检索结果-内蒙古大学图书馆

Solving Multiobjective Combinatorial Optimization via Learning to Improve Method

IEEE Transactions on Emerging Topics in Computational Intelligence 2025年第3期9卷 2122-2136页

作者： Ye, Te Zhang, Zizhen Zhang, Qingfu Chen, Jinbiao Wang, Jiahai Sun Yat-sen University School of Computer Science and Engineering Guangzhou510275 China Sun Yat-sen University Guangdong Key Laboratory of Big Data Analysis and Processing Guangzhou510275 China City University of Hong Kong Department of Computer Science Kowloon Hong Kong

Recently, neural combinatorial optimization (NCO) methods have been prevailing for solving multiobjective combinatorial optimization problems (MOCOPs). Most NCO methods are based on the "Learning to Construct"(L2C) paradigm, where the trained model(s) can directly generate a set of approximate Pareto optimal solutions. However, these methods still suffer from insufficient proximity and poor diversity towards the true Pareto front. In this paper, following the "Learning to Improve"(L2I) paradigm, we propose weight-related policy network (WRPN), a learning-based improvement method for solving MOCOPs. WRPN is incorporated into multiobjective evolutionary algorithm (MOEA) frameworks to effectively guide the search direction. A shared baseline for proximal policy optimization is presented to reduce variance in model training. A quality enhancement mechanism is designed to further refine the Pareto set during model inference. Computational experiments conducted on two classic MOCOPs, i.e., multiobjective traveling salesman problem and multiobjective vehicle routing problem, indicate that our method achieves remarkable results. Notably, our WRPN module can be easily integrated into various MOEA frameworks such as NSGA-II, MOEA/D and MOGLS, providing versatility and applicability across different problem domains. © 2025 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

PFed-NS: An Adaptive Personalized Federated Learning Scheme Through Neural Network Segmentation

引用

IEEE Transactions on computers 2025年第6期74卷 1936-1948页

作者： Liu, Yuchen He, Ligang Zhang, Zhigao Ren, Shenyuan University of Warwick Department of Computer Science CoventryCV4 7AL United Kingdom Wuhan University School of Computer Science Wuhan430072 China China Hubei Luojia Laboratory Wuhan430072 China Beijing Jiaotong University Key Laboratory of Big Data & Artificial Intelligence in Transportation Ministry of Education School of Computer and Information Technology Beijing100044 China

Federated Learning (FL) is typically deployed in a client-server architecture, which makes the Edge-Cloud architecture an ideal backbone for FL. A significant challenge in this setup arises from the diverse data feature distributions across different edge locations (i.e., non-IID data). In response, Personalized Federated Learning (PFL) approaches have been developed. Network segmentation-based PFL is an important approach to achieving PFL, in which the training network is divided into a global segment for server aggregation and a local segment maintained client-side. Existing methods determine the segmentation before the training, and the segmentation remains fixed throughout the PFL training. However, our investigation reveals that model representations vary as PFL progresses and the fixed segmentation may not deliver best performance across various training settings. To address this, we propose PFed-NS, a PFL framework based on adaptive network segmentation. This adaptive segmentation technique is composed of two elements: a mechanism for assessing divergence of clients’ probability density functions constructed from network layers’ outputs, and a model for dynamically establishing divergence thresholds, beyond which server aggregation is deemed detrimental. Further optimization strategies are proposed to reduce the computation and communication costs incurred by divergence modeling. Moreover, we propose a divergence-based BN strategy to optimize BN performance for network segmentation-based PFL. Extensive experiments have been conducted to compare PFed-NS against recent PFL models. The results demonstrate its superiority in enhancing model accuracy and accelerating convergence. © 1968-2012 IEEE.

关键词： Probability density function

来源：评论

学校读者我要写书评

暂无评论

Potato Disease Identification Method based on Improved DeeplabV3+ Network 24

Potato Disease Identification Method based on Improved Deepl...

引用

7th International Conference on Image and Graphics Processing, ICIGP 2024

作者： Duan, Xiaoyue Gu, Shenming School of Information Engineering Zhejiang Ocean University China Key Laboratory of Oceanographic Big Data Mining & Application of Zhejiang Province Zhejiang Ocean University China

ISBN: (纸本)9798400716720

In recent years, the advancement of deep learning technology and convolutional neural networks has presented novel solutions for the swift and precise detection of crop diseases. To address issues such as low segmentation accuracy in plant leaf images amidst complex backgrounds, inadequacies in recognition and detection models, insufficient data sets, and sluggish training speed, we propose a method for potato leaf image segmentation and recognition based on an enhanced DeeplabV3+ network. Specifically, we substitute the model encoder's backbone feature extraction network with MobileNetV2 to significantly reduce computational requirements while enhancing calculation speed. Then, the attention mechanism module was incorporated into the backbone feature extraction network and decoder to further enhance the model's edge recognition capability and improve its segmentation accuracy. Simultaneously, this study devised a classification model to achieve potato leaf disease segmentation and detection amidst complex backgrounds. In the initial stage, an enhanced DeeplabV3+ was employed for segmenting potato leaves in intricate environments. Subsequently, ResNet101, ResNet50, and ShuffleNetV2 classification models were utilized in the second stage for accurately classifying potato leaves. Finally, the two-stage model was integrated to effectively segment and detect potato leaf diseases amidst intricate backgrounds. The experimental findings demonstrate that the enhanced DeeplabV3+ segmentation model achieves a remarkable Mean Intersection over Union (MIoU) of 82.6%, an impressive Average Pixel Accuracy (MPA) of 90.65%, and a commendable accuracy rate of 94.72% for the second stage classification model. In conclusion, this approach presents a novel research avenue for accurately segmenting and detecting potato leaf diseases in challenging background scenarios, thereby contributing to advancements in this field. © 2024 ACM.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Contribution of Transmembrane Protein 68 to Triglyceride Synthesis and Lipid Droplet Formation Differs From Diacylglycerol Acyltransferase

引用

生物化学与生物物理进展 2025年第3期52卷 705-715页

作者： YU Qing FU Yang-Yang ZENG Fan-Si PANG Hui-Min HUANG Fei-Fei CHANG Ping-An Chongqing Key Laboratory of Big Data for Bio-Intelligence School of Life Health Information Science and EngineeringChongqing University of Posts and TelecommunicationsChongqing 400065China

Objective To characterize transmembrane protein 68(TMEM68)in an alternative triacylglycerol(TAG)biosynthesis pathway,and determine the interplay between TMEM68 and the canonical TAG synthesis enzyme acyl-CoA:diacylglycerol acyltransferase(DGAT).Methods Effects of exogenous fatty acid and monoacylglycerol on TAG synthesis and lipid droplet(LD)formation in TMEM68 overexpression and knockout cells treated with DGAT inhibitor or not were investigated by comparing LD morphology,Oil Red O staining,and measurement of TAG *** were stained with fluorescence dye and observed by confocal fluorescence *** levels were determined with an enzyme-based triglyceride assay *** of TMEM68 and DGAT1 was detected by co-expression and confocal fluorescence microscopy and their interaction was determined by ***-qPCR and immunoblotting assay were used to detect the expression of *** The synthesis of TAG catalyzed by TMEM68 was independent of DGAT *** exogenous fatty acids and monoacylglycerol promoted TAG synthesis mainly through DGAT in human neuroblastoma *** LDs formed by TMEM68 were different in morphology from those by *** addition,TMEM68 and DGAT1 colocalized in the same endoplasmic reticulum(ER)compartment but did not interact ***68 overexpression reduced the expression of DGAT1,the major DGAT enzyme involved in TAG synthesis,while TMEM68 knockout had little *** The TMEM68-mediated TAG synthesis pathway has distinct features from the canonical DGAT pathway,however,TMEM68 and DGAT may coregulate intracellular TAG levels.

关键词： triacylglycerol lipid droplet TMEM68 GAT

来源：评论

学校读者我要写书评

暂无评论

Machine Learning for Signal Demodulation in Underwater Wireless Optical Communications

引用

China Communications 2024年第5期21卷 297-313页

作者： Ma Shuai Yang Lei Ding Wanying Li Hang Zhang Zhongdan Xu Jing Li Zongyan Xu Gang Li Shiyin Department of Strategic and Advanced Interdisciplinary Research Peng Cheng LaboratoryShenzhen 518055China School of Information and Control Engineering China University of Mining and TechnologyXuzhou 221116China Data-driven Intelligent Information System Laboratory Shenzhen Research Institute of Big DataShenzhen 518172China Ocean College Zhejiang UniversityZhoushan 316021China State Key Laboratory of Millimeter Waves Southeast UniversityNanjing 210096China

The underwater wireless optical communication(UWOC)system has gradually become essential to underwater wireless communication *** other existing works on UWOC systems,this paper evaluates the proposed machine learningbased signal demodulation methods through the selfbuilt experimental *** on such a platform,we first construct a real signal dataset with ten modulation ***,we propose a deep belief network(DBN)-based demodulator for feature extraction and multi-class feature *** also design an adaptive boosting(Ada Boost)demodulator as an alternative scheme without feature filtering for multiple modulated ***,it is demonstrated by extensive experimental results that the Ada Boost demodulator significantly outperforms the other *** also reveals that the demodulator accuracy decreases as the modulation order increases for a fixed received optical power.A higher-order modulation may achieve a higher effective transmission rate when the signal-to-noise ratio(SNR)is higher.

关键词： AdaBoost DBN machine learning signal demodulation

来源：评论

学校读者我要写书评

暂无评论

A DL Intrusion Detection Method for Low-Rate Denial of Service Attacks

A DL Intrusion Detection Method for Low-Rate Denial of Servi...

引用

2023 Asia-Pacific Conference on Image Processing, Electronics and computers, IPEC 2023

作者： Wu, Yuhong Lv, Yunshan Hu, Xiangdong School of Big Data & Computer Science Engineering Chongqing College of Mobile Communication China School of Computer Science and Technology Chongqing University of Posts and Telecommunications Key Laboratory of Public Big Data Security Technology Chongqing Chongqing China School of Big Data & Computer Science Engineering Chongqing College of Mobile Communication Key Laboratory of Public Big Data Security Technology Chongqing China School of Automation Chongqing University of Posts and Telecommunications Chongqing China

ISBN: (纸本)9781665457033

Denial of service (DoS) attacks is among the most harmful network threats. Low-rate DoS (LDoS) attacks feature good concealment and are likely to evade detection by conventional signal analysis-based methods. A LDoS attack detection method that combines the convolutional neural network and the gated recurrent unit is proposed in the present work. To evaluate the proposed method, we collect real user traffic from a university's website, label the traffic with LDoS attacks as normal traffic. In order to get the real attack traffic, this paper has collected some new LDoS attack programs, such as slowloris, slowhttptest, pwnlloris, torhammer and httplog, which can be obtained from public channels, and constructed a LAN environment that conformed to the working scenario of the program. By running the attack program, the attack traffic could be obtained. In a large sample scenario, the average detection rate of our proposed method can reach 98.68%. This article focuses on the research: the average detection rate of the detection method proposed in this article in the case that the security agency can only intercept a small number of attack samples. That is, in the small sample scenario, the detection of the collected traffic shows that the average detection rate of the proposed detection method can reach 89.54% in a small sample scenario. Its effective detection rate is obviously superior to the existing similar detection methods. © 2023 IEEE.

关键词： Denial-of-service attack

来源：评论

学校读者我要写书评

暂无评论

Identification of Core Technological Topics in the New Energy Vehicle Industry: The SAO-BERTopic Topic Modeling Approach Based on Patent Text Mining 5

Identification of Core Technological Topics in the New Energ...

引用

Joint Workshop of the 5th Extraction and Evaluation of knowledge Entities from Scientific Documents and the 4th AI + Informetrics, EEKE-AII 2024

作者： Zhu, Jianxin Chuang, Yutong Wang, Zhinan Li, Yunke Harbin Engineering University School of Economics and Management 150001 China Key Laboratory of Big Data and Business Intelligence Technology Ministry of Industry and Information Technology China

In the new energy vehicle industry, precise identification of core technologies is the key to promoting innovation and maintaining market competitiveness. In this article, a comprehensive approach combining information weight method and SAO-BERTopic topic model is proposed to extract and analyze core technologies from large-scale patent data. Through in-depth analysis of Incopat patent database, we use the information weight method to select high-quality core patents from four dimensions: technological, strategic, law and market value. These selected patents form the basis of the research data, which is then applied to the SAO-BERTopic model, which combines the advanced semantic understanding capabilities of BERTopic with the fine-structured characteristics of SAO analysis, greatly improving the efficiency and accuracy of identifying complex technical topics. This innovation of this research is not only applicable to the analysis of the technological development of the new energy vehicle industry, but also can provide valuable reference for other high-tech industries such as biomedicine, renewable energy and information technology. These fields also need to identify core technologies from a large number of patents. SAO-BERTopic's structured analysis framework can help these industries to insight into technology development trends, identify innovation opportunities, and provide data-driven decision support for enterprises, research institutions and governments, thus playing an important role in technology planning and market commercialization. © Copyright 2024 for this paper by its authors.

关键词： Commerce

来源：评论

学校读者我要写书评

暂无评论

Discriminative latent semantics-preserving similarity embedding hashing for cross-modal retrieval

引用

Neural Computing and Applications 2024年第18期36卷 10655-10680页

作者： Chen, Yongfeng Tan, Junpeng Yang, Zhijing Cheng, Yongqiang Chen, Ruihan School of Information Engineering Guangdong University of Technology Guangdong Guangzhou510006 China School of Electronic and Information Engineering South China University of Technology Guangdong Guangzhou510641 China Guangdong Provincial Key Laboratory of Intellectual Property and Big Data Guangdong Guangzhou510665 China Department of Computer Science and Technology University of Hull HullHU6 7RX United Kingdom

Recently, there has been a significant increase in interest in cross-modal hashing technology. For hash code learning, most previous supervision methods use label information to create a similarity matrix in a straightforward manner. However, there are still the following challenges: (1) The asymmetric similarity matrix method only considers the similarity of labels, and the discriminant constraint in Hamming space is ignored;(2) there are optimization errors between the cross-modal semantic correlation of the Hamming space and the nonlinearity of the feature space;and (3) the cross-modal hash matrix is in a dynamic state during the optimization process, and the hash code is easily disturbed and there is bit uncertainty. To this end, we propose the Discriminative Latent Semantics-preserving Similarity Embedding Cross-modal Hashing (DLSSECH) method. Specifically, to reduce the quantization error, we introduce a non-asymmetric similarity decomposition based on orthogonal rotation bias. It can decompose the bitwise correlation of the learned hash code to capture more discriminative semantic information and compact hash code. In addition, to capture the cross-modal semantic correlation of nonlinear feature transformation and reduce quantization error, we propose the latent correlation error matrices based on orthogonal rotation decomposition. The model maintains the maximum difference in semantic dependencies between the projected data and its semantic representation. Finally, we use a sparse common hash matrix and non-asymmetric similarity decomposition factors to solve the uncertainty of the dynamic changes of the hash code. The effectiveness of DLSSECH has demonstrated through experiments on four cross-modal retrieval datasets, where it outperformed some state-of-the-art methods. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Cross-modal image text retrieval based on Cross-optimization and Joint strategy

Cross-modal image text retrieval based on Cross-optimization...

引用

2024 International Conference on Image, Signal Processing, and Pattern Recognition, ISPP 2024

作者： Zheng, Wei Han, Na Hu, Yan Kang, Peipei Zhang, Yue Yang, Aqing Zhang, Lei School of Computer Science Guangdong Polytechnic Normal University China Guangdong Provincial Key Laboratory of Intellectual Property and Big Data China School of Computer Science and Technology Guangdong University of Technology China School of Information Engineering East University of Heilongjiang China

ISBN: (纸本)9781510680425

In the fields of computer vision and natural language processing, cross-modal retrieval is of great importance that cannot be ignored. In existing multi-granularity alignment methods, significant progress has been made by globally aligning images and sentences or locally aligning regions and words. However, most of these methods rely on attention mechanisms, which may be affected by noise and cannot balance the differences between modalities, resulting in suboptimal image-text alignment. In addition, training with attention mechanisms consumes a lot of computational resources, and the retrieval process is time-consuming. To address these challenges, we develop a novel multigranularity image-text alignment model, which we call DFVLM. First, we independently train the intra-modal encoder and cross-modal encoder to reduce interference between different modal encoders and better learn intra-modal and inter-modal information. Then, we propose a joint strategy to combine the intra-modal and inter-modal information to better capture the information between image and text. In addition, we introduce a hard negative pair-based method to train the performance of the intra-modal multi-granularity encoder without using attention mechanisms. Unlike existing methods, We treat image and text retrieval as a two-way process and delve into the intrinsic connections between the two matrices. We propose a cross-validation method to optimize the retrieval result based on common image-to-text alignment scenarios in daily life, i.e., during the image-to-text retrieval process, a text-to-image reverse verification is performed on the retrieved corresponding text. Through extensive qualitative experiments and analysis, our approach performs well on both Flickr30K and MSCOCO datasets and also significantly reduces the time required for testing. © 2024 SPIE.

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

Detecting the One-Shot Dummy Attack on the Power Industrial Control Processes With an Unsupervised data-Driven Approach

引用

IEEE/CAA Journal of Automatica Sinica 2023年第2期10卷 550-553页

作者： Zhenyong Zhang Yan Qin Jingpei Wang Hui Li Ruilong Deng the State Key Laboratory of Public Big Data and the College of Computer Science and Technology Guizhou UniversityGuiyang 550025China the School of Chemical and Biomedical Engineering Nanyang Technological UniversitySingapore 639798Singapore the State Key Laboratory of Industrial Control Technology and the College of Control Science and Engineering Zhejiang UniversityHangzhou 310027China IEEE

Dear Editor,Dummy attack(DA), a deep stealthy but impactful data integrity attack on power industrial control processes, is recently recognized as hiding the corrupted measurements in normal measurements. In this letter, targeting a more practical case, we aim to detect the oneshot DA, with the purpose of revealing the DA once it is ***, we first formulate an optimization problem to generate one-shot DAs. Then, an unsupervised data-driven approach based on a modified local outlier factor(MLOF) is proposed to detect them.

关键词： Dummy Power letter

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：