检索结果-内蒙古大学图书馆

An Improved Convolutional Neural Networks: Quantum Pseudo-Transposed Convolutional Neural Networks

IEEE Access 2025年 13卷 37108-37117页

作者： Hai, Li Liang, Chen Yaming, Hao Wenli, Yu Fengquan, Shi Shandong Technology and Business University School of Information and Electronic Engineering Yantai264005 China Shandong Technology and Business University School of Computer Science and Technology Yantai264005 China

Recent advancements in quantum machine learning have spurred the development of hybrid quantum-classical convolutional neural networks (HQCCNNs), which have demonstrated promising potential for image classification tasks. Building on the operational principles of classical transposed convolutional neural networks (CNNs), we introduce a novel quantum variant: the Quantum Pseudo-Transposed Convolutional Neural Network (QPTCNN). The QPTCNN adapts the concept of classical transposed CNNs to the quantum domain, leveraging a hybrid quantum-classical framework that combines a quantum convolutional layer with a classical fully connected layer. In the QPTCNN, the quantum convolutional layer emulates a transposed convolution operation, ensuring that the output feature map retains the same dimensions as the input image. This is accomplished using rotational angle encoding and a ring-structured quantum circuit, interconnected by two-qubit control gates such as CNOT and CRY gates, facilitating efficient quantum convolution. We evaluated the performance of the QPTCNN on the MNIST and Fashion-MNIST datasets, with two distinct versions of the model: Model A, which utilizes a CNOT-gate entanglement circuit, and Model B, which employs a CRY-gate entanglement circuit. The results demonstrate that both Model A and Model B achieve strong performance across the datasets. However, Model A outperforms Model B, achieving higher classification accuracy and lower loss compared to earlier models. These findings suggest that the QPTCNN is highly capable of learning and extracting relevant feature information from input images, making it well-suited for high-performance image classification tasks. This work represents a significant advancement in quantum-enhanced image classification. © 2013 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Design and Optimization of a Novel Multiheater Microhotplate With Thermal Crosstalk

引用

IEEE Sensors Journal 2025年第9期25卷 15893-15905页

作者： Yang, Youpeng Wei, Guangfen He, Aixiang Jiao, Shasha Zhang, Wei Gao, Cong Shandong Technology and Business University School of Information and Electronic Engineering Yantai264005 China Shandong Technology and Business University School of Computer Science and Technology Yantai264005 China

As demand grows for high-sensitivity and high-selectivity gas sensors in the Internet of Things (IoT) andenvironmental monitoring fields, metal oxide semiconductor(MOX) gas sensors based on microhotplates have attractedsignificant research interest due to their low-power consump-tion and high integration. While array designs can addressthe issue of low sensor selectivity, integrating multiple sen-sor elements may result in increased power *** overcome this limitation, this article proposes a novel mul-tiheater integrated microhotplate design based on thermalcrosstalk, integrating multiple heaters on a single micro-hotplate to generate encodable multiple temperature *** COMSOL finite element simulation, the effects ofheater sizes, geometry, and thickness on the performance of the microhotplate were systematically studied. The resultsindicate that with a heater radius of 100 μm, the microhotplate exhibits excellent temperature uniformity at a workingtemperature of 300 ◦C, with more than 70% of the heating area having a temperature gradient of less than 10 ◦***, the optimized heater geometry significantly reduces the mechanical deformation of the microhotplate,ensuring structural stability. Most notably, the thermal crosstalk effect effectively reduces the power consumption of themicrohotplate. At a working temperature of 300 ◦C, the dual-heater design consumes approximately 4.68 mW less powerthan the single-heater configuration. By introducing an etched window design, the thermal efficiency of the microhotplateis further enhanced, with a peak temperature increase of approximately 40 ◦C. The innovative multiheater microhotplatedesign based on thermal crosstalk achieves low-power temperature control across all heaters, laying a solid foundationfor the development and widespread adoption of high-performance portable gas sensors. © 2001-2012 IEEE.

关键词： Oxide semiconductors

来源：评论

学校读者我要写书评

暂无评论

MSA: Mamba Semantic Alignment Networks for Remote Sensing Change Detection

引用

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 2025年 18卷 10625-10639页

作者： Huang, Zhenyang Duan, Peng Yuan, Genji Li, Jinjiang Shandong Technology and Business University School of Information and Electronic Engineering Yantai264005 China Shandong Technology and Business University School of Computer Science and Technology Yantai264005 China

With the rapid advancement of Earth observation technologies, remote sensing change detection (CD) has become a crucial method for monitoring surface changes. It is widely used in areas, such as urban expansion, disaster assessment, and resource detection. Current deep learning-based CD methods typically extract feature information from remote sensing images through downsampling and then aggregate early features with deeper ones during upsampling. However, directly aggregating these features without addressing spatial misalignment due to resolution changes can compromise the accuracy of change detection. In addition, there is a need to address the challenge of inadequate long-range dependency modeling in image processing. To tackle these challenges, this article proposes Mamba semantic alignment networks (MSA) for remote sensing CD. MSA introduces the semantic offset correction block, which corrects spatial misalignment during feature aggregation by incorporating a learnable semantic offset map, thereby reducing classification errors caused by feature mismatches. Furthermore, MSA incorporates the global dependency enhancement block, leveraging the Mamba architecture and the lossless downsampling and reversibility of wavelet transforms to significantly enhance global feature modeling. We evaluated MSA on three datasets, and the experimental results demonstrate that MSA outperforms mainstream methods across all three datasets. © 2008-2012 IEEE.

关键词： Change detection

来源：评论

学校读者我要写书评

暂无评论

Deep Reinforcement Learning and Nash Equilibrium Game based task offloading optimization strategy for the IoV

引用

IEEE Sensors Journal 2025年第10期25卷 18384-18393页

作者： He, Chao Jiang, Wenhui Wang, Xing Wang, Wanting Wang, Jiacheng Xie, Xin Chongqing Three Gorges University School of Electronic and Information Engineering Chongqing404100 China Nanyang Technological University School of Computer Science and Engineering Nanyang Avenue 639798 Singapore Chongqing University of Posts and Telecommunications School of Automation Chongqing400065 China

With the rapid development of 5G, the amount of data generated by the Internet of Vehicles (IoV) is growing explosively, which affects the efficiency of vehicle task offloading. To solve this problem, this paper proposes a task offloading strategy based on Regional Roadside Unit Fusion (RRF) algorithm. Firstly, with the help of sensors to collect vehicle historical trajectory data as input, the Deep Reinforcement Learning (DRL) technique combined with Convolutional Neural Network (CNN) and path planning algorithm is used to accurately project the dynamic behavior of vehicles. Secondly, when a Roadside Unit Fusion (RSU) is unable to handle a task, the Automatic Tagging (AT) algorithm will tag the RSU, preparing for the optimization of subsequent RSU. Finally, the Nash Equilibrium Game (NEG) algorithm is used to optimize RSU resource allocation to maximize task offloading efficiency and reduce energy consumption. Experimental results show that the RRF algorithm proposed in this paper has significant advantages over the standard Full DRL algorithm and the Non-Orthogonal Multiple Access (NOMA) algorithm. In terms of task offloading delay, it reduces 11.2% and 9.6%, respectively, energy consumption is reduced by 12.7% and 8.2%, respectively, and the task offloading success rate is improved to 92%, which is significantly better than the comparison algorithms. The algorithm provides a better solution for the practical application of Mobile Edge Computing (MEC) in the IoV. © 2001-2012 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A Lightweight Fault Diagnosis Method Based on Knowledge Distillation Under Time-Varying Rotational Speeds

引用

IEEE Access 2025年 13卷 36374-36387页

作者： Yang, Xiwang Wang, Yarong Gao, Lele Luo, Jia Jing, Licheng Huang, Jinying Liu, Guangpu Yang, Chenfeng Shanxi University of Electronic Science and Technology School of Information and Communication Engineering Linfen041000 China North University of China School of Computer Science and Technology Taiyuan030051 China North University of China School of Mechanical and Electrical Engineering Taiyuan030051 China North University of China School of Energy and Power Engineering Taiyuan030051 China North University of China School of Mechanical Engineering Taiyuan030051 China

Intelligent fault diagnosis models for engineering must meet real-time requirements while satisfying high diagnosis rates. However, collecting a large amount of fault information of different working conditions in the engineering environment is difficult, and the lack of fault data is a problem that has been difficult to solve. This paper presents a lightweight fault diagnosis model for wind turbine gearboxes based on transfer learning and knowledge distillation. The model is designed to be lightweight and can be used under time-varying rotational speeds. First, TL-ResPConv based on transfer learning and partial convolution was designed as a teacher network model to train a lightweight student network model for fault diagnosis based on TL-ResPConv-KD. Then, the student model obtained through knowledge distillation training is used as the distilled fault diagnosis model. Finally, the performance of the different models is validated and compared on a laboratory simulated planetary gearbox dataset. The experiments prove that the proposed model outperforms other models in fault recognition, achieving a balance between lightweight and high diagnostic accuracy. This makes it a practical network with higher diagnostic accuracy for fault diagnosis under time-varying rotational speeds, meeting the demand for lightweight deployment of edge equipment. © 2013 IEEE.

关键词： Probability distributions

来源：评论

学校读者我要写书评

暂无评论

Unveiling the Superiority of Unsupervised Learning on GPU Cryptojacking Detection: Practice on Magnetic Side Channel-based Mechanism

引用

IEEE Transactions on Information Forensics and Security 2025年 20卷 4874-4889页

作者： Cao, Hangcheng Xu, Guowen He, Ziyang Shi, Shaoqing Xu, Shengmin Wu, Cong Ning, Jianting Hunan University College of Computer Science and Electronic Engineering China City University of Hong Kong Department of Computer Science Hong Kong University of Electronic Science and Technology of China School of Computer Science and Engineering Chengdu China Zhengzhou University School of Cyber Science and Engineering Zhengzhou China Zhengzhou University International College Zhengzhou China Fujian Provincial Key Laboratory of Network Security and Cryptology College of Computer and Cyber Security Fuzhou China Nanyang Technological University College of Computing and Data Science Singapore Wuhan University School of Cyber Science and Engineering Wuhan China City University of Macau Faculty of Data Science China

Ample profits of GPU cryptojacking attract hackers to recklessly invade victims’ devices, for completing specific cryptocurrency mining tasks. Such malicious invasion undoubtedly obstructs normal device usage and wastes computation resources. To resist the threat of GPU cryptojacking, existing works aim to timely detect and clear away it, by distinguishing the dissimilitude between it and legitimate applications. However, these detection mechanisms inappropriately rely on two conflict cornerstones, manifested in leveraging mutable samples of illegitimate cryptojacking to design supervision-based detection models requiring samples with stable patterns. This limitation compromises the practicability of existing detection mechanisms in the face of mutable cryptojacking samples. To fill the gap, we explore the superiority of unsupervised learning in handling this issue and further propose an unsupervised manner-enabled detection mechanism named MagInspector, only using legitimate applications’ magnetic signatures from GPU side channels for model construction. MagInspector innovates in training an unsupervised autoencoder network by an adversarial mode that well learns the stable signature patterns of legitimate applications, while incompatible with mutable cryptojacking ones. In the process of model training, we elaborately extract mutual energy cumulation distribution features to represent legitimate applications to overcome the impact of their inter-type differences. Meanwhile, a locality sensitive hashing-driven outlier removal algorithm is designed to enhance MagInspector’s robustness to the noise samples. Finally, extensive experiments are conducted on GPUs covering four generations of common NVIDIA architectures and two generations of AMD architectures;the results show that applying MagInspector to mutable cryptojacking signature detection achieves a significant average accuracy improvement of 25.5% and 17.8%, respectively. © 2005-2012 IEEE.

关键词： Unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

Deep Reinforcement Learning-Based Continuous Workflows Scheduling in Heterogeneous Environments

引用

IEEE Internet of Things Journal 2025年第10期12卷 14036-14050页

作者： Wang, Zhi Zhan, Wenhan Duan, Hancong Min, Geyong Huang, Hualong University of Electronic Science and Technology of China School of Computer Science and Engineering Chengdu611731 China University of Exeter College of Engineering Mathematics and Physical Sciences Department of Computer Science ExeterEX4 4QF United Kingdom

Workflow scheduling plays a critical role in optimizing completion time and throughput in distributed cloud environments, leveraging the parallelism of heterogeneous computing resources. However, existing workflow scheduling algorithms often fall short due to heuristic limitations and the challenges in adaptability within heterogeneous settings, leading to suboptimal scheduling solutions. In this paper, we present a novel deep reinforcement learning (DRL) framework tailored for continuous workflow scheduling in heterogeneous environments. First, we propose an intelligent scheduler that updates the policy network through interactions with a multi-tenant environment, triggered by scheduling events. Next, the framework incorporates a Graph Attention Network (GAT) and a self-attention MultiLayer Perceptron (MLP) to preprocess the workflow topology and embed dynamic features of ready tasks and available processors into the state input at each scheduling step. Additionally, a k-dimensional tree-based k-nearest neighbors (kNN) algorithm is employed to map the output action vector to a pair of executed ready task and processor, facilitating the transition from continuous to discrete action spaces and addressing challenges associated with dynamic action spaces. Experimental results demonstrate that our method converges effectively in continuous workflow scheduling scenarios and significantly outperforms the best-known methods in terms of average makespan and load balancing efficiency. © 2014 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Enhanced Structure-Aware Graph Embedding for Workflow Recommendation 3rd

Enhanced Structure-Aware Graph Embedding for Workflow Recomm...

引用

3rd International Conference on Machine Learning, Cloud Computing and Intelligent Mining, MLCCIM 2024

作者： Tang, Rui Zhao, Taiyin School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu611731 China

ISBN: (纸本)9789819624676

With the continuous development of informatization, the system operation in today's society is becoming increasingly efficient, which prompts individuals and teams to continuously optimize their workflow design and improve the efficiency of workflow modeling processes. To address these issues, we deeply explore the attribute information of workflows and apply it to workflow recommendations. This article proposes an enhanced Structure Aware Graph Embedding method (ESAGE) applied to workflow recommendation, which utilizes the node attribute information of the workflow graph to enhance the graph structure information, generate feature embeddings for the workflow to calculate the similarity between workflows. Based on this, we have achieved efficient recommendation in the workflow. In addition, this method is also applicable to generating feature embeddings for nodes in other graph-structured data. Finally, based on real datasets, we designed and conducted experiments, and the experimental results showed the effectiveness of this method in the practical process. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Graph embeddings

来源：评论

学校读者我要写书评

暂无评论

A Dynamic Symmetric Searchable Encryption Scheme for Rapid Conjunctive Queries 24th

A Dynamic Symmetric Searchable Encryption Scheme for Rapid ...

引用

24th International Conference on Algorithms and Architectures for Parallel Processing, ICA3PP 2024

作者： Zou, Zekun Gao, Ruirui Li, Xiong Wang, Xiaofen Zhang, Xiaosong School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu611731 China

ISBN: (纸本)9789819615445

Dynamic Symmetric Searchable Encryption (DSSE) enables secure storage of encrypted data on a cloud server while supporting dynamic updates and keyword searches. Two key security concepts, forward privacy and backward privacy, ensure that searchable encryption minimizes the exposure of sensitive information to the server. Currently, most existing searchable encryption schemes only support single-keyword queries, and reusing single-keyword searches to achieve conjunctive keyword queries will significantly decrease query efficiency. This paper proposes a novel approach to tackle this issue, which enhances the efficiency of conjunctive keyword queries by including all keyword information of documents within a single encrypted entry. Furthermore, by merging the previously updated keys into subsequent updated data, the client’s computational overhead during queries is reduced as well as the amount of state information that the client needs to maintain. Compared with ODXT, our scheme achieves less information leakage, higher query efficiency, and lower communication overhead. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

CO-LEACH: Cooperative Data Collection Protocol for Data-Heterogeneous WSNs 18th

CO-LEACH: Cooperative Data Collection Protocol for Data-Het...

引用

18th International Conference on Wireless Artificial Intelligent Computing Systems and Applications, WASA 2024

作者： Zhang, Xinyu Wu, Gaojie Fu, Luwei Mao, Wenliang Zhao, Zhiwei School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu611731 China

ISBN: (纸本)9783031714696

Wireless Sensor Networks (WSNs) are indispensable for providing crucial sensing and data collection services to IoT systems. The diversity of IoT applications leads to data heterogeneity, which has become a prevalent characteristic of WSNs. Most existing work can only cope with data-homogeneous WSNs, where the same policy for clustering and data processing is applied to all network nodes. However, in practical IoT systems, different policies are desired to support the diverse types of IoT data. The neglect of data heterogeneity may result in poor data collection performance or even non-functionality. In this paper, we propose a cooperative data collection protocol (CO-LEACH) for data-heterogeneous WSNs. Specifically, 1) To mitigate the impact of data heterogeneity, CO-LEACH introduces an efficient cooperative clustering algorithm, in which heterogeneous nodes jointly form clusters, nodes with a higher proportion of intra-cluster type are more likely to be elected as cluster heads, thereby improving data collection efficiency. 2) To cope with the heterogeneous data, CO-LEACH presents an adaptive reprogramming strategy, making cluster heads capable of processing heterogeneous data. Simulation experiments show that CO-LEACH makes a better energy balance of all devices than state-of-the-art protocols, and prolongs the network lifetime by 3–5 times in data-heterogeneous WSNs. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Data collection

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：