检索结果-内蒙古大学图书馆

Cross-Scale Bilevel Aggregation for Multi-exposure Fusion via Conditional Generative Adversarial Network 9th

Cross-Scale Bilevel Aggregation for Multi-exposure Fusion v...

9th International Conference on Internet of Things, ICIOT 2024, Held as Part of the Services Conference Federation, SCF 2024

作者： Wang, Longchun Yu, Mali Zhang, Hai Yang, Taojun Leng, Qingming Dong, Xiwei Guo, Jingjuan Wang, Guangxing School of Computer and Big Data Science Jiujiang University Jiujiang332005 China School of Information Management Jiangxi University of Finance and Economics Nanchang330013 China Jiujiang Engineering Research Center for Collaborative Cyber Security Protection Jiujiang332005 China School of Electronic and Information Engineering Jiujiang University Jiujiang332005 China

ISBN: (纸本)9783031770029

The aim of multi-exposure fusion (MEF) is to generate a high-dynamic-range-like image from images captured by common cameras under different exposure settings. The existing generative adversarial network (GAN)-based MEF methods have achieved fair performance. However, when addressing extremely exposed images, these existing GAN-based MEF methods still face numerous limitations, such as detail loss and unnatural visual effects, partly due to their inability to sufficiently aggregate multiscale context information. To address these limitations, we propose a cross-scale bilevel aggregation-based conditional generative adversarial network (CBA-cGAN). The generator contains two subnetworks: a cross-scale aggregation GhostNetV2 (CSA-GhostNetV2) for capturing and integrating multiscale context information with long-range dependencies and a U-shaped local network (ULN) for extracting local features. To fully aggregate short-range and multiscale long- range dependencies, we propose the use of bilevel aggregation in CSA-GhostNetV2: intrablock and interblock aggregation. The intrablock aggregation scheme based on dilated convolution and GhostBlockV2 is designed to aggregate multiscale context information and long-range relationships. The interblock aggregation scheme is designed to balance local and global contextual information during the fusion process by combining features acquired from CSA-GhostNetV2 and the ULN. Additionally, to better align the visual effect with human perception, we take the average of the input images for the discriminator. We compare the performance of the proposed method with that of traditional and deep learning methods on two publicly available datasets in terms of six objective metrics. Extensive experiments are conducted to demonstrate that the proposed CBA-cGAN outperforms the existing state-of-the-art methods in retaining local details while preserving overall visual effects. © The Author(s), under exclusive license to Springer Nature Switzerla

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Generating Targeted Universal Adversarial Perturbation against Automatic Speech Recognition via Phoneme Tailoring

Generating Targeted Universal Adversarial Perturbation again...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Zhang, Yujun Chen, Yanqu Wang, Jiakai Hu, Jin Tao, Renshuai Liu, Xianglong State Key Laboratory of Complex & Critical Software Environment Beihang University China School of Computer Science and Engineering Beihang University China College of Computer Science Beijing University of Technology China Zhongguancun Laboratory China School of Computer and Information Technology Beijing Jiaotong University China Institute of Data Space Hefei Comprehensive National Science Center China

ISBN: (纸本)9798350368741

There is a growing concern about adversarial attacks against automatic speech recognition (ASR) systems. Although research into targeted universal adversarial examples (AEs) has progressed, current methods are constrained by inefficient exploitation of audio features, demonstrating insufficient attack ability and robustness in the physical world. To solve this problem, we propose a phoneme-tailored attack (PTA) to improve the quality of the generated AEs. Specifically, to improve attack ability, we propose a Diverse Audio Composition Enrichment method, which enhances the utilization of audio features through phoneme-level slicing and recombination. To adapt AEs to complex environments, we propose a Natural Noise Pattern Guidance method to align AEs with natural noise patterns to improve their robustness. Experiments show that our method achieves an average accuracy of more than 72.34% and 98% with and without a norm constraint, and also demonstrates excellent performance in terms of generalization across datasets and resilience to MP3 compression. © 2025 IEEE.

关键词： Automatic speech recognition natural noise pattern phoneme-level targeted universal adversarial examples

来源：评论

学校读者我要写书评

暂无评论

Language-guided Image Reflection Separation

Language-guided Image Reflection Separation

引用

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Haofeng Zhong Yuchen Hong Shuchen Weng Jinxiu Liang Boxin Shi National Key Laboratory for Multimedia Information Processing School of Computer Science Peking University National Engineering Research Center of Visual Technology School of Computer Science Peking University AI Innovation Center School of Computer Science Peking University

ISBN: (数字)9798350353006

ISBN: (纸本)9798350353013

This paper studies the problem of language-guided re-flection separation, which aims at addressing the ill-posed reflection separation problem by introducing language de-scriptions to provide layer content. We propose a unified framework to solve this problem, which leverages the cross-attention mechanism with contrastive learning strategies to construct the correspondence between language descriptions and image layers. A gated network design and a ran-domized training strategy are employed to tackle the rec-ognizable layer ambiguity. The effectiveness of the pro-posed method is validated by the significant performance advantage over existing reflection separation methods on both quantitative and qualitative comparisons.

关键词： Training computer vision Annotations Natural languages Contrastive learning Logic gates Reflection

来源：评论

学校读者我要写书评

暂无评论

Optimizing Healthcare Delivery through CloudBased Clinical Decision Support Systems

Optimizing Healthcare Delivery through CloudBased Clinical D...

引用

2024 OPJU International Technology Conference on Smart Computing for Innovation and Advancement in Industry 4.0, OTCON 2024

作者： Ebenezar, U. Samson Jenila Vennila, G. Balakrishnan, T. Suresh Krishnan, Prabhakar Saveetha Institute of Medical and Technical Sciences Saveetha University Saveetha School of Engineering Department of Computer Science and Engineering Tamil Nadu Chennai India Care College of Engineering Department of Computer Science and Engineering Tamil Nadu Trichy India Mohan Babu University School of Computing Department of Ai & Ml Andhra Pradesh Tirupati India Saveetha Institute of Medical and Technical Sciences Saveetha University Saveetha School of Engineering Department of Computer Science and Engineering Chennai India Amrita Vishwa Vidyapeetham Center for Cyber Security Systems and Networks Amritapuri-campus India

ISBN: (纸本)9798350373783

The experiment aims at evaluating how four machine learning algorithms work: Decision Tree, Random Forest, SVM, and Neural Network, in a task, of binary classification. Models were evaluated using comprehensive metrics such as accuracy, precision, recall, F-score and AUC that revealed a good diagnostic performance. Models selected for the experiment and parameters tested on real datasets show steady changes in performance. The Neural Network model remained on top, achieving accuracy, precision, recall, F-score and AUC-ROC of 91.8%, 92.1%, 91.3%, 91.7% and 0.93 respectively. Random Forest displayed a competitive accordance along with the accuracy of 88.6%, precision of 89.2%, recall of 88.7%, and the F-score of 89.0%, and the AUC of 0.91. From the results, Support Vector Machine and Decision Tree models posted lower accuracy, recall, and fscores compared to the Neural Network and Random Forest models in most cases. These reveal the need for the improvement of the suitable machine learning model that comes from specific task needs as well as the consideration of each one of factors including accuracy, interpretability, and computational efficiency. Extensive and development of the model may result in more comprehensive understanding of how the model leads classifying of real-world applications such as medical diagnoses, fraud detection and target marketing. © 2024 IEEE.

关键词： Random forests

来源：评论

学校读者我要写书评

暂无评论

Visible Light Positioning Method for Indoor Line-of-Sight Scene Based on Time Series Coding 9

Visible Light Positioning Method for Indoor Line-of-Sight Sc...

引用

9th IEEE International Conference on data science in Cyberspace, DSC 2024

作者： Yu, Yonghao Guo, Yongde Zhao, Dawei Fu, Kexue Tang, Yongwei Wang, Qinyuan City University of Macau Faculty of Data Science China Qilu University of Technology Shandong Academy of Sciences Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center National Supercomputer Center in Jinan Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China Shandong University School of Mechanical Engineering Key Laboratory of High Efficiency and Clean Mechanical Manufacture Ministry of Education Jinan China

ISBN: (纸本)9798350391367

Existing indoor visible light positioning methods, such as those based on fingerprint database algorithms, are complex to construct and consume significant resources. In response to this issue, the text proposes an end-to-end indoor visible light positioning method that samples continuous signals at a small number of random coordinate positions in indoor environments. Through a deep learning method based on attention mechanism, a nonlinear mapping between the received signal and spatial position coordinates is constructed, and the time series properties in the received signal are encoded as a feature in the data. The method is further improved by combining residual networks to improve accuracy. Through simulation experiments, it has been verified that the deep learning based localization method described in this article can achieve sub meter level accuracy on small sample datasets, and it has been proven that time series features have a significant impact on localization accuracy in indoor visible light positioning algorithms based on deep learning. © 2024 IEEE.

关键词： Time series

来源：评论

学校读者我要写书评

暂无评论

FedMoS: Taming Client Drift in Federated Learning with Double Momentum and Adaptive Selection 42

FedMoS: Taming Client Drift in Federated Learning with Doubl...

引用

42nd IEEE International Conference on computer Communications, INFOCOM 2023

作者： Wang, Xiong Chen, Yuxin Li, Yuqing Liao, Xiaofei Jin, Hai Li, Bo Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Wuhan China Wuhan University School of Cyber Science and Engineering Wuhan China Hong Kong University of Science and Technology Department of Computer Science and Engineering Hong Kong

ISBN: (纸本)9798350334142

Federated learning (FL) enables massive clients to collaboratively train a global model by aggregating their local updates without disclosing raw data. Communication has become one of the main bottlenecks that prolongs the training process, especially under large model variances due to skewed data distributions. Existing efforts mainly focus on either single momentum-based gradient descent, or random client selection for potential variance reduction, yet both often lead to poor model accuracy and system efficiency. In this paper, we propose FedMoS, a communication-efficient FL framework with coupled double momentum-based update and adaptive client selection, to jointly mitigate the intrinsic variance. Specifically, FedMoS maintains customized momentum buffers on both server and client sides, which track global and local update directions to alleviate the model discrepancy. Taking momentum results as input, we design an adaptive selection scheme to provide a proper client representation during FL aggregation. By optimally calibrating clients' selection probabilities, we can effectively reduce the sampling variance, while ensuring unbiased aggregation. Through a rigid analysis, we show that FedMoS can attain the theoretically optimal O(T - 2/3) convergence rate. Extensive experiments using real-world datasets further validate the superiority of FedMoS, with 58%-87% communication reduction for achieving the same target performance compared to state-of-the-art techniques. © 2023 IEEE.

关键词： Momentum

来源：评论

学校读者我要写书评

暂无评论

A novel hybrid butterfly optimization algorithm for feature selection with sine cosine velocity in the high-dimensional classification data

引用

Journal of Intelligent and Fuzzy Systems 2024年第5-6期47卷 369-391页

作者： Zhang, Li Chen, Xiaobo Key Laboratory of Data Science and Intelligence Education Hainan Normal University Ministry of Education Haikou Hainan China School of Computer Engineering Jiangsu University of Technology Changzhou Jiangsu China Changzhou City Center Branch People's Bank of China Changzhou Jiangsu China Key Laboratory of Symbolic Computation and Knowledge Engineering Ministry of Education Jilin University Changchun China

aiming at the shortcomings of the traditional butterfly optimization algorithm in solving the high-dimensional classification feature selection problem, which has low convergence and is prone to fall into local optimal solutions, a new hybrid butterfly optimization algorithm is proposed, i.e., HBOA-SCV (A novel hybrid butterfly optimization algorithm with sine cosine velocity). The algorithm is applied to solve a high-dimensional classification feature selection problem. Firstly, the algorithm's global exploration and local exploitation ability can be dynamically balanced by introducing inertia weight coefficients w based on multiple learning strategies. Secondly, using the updated speed position formula of the sine-cosine acceleration strategy, individual butterflies' autonomous search ability and convergence speed can be further improved. Finally, according to the fitness value of each butterfly individual, the moving step length and direction of the butterfly individual are automatically adjusted better to fit the actual search process of the butterfly individual, increase the search ability in the global range, and avoid the algorithm from falling into the local optimum. To verify the algorithm's effectiveness, 18 high-dimensional classification numbers are selected to carry out simulation and comparison experiments between HBOA-SCV and traditional BOA algorithm, five improved BOA algorithms and other comparative algorithms for high-dimensional classification data successively. The experimental results show that the average fitness value and classification accuracy of the HBOA-SCV algorithm are better than the comparison algorithm, thus verifying the superiority of the HBOA-SCV algorithm. © 2024 - IOS Press. All rights reserved.

关键词： Optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

PPA: Preference Profiling Attack Against Federated Learning 30

PPA: Preference Profiling Attack Against Federated Learning

引用

30th Annual Network and Distributed System Security Symposium, NDSS 2023

作者： Zhou, Chunyi Gao, Yansong Fu, Anmin Chen, Kai Dai, Zhiyang Zhang, Zhi Xue, Minhui Zhang, Yuqing School of Computer Science and Engineering Nanjing University of Science and Technology China State Key Laboratory of Information Security Institute of Information Engineering Chinese Academy of Science China Data61 CSIRO Syndey Australia National Computer Network Intrusion Protection Center University of Chinese Academy of Science China

ISBN: (纸本)1891562835

—Federated learning (FL) trains a global model across a number of decentralized users, each with a local dataset. Compared to traditional centralized learning, FL does not require direct access to local datasets and thus aims to mitigate data privacy concerns. However, data privacy leakage in FL still exists due to inference attacks, including membership inference, property inference, and data inversion. In this work, we propose a new type of privacy inference attack, coined Preference Profiling A ttack (PPA), t hat accurately profiles t he p rivate p references o f a l ocal u ser, e .g., m ost liked (disliked) items from the client’s online shopping and most common expressions from the user’s selfies. I n g eneral, P PA can profile t op-k (i.e., k = 1, 2, 3 a nd k = 1 i n p articular) preferences contingent on the local client (user)’s characteristics. Our key insight is that the gradient variation of a local user’s model has a distinguishable sensitivity to the sample proportion of a given class, especially the majority (minority) class. By observing a user model’s gradient sensitivity to a class, PPA can profile the sample proportion of the class in the user’s local dataset, and thus the user’s preference of the class is exposed. The inherent statistical heterogeneity of FL further facilitates PPA. We have extensively evaluated the PPA’s effectiveness using four datasets (MNIST, CIFAR10, RAF-DB and Products-10K). Our results show that PPA achieves 90% and 98% top-1 attack accuracy to the MNIST and CIFAR10, respectively. More importantly, in real-world commercial scenarios of shopping (i.e., Products-10K) and social network (i.e., RAF-DB), PPA gains a top-1 attack accuracy of 78% in the former case to infer the most ordered items (i.e., as a commercial competitor), and 88% in the latter case to infer a victim user’s most often facial expressions, e.g., disgusted. The top-3 attack accuracy and top-2 accuracy is up to 88% and 100% for the Products-10K and RAF-DB, re

关键词：

来源：评论

学校读者我要写书评

暂无评论

Differentially Private and Heterogeneity-Robust Federated Learning With Theoretical Guarantee

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第12期5卷 6369-6384页

作者： Wang, Xiuhua Wang, Shuai Li, Yiwei Fan, Fengrui Li, Shikang Lin, Xiaodong Huazhong University of Science and Technology Hubei Key Laboratory of Distributed System Security Hubei Engineering Research Center on Big Data Security School of Cyber Science and Engineering Wuhan430074 China University of Electronic Science and Technology of China National Key Laboratory of Wireless Communications Chengdu611731 China Xiamen University of Technology Fujian Key Laboratory of Communication Network and Information Processing Xiamen361024 China University of Guelph School of Computer Science GuelphONN1G 2W1 Canada

Federated learning (FL) is a popular distributed paradigm where enormous clients collaboratively train a machine learning (ML) model under the orchestration of a central server without knowing the clients' private raw data. The development of effective FL algorithms faces multiple practical challenges including data heterogeneity and clients' privacy protection. Despite that numerous attempts have been made to deal with data heterogeneity or rigorous privacy protection, none have effectively tackled both issues simultaneously. In this article, we propose a differentially private and heterogeneity-robust FL algorithm, named DP-FedCVR to mitigate the data heterogeneity by following the client-variance-reduction strategy. Besides, it adopts a sophisticated differential privacy (DP) mechanism where the privacy-amplified strategy is applied, to achieve a rigorous privacy protection guarantee. We show that the proposed DP-FedCVR algorithm maintains its heterogeneity-robustness though DP noises are incorporated, while achieving a sublinear convergence rate for a nonconvex FL problem. Numerical experiments based on image classification tasks are presented to demonstrate that DP-FedCVR provides superior performance over the benchmark algorithms in the presence of data heterogeneity and various DP privacy budgets. © 2020 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Tensor and Minimum Connected Dominating Set based Confident Information Coverage Reliability Evaluation for IoT

IEEE Transactions on Sustainable Computing

引用

IEEE Transactions on Sustainable Computing 2024年第03期10卷 547-561页

作者： Xiao, Ziheng Zhu, Chenlu Feng, Wei Liu, Shenghao Deng, Xianjun Lu, Hongwei Yang, Laurence T. Park, Jong Hyuk Hubei Key Laboratory of Distributed System Security Hubei Engineering Research Center on Big Data Security School of Cyber Science and Engineering Huazhong University of Science and Technology Hubei Wuhan430074 China Artificial Intelligence and Intelligent Transportation Joint Technical Center of HUST and Hubei Chutian Intelligent Transportation Company Ltd United States Network and Industrial Control Information Security Technology Department China Nuclear Power Operation Technology Company Ltd. 1011 Xiongchu Avenue Hubei Wuhan430070 China The Department of Computer Science and Engineering Seoul National University of Science and Technology Seoul01811 Korea Republic of

Internet of Things (IoT) reliability evaluation contributes to the sustainable computing and enhanced stability of the network. Previous algorithms usually evaluate the reliability of IoT by enumenating the states of nodes and networks, which are difficult to handle IoT with hundreds of nodes because the computational cost. In this paper, a novel algorithm, TMCRA, is proposed to evaluate the reliability of IoT in complex network environment, which consider both coverage and connectivity. For coverage, TMCRA employs the Confident Information Coverage (CIC) model to divide the target area into independent grids and calculates the coverage rate. In terms of connectivity, TMCRA forming the Virtual Backbone Network (VBN) based on two proposed methods: TMA and MGIN, and evaluate connectivity by analyzing the VBN rather than the whole network. The TMA and MGIN are two algorithms for constructing Minimum Connected Dominant Sets (MCDS), which are suitable for different scale networks. Finally, based on the data of coverage and connectivity, TMCRA utilizes tensors for the unified modeling and representation of network structure, and calculates IoT reliability based on the tensors. Simulations are carried out for various sizes of IoT to show the advantages and effectiveness of the proposed approach in reliability evaluation. © 2016 IEEE.

关键词： data reliability

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：