out of date,ignoring the urgent demand of software industry development for advanced technology application;the traditional project training inspects the mastery of knowledge in each semester Degree,ignores the incremental of a progressive training system.In view of the above problems,this study proposes an Application-Oriented Software Engineering Semester Training System.Practice has proved that the construction of the training system can effectively improve the quality of teaching,so as to further improve the comprehensive quality of students.
文摘In view of the increasingly rapid development of global economic integration and combined with the existing modes of training international software engineering talents in China,this paper deeply analyzes and obtains the existing problems in the current teaching process,and proposes various teaching reform measures under the guidance of CDIO higher engineering education thought.Through many years of teaching practice experience,we can find that our reform has achieved remarkable results.
基金supported by the National Natural Science Foundation of China(Grant Nos.62272120,62106030,U20B2046,62272119,61972105)the Technology Innovation and Application Development Projects of Chongqing(Grant Nos.cstc2021jscx-gksbX0032,cstc2021jscxgksbX0029).
文摘In recent years,with the great success of pre-trained language models,the pre-trained BERT model has been gradually applied to the field of source code understanding.However,the time cost of training a language model from zero is very high,and how to transfer the pre-trained language model to the field of smart contract vulnerability detection is a hot research direction at present.In this paper,we propose a hybrid model to detect common vulnerabilities in smart contracts based on a lightweight pre-trained languagemodel BERT and connected to a bidirectional gate recurrent unitmodel.The downstream neural network adopts the bidirectional gate recurrent unit neural network model with a hierarchical attention mechanism to mine more semantic features contained in the source code of smart contracts by using their characteristics.Our experiments show that our proposed hybrid neural network model SolBERT-BiGRU-Attention is fitted by a large number of data samples with smart contract vulnerabilities,and it is found that compared with the existing methods,the accuracy of our model can reach 93.85%,and the Micro-F1 Score is 94.02%.
基金supported by National Key Research and Development Program of China (2019YFB2102500)China Postdoctoral Science Foundation (2021M700533)+1 种基金Natural Science Basic Research Program of Shaanxi Province of China (2021JQ-289,2020JQ-855)Social Science Fund of Shaanxi Province of China (2019S044).
文摘With the extensive integration of the Internet,social networks and the internet of things,the social internet of things has increasingly become a significant research issue.In the social internet of things application scenario,one of the greatest challenges is how to accurately recommend or match smart objects for users with massive resources.Although a variety of recommendation algorithms have been employed in this field,they ignore the massive text resources in the social internet of things,which can effectively improve the effect of recommendation.In this paper,a smart object recommendation approach named object recommendation based on topic learning and joint features is proposed.The proposed approach extracts and calculates topics and service relevant features of texts related to smart objects and introduces the“thing-thing”relationship information in the internet of things to improve the effect of recommendation.Experiments show that the proposed approach enables higher accuracy compared to the existing recommendation methods.
基金supported by the Fundamental Research Funds for the Central Universities(2021RC239)the Postdoctoral Science Foundation of China(2021 M690338)+3 种基金theHainan Provincial Natural Science Foundation of China(620RC562,2019RC096,620RC560)the Scientific Research Setup Fund of Hainan University(KYQD(ZR)1877)the Program of Hainan Association for Science and Technology Plans to Youth R&D Innovation(QCXM201910)the National Natural Science Foundation of China(61802092,62162021).
文摘With the expansion of network services,large-scale networks have progressively become common.The network status changes rapidly in response to customer needs and configuration changes,so network configuration changes are also very frequent.However,no matter what changes,the network must ensure the correct conditions,such as isolating tenants from each other or guaranteeing essential services.Once changes occur,it is necessary to verify the after-changed network.Whereas,for the verification of large-scale network configuration changes,many current verifiers show poor efficiency.In order to solve the problem ofmultiple global verifications caused by frequent updates of local configurations in large networks,we present a fast configuration updates verification tool,FastCUV,for distributed control planes.FastCUV aims to enhance the efficiency of distributed control plane verification for medium and large networks while ensuring correctness.This paper presents a method to determine the network range affected by the configuration change.We present a flow model and graph structure to facilitate the design of verification algorithms and speed up verification.Our scheme verifies the network area affected by obtaining the change of the Forwarding Information Base(FIB)before and after.FastCUV supports rich network attributes,meanwhile,has high efficiency and correctness performance.After experimental verification and result analysis,our method outperforms the state-of-the-art method to a certain extent.
基金the National Natural Science Foundation of China(Grant Nos.61972016 and 62032016)the Beijing Nova Program(Grant No.20220484106)。
文摘We present a novel approach for the prediction of crystal material properties that is distinct from the computationally complex and expensive density functional theory(DFT)-based calculations.Instead,we utilize an attention-based graph neural network that yields high-accuracy predictions.Our approach employs two attention mechanisms that allow for message passing on the crystal graphs,which in turn enable the model to selectively attend to pertinent atoms and their local environments,thereby improving performance.We conduct comprehensive experiments to validate our approach,which demonstrates that our method surpasses existing methods in terms of predictive accuracy.Our results suggest that deep learning,particularly attention-based networks,holds significant promise for predicting crystal material properties,with implications for material discovery and the refined intelligent systems.
基金the research outcomes of a blended top-tier undergraduate course in Henan ProvinceData Structures and Algorithms(Jiao Gao[2022]324)a research-based teaching demonstration course in Henan Province-Data Structures and Algorithms(Jiao Gao[2023]36)a model course of ideological and political education of Anyang Normal University-Data Structures and Algorithms(No.YBKC20210012)。
文摘To capitalize on the primary role of major course teaching and to facilitate students’understanding of abstract concepts in the data structure course,it is essential to increase their interest in learning and develop case studies that highlight fine traditional culture.By incorporating these culture-rich case studies into classroom instruction,we employ a project-driven teaching approach.This not only allows students to master professional knowledge,but also enhances their abilities to solve specific engineering problems,ultimately fostering cultural confidence.Over the past few years,during which educational reforms have been conducted for trial runs,the feasibility and effectiveness of these reform schemes have been demonstrated.
基金supported by the National Key Research and Development Program of China[grant number 2020YFA0608000]the National Natural Science Foundation of China[grant number 42075141]+2 种基金the Meteorological Joint Funds of the National Natural Science Foundation of China[grant number U2142211]the Key Project Fund of the Shanghai 2020“Science and Technology Innovation Action Plan”for Social Development[grant number 20dz1200702]the first batch of Model Interdisciplinary Joint Research Projects of Tongji University in 2021[grant number YB-21-202110].
基金supported by the National Natural Science Foundation of China(No.62176034)the Science and Technology Research Program of Chongqing Municipal Education Commission(No.KJZD-M202300604)the Natural Science Foundation of Chongqing(Nos.cstc2021jcyj-msxmX0518,2023NSCQ-MSX1781).
文摘Automatic crack detection of cement pavement chiefly benefits from the rapid development of deep learning,with convolutional neural networks(CNN)playing an important role in this field.However,as the performance of crack detection in cement pavement improves,the depth and width of the network structure are significantly increased,which necessitates more computing power and storage space.This limitation hampers the practical implementation of crack detection models on various platforms,particularly portable devices like small mobile devices.To solve these problems,we propose a dual-encoder-based network architecture that focuses on extracting more comprehensive fracture feature information and combines cross-fusion modules and coordinated attention mechanisms formore efficient feature fusion.Firstly,we use small channel convolution to construct shallow feature extractionmodule(SFEM)to extract low-level feature information of cracks in cement pavement images,in order to obtainmore information about cracks in the shallowfeatures of images.In addition,we construct large kernel atrous convolution(LKAC)to enhance crack information,which incorporates coordination attention mechanism for non-crack information filtering,and large kernel atrous convolution with different cores,using different receptive fields to extract more detailed edge and context information.Finally,the three-stage feature map outputs from the shallow feature extraction module is cross-fused with the two-stage feature map outputs from the large kernel atrous convolution module,and the shallow feature and detailed edge feature are fully fused to obtain the final crack prediction map.We evaluate our method on three public crack datasets:DeepCrack,CFD,and Crack500.Experimental results on theDeepCrack dataset demonstrate the effectiveness of our proposed method compared to state-of-the-art crack detection methods,which achieves Precision(P)87.2%,Recall(R)87.7%,and F-score(F1)87.4%.Thanks to our lightweight crack detectionmodel,the parameter count of the model in real-world detection scenarios has been significantly reduced to less than 2M.This advancement also facilitates technical support for portable scene detection.
基金supported by the Sichuan Science and Technology Program (No.2019YJ0356).
文摘Accurate automatic segmentation of gliomas in various sub-regions,including peritumoral edema,necrotic core,and enhancing and non-enhancing tumor core from 3D multimodal MRI images,is challenging because of its highly heterogeneous appearance and shape.Deep convolution neural networks(CNNs)have recently improved glioma segmentation performance.However,extensive down-sampling such as pooling or stridden convolution in CNNs significantly decreases the initial image resolution,resulting in the loss of accurate spatial and object parts information,especially information on the small sub-region tumors,affecting segmentation performance.Hence,this paper proposes a novel multi-level parallel network comprising three different level parallel subnetworks to fully use low-level,mid-level,and high-level information and improve the performance of brain tumor segmentation.We also introduce the Combo loss function to address input class imbalance and false positives and negatives imbalance in deep learning.The proposed method is trained and validated on the BraTS 2020 training and validation dataset.On the validation dataset,ourmethod achieved a mean Dice score of 0.907,0.830,and 0.787 for the whole tumor,tumor core,and enhancing tumor core,respectively.Compared with state-of-the-art methods,the multi-level parallel network has achieved competitive results on the validation dataset.
文摘Graphics Processing Units(GPUs)are used to accelerate computing-intensive tasks,such as neural networks,data analysis,high-performance computing,etc.In the past decade or so,researchers have done a lot of work on GPU architecture and proposed a variety of theories and methods to study the microarchitectural characteristics of various GPUs.In this study,the GPU serves as a co-processor and works together with the CPU in an embedded real-time system to handle computationally intensive tasks.It models the architecture of the GPU and further considers it based on some excellent work.The SIMT mechanism and Cache-miss situation provide a more detailed analysis of the GPU architecture.In order to verify the GPU architecture model proposed in this article,10 GPU kernel_task and an Nvidia GPU device were used to perform experiments.The experimental results showed that the minimum error between the kernel task execution time predicted by the GPU architecture model proposed in this article and the actual measured kernel task execution time was 3.80%,and the maximum error was 8.30%.
基金Project supported by the National Basic Research Program of China(Grant No.2013CB328903)the Special Fund of 2011 Internet of Things Development of Ministry of Industry and Information Technology,China(Grant No.2011BAJ03B13-2)+1 种基金the National Natural Science Foundation of China(Grant No.61473050)the Key Science and Technology Program of Chongqing,China(Grant No.cstc2012gg-yyjs40008)
文摘Cascading failure can cause great damage to complex networks, so it is of great significance to improve the network robustness against cascading failure. Many previous existing works on load-redistribution strategies require global information, which is not suitable for large scale networks, and some strategies based on local information assume that the load of a node is always its initial load before the network is attacked, and the load of the failure node is redistributed to its neighbors according to their initial load or initial residual capacity. This paper proposes a new load-redistribution strategy based on local information considering an ever-changing load. It redistributes the loads of the failure node to its nearest neighbors according to their current residual capacity, which makes full use of the residual capacity of the network. Experiments are conducted on two typical networks and two real networks, and the experimental results show that the new load-redistribution strategy can reduce the size of cascading failure efficiently.
