检索结果-内蒙古大学图书馆

Applying SVR-PPSO, SVR-SSO, and SVR-BBO to estimate california bearing capacity of stabilized pond ash using admixtures

引用

Journal of Intelligent and Fuzzy Systems 2024年第4期46卷 8311-8327页

作者： Teng, Wei Li, Yan Sun, Hongxing Chen, Haojie School of Computer Science and Software Engineering University of Science and Technology Liaoning Liaoning Anshan China School of Electronic and Information Engineering University of Science and Technology Liaoning Liaoning Anshan China School of Information Technology Nanchang Vocational University Jiangxi Nanchang China School of Intelligent Construction Luzhou Vocational and Technical College Sichuan Luzhou China

In the present study, three hybrid models include support vector regression-salp swarm optimization (SVR-SSO), support vector regression-biogeography-based (SVR-BBO), and support vector regression-phasor particle swarm optimization (SVR- PPSO) was applied to forecast pond ash's CBR value modified with lime sludge (LS) and lime (LI). In the developed models, five variables were selected as inputs. It can result that the developed integrated models have R2 bigger than 0.9952. It means the agreement between observed and forecasted values by hybrid models is mainly similar to represent the highest accuracy. In both the training and testing stages, PSO-SVR results from better performance than the BBO-SVR model, with R2, RMSE, MAE, and PI equal to 0.9983, 0.6439, 0.3181, and 0.0081 for training data, and 0.9975, 0.7319, 0.4135, and 0.0141 for testing data, respectively. So, by considering the OBJ index, the OBJ value for PSO-SVR is 12.966, lower than BBO-SVR at 16.9957. Therefore, the PSO-SVR model outperforms another model to estimate the CBR of pond ash modified with LI and LS, consequently being recognized as the proposed model that makes it to be used for practical applications. © 2024 - IOS Press. All rights reserved.

关键词： Lime

来源：评论

学校读者我要写书评

暂无评论

DMF: A Deep Multimodal Fusion-Based Network Traffic Classification Model

引用

Computers, Materials & Continua 2025年第5期83卷 2267-2285页

作者： Xiangbin Wang Qingjun Yuan Weina Niu Qianwei Meng Yongjuan Wang Chunxiang Gu Henan Key Laboratory of Network Cryptography Technology Information Engineering UniversityZhengzhou450001China School of Computer Science and Engineering University of Electronic Science and Technology of ChinaChengdu611731China

With the rise of encrypted traffic,traditional network analysis methods have become less effective,leading to a shift towards deep learning-based *** these,multimodal learning-based classification methods have gained attention due to their ability to leverage diverse feature sets from encrypted traffic,improving classification ***,existing research predominantly relies on late fusion techniques,which hinder the full utilization of deep features within the *** address this limitation,we propose a novel multimodal encrypted traffic classification model that synchronizes modality fusion with multiscale feature ***,our approach performs real-time fusion of modalities at each stage of feature extraction,enhancing feature representation at each level and preserving inter-level correlations for more effective *** continuous fusion strategy improves the model’s ability to detect subtle variations in encrypted traffic,while boosting its robustness and adaptability to evolving network *** results on two real-world encrypted traffic datasets demonstrate that our method achieves a classification accuracy of 98.23% and 97.63%,outperforming existing multimodal learning-based methods.

关键词： Deep fusion intrusion detection multimodal learning network traffic classification

来源：评论

学校读者我要写书评

暂无评论

A game incentive mechanism for energy efficient federated learning in computing power networks

引用

Digital Communications and Networks 2024年第6期10卷 1741-1747页

作者： Xiao Lin Ruolin Wu Haibo Mei Kun Yang School of Information and Communication Engineering University of Electronic Science and Technology of ChinaChengduChina Chengdu Tongfei Technology Co.LtdChengduChina

Computing Power Network(CPN)is emerging as one of the important research interests in beyond 5G(B5G)or *** paper constructs a CPN based on Federated Learning(FL),where all Multi-access Edge Computing(MEC)servers are linked to a computing power center via wireless *** this FL procedure,each MEC server in CPN can independently train the learning models using localized data,thus preserving data ***,it is challenging to motivate MEC servers to participate in the FL process in an efficient way and difficult to ensure energy efficiency for MEC *** address these issues,we first introduce an incentive mechanism using the Stackelberg game framework to motivate MEC ***,we formulate a comprehensive algorithm to jointly optimize the communication resource(wireless bandwidth and transmission power)allocations and the computation resource(computation capacity of MEC servers)allocations while ensuring the local accuracy of the training of each MEC *** numerical data validates that the proposed incentive mechanism and joint optimization algorithm do improve the energy efficiency and performance of the considered CPN.

关键词： Computing power network Federated learning Energy efficiency Stackelberg game Resource allocation

来源：评论

学校读者我要写书评

暂无评论

MPGA-YOLOv8: A Leaf Disease Detection Network Combining Multi-Scale Progressive Feature Network and Global Attention

IAENG International Journal of Computer Science

引用

IAENG International Journal of Computer science 2025年第2期52卷 436-446页

作者： Xu, Jiacheng Yang, Yonghui Chu, Maoxiang Student of School of Electronic Information University of Science and Technology Liaoning Anshan114051 China School of Control Science and Engineering University of Science and Technology Liaoning Anshan114051 China

Considering the diversity of natural conditions, including fog, low light, and strong light, as well as the impact of various diseases on leaves, we propose an improved apple leaf disease detection method based on the YOLOv8n model. This method first uses a Multi-Scale Progressive Feature Network(MSPN) as the neck network to integrate multiple morphological lesions and enhance the model’s information integration capability. Then, a Global Self Attention Module (GSAM) is applied at the end of the backbone to help the model capture spatial relationships in the image and reduce interference from extremely complex background conditions. Next, we replace the convolution blocks in the backbone with Receptive Field-Focused Convolution Block (RFFconv) to effectively improve the model’s recognition accuracy through shared receptive field weights. Finally, we add a small object detection layer for small lesions to enhance the model’s generalization performance. Experimental results show that the proposed MPGA-YOLOv8 model effectively detects six types of apple leaf diseases in complex backgrounds, including healthy samples, with an average mAP accuracy of 74.3%. Compared to classic models like SSD, Faster RCNN, YOLOv3 tiny, YOLOv5, and YOLOv6, the mAP performance improves by 20.9%, 38.4%, 9.8%, 10.0%, and 14.3%, respectively. This model enables rapid and accurate detection and recognition of apple leaf diseases, providing viable technical support and solutions for disease prevention and control. © (2025), (International Association of Engineers). All rights reserved.

关键词： Disease control

来源：评论

学校读者我要写书评

暂无评论

Towards Accurate Alzheimer's Disease Diagnosis: Integrating Focused Linear Attention in Deep Learning Frameworks 8

Towards Accurate Alzheimer's Disease Diagnosis: Integrating ...

引用

8th International Artificial Intelligence and Data Processing Symposium, IDAP 2024

作者： Sam, Francis Qin, Zhiguang Addo, Daniel Arhin, Joseph Roger Ayivi, Williams Kwabena, Sarpong Muoka, Gladys Wavinya University of Electronic Science and Technology of China School of Information and Software Engineering Chengdu610054 China University of Electronic Science and Technology of China School of Information and Communication Engineering Chengdu611731 China

ISBN: (纸本)9798331531492

The early stage and accurate diagnosis of Alzheimer's Disease (AD) in neuroimaging remains a significant challenge. We introduce an innovative deep learning framework that incorporates a Focused Linear Attention (FLA) module to enhance the diagnosis of AD through medical imaging in this study. Our approach the strength of linear and softmax attention processes to enable the model to detect structural abnormalities and subtle patterns in brain pictures indicating various phases of cognitive decline. Our model achieves an accuracy of 93.21 %, together with a sensitivity of 92.74 % and a precision of 94.01 % in differentiating between non-demented, mild to moderately demented, and badly demented individuals according to experimental data. Our framework enhances overall diagnostic performance in multiclass classification by 1.85 % when compared to traditional machine learning models and baseline deep learning architectures. The incorporation of FLA greatly enhances these improvements, transforming our model into a powerful resource for the early diagnosis and intervention in AD. The results highlight how advanced deep learning methods, particularly Focused Linear Attention, could transform the early detection of Alzheimer's Disease. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Steel Plate Surface Defects Classification Method using Multiple Hyper-planes Twin Support Vector Machine with Additional information

引用

engineering Letters 2023年第3期31卷 1016-1024页

作者： Maoxiang, Chu Zixuan, Zhai Liming, Liu Guanghu, Liu School of Electronic and Information Engineering in University of Science and Technology Liaoning China

Focusing on the actual situation of steel surface defects, a novel multiple hyper-planes twin support vector machine with additional information (MHTSVM) is proposed. Similar to twin support vector machine (TSVM), MHTSVM also uses the nonparallel hyper-planes to deal with the classification problem. Differently, the MHTSVM model is trained with expert samples and rough samples respectively, which makes the MHTSVM more applicable. In addition, MHTSVM introduces the gaussian weight information obtained with K-nearest neighbors (KNN) method, which makes MHTSVM deal with noise samples and reliable. Meanwhile, the within-class scatter informat ion is embedded into MHTSVM model. The scatter information can be used to balance the samples distribution in defect dataset. The above two kinds of additional information make the MHTSVM have perfect classification performance in accuracy. Lastly, the MHTSVM algorithm is used to classify six types of steel surface defects. And a series of numerical experiments are performed. The experimental results proved that the novel MHTSVM model has perfect ability, especially for corrupted defect samples. © 2023, International Association of Engineers. All rights reserved.

关键词： Surface defects

来源：评论

学校读者我要写书评

暂无评论

A deep Q-learning model for sequential task offloading in edge AI systems

引用

Intelligent and Converged Networks 2024年第3期5卷 207-221页

作者： Dong Liu Shiheng Gu Xinyu Fan Xu Zheng Department of Computer Science and Engineering University of Electronic Science and Technology of ChinaChengdu 610000China Shenzhen Institute for Advanced Study University of Electronic Science and Technology of ChinaShenzhen 515100China Department of Information and Software Engineering University of Electronic Science and Technology of ChinaChengdu 610000China

Currently,edge Artificial Intelligence(AI)systems have significantly facilitated the functionalities of intelligent devices such as smartphones and smart cars,and supported diverse applications and *** fundamental supports come from continuous data analysis and computation over these *** the resource constraints of terminal devices,multi-layer edge artificial intelligence systems improve the overall computing power of the system by scheduling computing tasks to edge and cloud servers for *** efforts tend to ignore the nature of strong pipelined characteristics of processing tasks in edge AI systems,such as the encryption,decryption and consensus algorithm supporting the implementation of Blockchain ***,this paper proposes a new pipelined task scheduling algorithm(referred to as PTS-RDQN),which utilizes the system representation ability of deep reinforcement learning and integrates multiple dimensional information to achieve global task ***,a co-optimization strategy based on Rainbow Deep Q-Learning(RainbowDQN)is proposed to allocate computation tasks for mobile devices,edge and cloud servers,which is able to comprehensively consider the balance of task turnaround time,link quality,and other factors,thus effectively improving system performance and user *** addition,a task scheduling strategy based on PTS-RDQN is proposed,which is capable of realizing dynamic task allocation according to device *** results based on many simulation experiments show that the proposed method can effectively improve the resource utilization,and provide an effective task scheduling strategy for the edge computing system with cloud-edge-end architecture.

关键词： edge computing task scheduling reinforcement learning Rainbow Deep Q-Learning(RainbowDQN)

来源：评论

学校读者我要写书评

暂无评论

Dynamic Resolution Guidance for Facial Expression Recognition 27th

Dynamic Resolution Guidance for Facial Expression Recogniti...

引用

27th International Conference on Pattern Recognition, ICPR 2024

作者： Wang, Songpan Li, Xu School of Information and Software Engineering University of Electronic Science and Technology of China Sichuan Chengdu610054 China

ISBN: (纸本)9783031781247

Facial expression recognition (FER) plays a crucial role in human-computer interaction and emotion analysis. However, recognizing expressions in low-resolution images remains a significant challenge. This paper introduces a practical method called Dynamic Resolution Guidance for Facial Expression Recognition (DRGFER) to effectively recognize facial expressions in images with varying resolutions without compromising the accuracy of the FER model. Our framework comprises two main components: the Resolution Recognition Network (RRN) and the Multi-Resolution Adaptation Facial Expression Recognition Network (MRAFER). The RRN determines the resolution of the input image, and the MRAFER assigns the image to the most suitable facial expression recognition network according to its resolution. We evaluated the performance of DRGFER on two widely used datasets, RAF-DB and FERPlus. The results demonstrate that our method maintains optimal model performance at each resolution and outperforms alternative resolution-handling approaches. The proposed framework exhibits robustness against variations in both resolution and facial expressions, offering a promising solution for real-world applications. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

Improving Chinese Emotion Classification Based on Bilingual Feature Fusion 27th

Improving Chinese Emotion Classification Based on Bilingual...

引用

27th International Conference on Pattern Recognition, ICPR 2024

作者： Lan, Haocheng Ou, Jie Wang, Zhaokun Tian, Wenhong School of Information and Software Engineering University of Electronic Science and Technology of China Sichuan Chengdu610054 China

ISBN: (纸本)9783031781186

The growing popularity of Chinese social media platforms such as Sina Weibo has created a large number of user generated text content, which is of great value for understanding public emotions. However, the existence of mixed languages in these texts, especially Chinese and English, and mixed expressions pose a major challenge to current emotion classification methods. To address these issues, we propose a Bilingual Feature Fusion Network (BFFN) that leverages the multilingual capabilities of pre-trained language models to enhance the semantic feature extraction of Chinese text. Additionally, we introduce a Bilingual Cross Attention Mechanism (BCAM) that utilizes emotional features as the primary factor to capture cross-lingual emotional information effectively. Furthermore, we employ a lightweight fine-tuning approach that combines Low-Rank Adaptation (LoRA) and Embedding Fine-tuning (LEF) to reduce the complexity of fine-tuning model weights for downstream tasks. Extensive experiments on various datasets demonstrate the superiority of our proposed method, outperforming state-of-the-art models like ERNIE by 1.43% in accuracy. Our work contributes to the advancement of emotion classification in the context of mixed-language communication culture and provides a practical solution for real-world applications. Our code has been published on the open source community Github (https://***/oujieww/BFFN). © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Vehicle Detection Algorithm in Complex Scenes Based on Improved YOLOv8

IAENG International Journal of Computer Science

引用

IAENG International Journal of Computer science 2025年第4期52卷 886-893页

作者： Yu, Qing Ouyang, Xinyu Su, Bochao Zhao, Nannan You, Hongman School of Electronic and Information Engineering University of Science and Technology Liaoning Liaoning Anshan114051 China School of Electronic and Information Engineering University of Science and Technology Liaoning Liaoning Anshan114051 China Institute of Future Technology Shenzhen Polytechnic University Nanshan District Shenzhen518000 China

An algorithm based on improved YOLOv8 for detecting densely small-scale vehicles in complex scenes is designed. Using YOLOv8s as a baseline model, the Global Attention Module (GAM) is first introduced into the backbone network of YOLOv8 to improve the ability to extract detailed features in complex scenes. Secondly, the Efficient RepGFPNASFF network structure is used instead of the original neck network and detection heads to enhance the fusion of shallow detail features and deeper high-level features, and weighting parameters are introduced to improve the network’s interest in dense, small-scale vehicles based on multilevel functionality. Then, an inner-SIoU loss function is adopted, and the size of the auxiliary border is controlled by adjusting the scale factor ratio, which improves the model’s detection performance for overlapping occluded vehicle targets. Finally, the robustness of the model in complex scenes is verified in different datasets, and trained and tested in a self-built Complex Vehicle Dataset (CVD) using transfer learning. The experimental results show that compared with the original model, the precision, recall and mAP@0.5 of the optimised model are improved by 4.3%, 6.2% and 6.0%, respectively, and its detection effect is significantly improved, which proves the effectiveness and reliability of the algorithm. © (2025.0), (International Association of Engineers). All Rights Reserved.

关键词： complex scenes Global Attention Mechanism (GAM) Improved YOLOv8 loss function vehicle detection

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：