检索结果-内蒙古大学图书馆

SSRN 2023年

作者： Li, Jia Wang, Chenxu Wang, Xiaochun Wang, Xiali School of Software Engineering Xi’an Jiaotong University Xi’an710049 China MOE Key Lab of Intelligent Network and Network Security Xi’an Jiaotong University China Department of Computer Science Chang’an University Xi’an China

Outlier detection is a critical method in data mining as it has a wide range of practical use in various fields. Spectral clustering has attracted much attention for its advantage of coping with the curse of dimensionality in high-dimensional spaces. This paper presents a new spectral analysis-based outlier detection (SAOD) framework. We first construct the $k$ nearest neighbors ($k$NN) graphs of each sample and compute the eigenvalues of the Laplacian matrices of these graphs. Our findings indicate that the distribution of eigenvalues for outliers differs from that of inliers. Based on this observation, we employ the multivariate Gaussian kernel density estimation to empirically estimate the eigenvalues distribution of a given dataset. We then introduce a new outlier factor, based on the eigenvalue probabilities of samples to assess their outlierness degree. The experimental results, obtained from both synthetic and real-world datasets, show that our spectral analysis-based outlier detection method outperforms the state-of-the-art approaches. © 2023, The Authors. All rights reserved.

关键词： Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Federated Learning for Diffusion Models

arXiv

引用

arXiv 2025年

作者： Peng, Zihao Wang, Xijun Chen, Shengbo Rao, Hong Shen, Cong School of Mathematics and Computer Science Nanchang University Nanchang330000 China School of Electronics and Information Technology Sun Yat-sen University Guangzhou510006 China School of Software Nanchang University Nanchang330000 China Charles L. Brown Department of Electrical and Computer Engineering University of Virginia CharlottesvilleVA22904 United States

Diffusion models are powerful generative models that can produce highly realistic samples for various tasks. Typically, these models are constructed using centralized, independently and identically distributed (IID) training data. However, in practical scenarios, data is often distributed across multiple clients and frequently manifests non-IID characteristics. Federated Learning (FL) can leverage this distributed data to train diffusion models, but the performance of existing FL methods is unsatisfactory in non-IID scenarios. To address this, we propose FEDDDPM—Federated Learning with Denoising Diffusion Probabilistic Models, which leverages the data generative capability of diffusion models to facilitate model training. In particular, the server uses well-trained local diffusion models uploaded by each client before FL training to generate auxiliary data that can approximately represent the global data distribution. Following each round of model aggregation, the server further optimizes the global model using the auxiliary dataset to alleviate the impact of heterogeneous data on model performance. We provide a rigorous convergence analysis of FEDDDPM and propose an enhanced algorithm, FEDDDPM+, to reduce training overheads. FEDDDPM+ detects instances of slow model learning and performs a one-shot correction using the auxiliary dataset. Experimental results validate that our proposed algorithms outperform the state-of-the-art FL algorithms on the MNIST, CIFAR10 and CIFAR100 datasets. Copyright © 2025, The Authors. All rights reserved.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Latent fingerprint enhancement and segmentation technique based on hybrid edge adaptive DTV model 2

Latent fingerprint enhancement and segmentation technique ba...

引用

2nd International Conference on Smart Computing and Electronic Enterprise, ICSCEE 2021

作者： Hilles, Shadi M S Liban, Abdilahi Miaikil, Othman A.M. Mahmoud Altrad, Abdullah Baker El-Ebiary, Yousef A. Hilles, Mohanad M Contreras, Jennifer Istanbul OKAN University Software Engineering Department Istanbul Turkey Hargeisa University Computer Science Department Hargeisa Somalia Al-Madinah International University Computer Science Department Kuala Lumpur Malaysia Universiti Sultan Zainal Abidin Faculty of Informatics and Computing Trengganu Malaysia University College of Applied Science Information Technology Department Gaza Palestine FEU Institute of Technology Department of Information Technology Manila Philippines

ISBN: (纸本)9781665432221

Image enhancement and segmentation is widely used for fingerprint identification and authorization in biometrics devices, criminal scene is most challenges due to low quality of fingerprint, the most significant efforts is to develop algorithm for latent fingerprint enhancement which become challenging problem due to the complex and existing problem for instance, developing algorithms of latent fingerprint is able to extract features of image blocks and removing overlapping and isolate the poor and noisy background. however, it's still challenging and interested problem specifically latent fingerprint enhancement and segmentation. The aim study of this paper is to propose latent fingerprint enhancement and segmentation based on hybrid model and Chan-Vese method for segmentation, in order to reduce low image quality and increase the accuracy of fingerprint. The desired characteristics of intended technique are adaptive, effective and accurate, hybrid model of edge adaptive direction achieves accurate latent fingerprint enhancement and segmentation, the target needs to improve feature detection and performance, this research has proposed system architecture of research method in fingerprint enhancement and segmentation where is the method content two stages, the first is normalization and second is reconstruction, using EDTV model is required for adaptive noise, in addition Chan-vase technique contributed for identification of fingerprint image features, the result and testing using RMSE with three categories of fingerprint images good, bad and ugly show better performance for all three categories, as well RMSE shows the average of good latent fingerprint before and after enhancement. Latent Fingerprint Enhancement and Segmentation Technique Based on Hybrid Model Edge Adaptive Directional Total Variation © 2021 IEEE.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

Stochastic Approximation Approaches to Group Distributionally Robust Optimization and Beyond

arXiv

引用

arXiv 2023年

作者： Zhang, Lijun Bai, Haomin Zhao, Peng Yang, Tianbao Zhou, Zhi-Hua National Key Laboratory for Novel Software Technology Nanjing University China School of Artificial Intelligence Nanjing University China Department of Computer Science and Engineering Texas A&M University College Station United States

This paper investigates group distributionally robust optimization (GDRO) with the goal of learning a model that performs well over m different distributions. First, we formulate GDRO as a stochastic convex-concave saddle-point problem, which is then solved by stochastic mirror descent (SMD) with m samples in each iteration, and attain a nearly optimal sample complexity. To reduce the number of samples required in each round from m to 1, we cast GDRO as a two-player game, where one player conducts SMD and the other executes an online algorithm for non-oblivious multi-armed bandits, maintaining the same sample complexity. Next, we extend GDRO to address scenarios involving imbalanced data and heterogeneous distributions. In the first scenario, we introduce a weighted variant of GDRO, enabling distribution-dependent convergence rates that rely on the number of samples from each distribution. We design two strategies to meet the sample budget: one integrates non-uniform sampling into SMD, and the other employs the stochastic mirror-prox algorithm with mini-batches, both of which deliver faster rates for distributions with more samples. In the second scenario, we propose to optimize the average top-k risk instead of the maximum risk, thereby mitigating the impact of outlier distributions. Similar to the case of vanilla GDRO, we develop two stochastic approaches: one uses m samples per iteration via SMD, and the other consumes k samples per iteration through an online algorithm for non-oblivious combinatorial semi-bandits. Copyright © 2023, The Authors. All rights reserved.

关键词： Budget control

来源：评论

学校读者我要写书评

暂无评论

On the comparison of optimal NOMA and OMA in a paradigm shift of emerging technologies

TechRxiv

引用

TechRxiv 2022年

作者： Ghosh, Joydev Ra, In-Ho Singh, Saurabh Haci, Hüseyin Al-Utaibi, Khaled A. Sait, Sadiq M. School of Computer Science and Robotics Tomsk Polytechnic University Tomsk Russia School of Computer Information and Communication Engineering Kunsan National University Gunsan Korea Republic of Department of Industrial & Systems Engineering Dongguk University Korea Republic of Department of Electrical and Electronics Engineering Near East University Mersin 10 Turkey Computer Science and Software Engineering Department University of Ha’il Saudi Arabia Department of Computer Engineering King Fahd University of Petroleum & Minerals Dhahran Saudi Arabia

Non-orthogonal multiple access (NOMA) is a better multiple access technique than orthogonal multiple access (OMA), precisely orthogonal frequency division multiple access (OFDMA) scheme, at the conceptual level for fifth-generation (5G) networks and beyond 5G (B5G) networks. We investigate the potentials of the schemes by comparing the proposed NOMA scheme with the traditional cooperative communication NOMA (CCNOMA) scheme, rather than the comparison between NOMA and OMA only. To probe the effectiveness of NOMA as a multiple access technique, we propose a novel NOMA scheme considering two adjacent BSs with a special design of the transceiver architecture. The proposed scheme provides a reasonable data rate to both near user (NU) and far user (FU) without compromising the quality of service (QoS) to anyone of them. The conclusive analyses on the optimization framework of multi-user sum rate, capacity, transmit power, spectral efficiency (SE), and energy efficiency (EE) trade-off for NOMA and OFDMA schemes have been established to a succession of derivations. Under the analytical optimization framework, we also prove quite a few properties for them. Simulation results confirm the theoretical findings and show that the two schemes can efficiently approach the optimal power allocation, minimization of power consumption, and optimal SE-EE trade-off, and the proposed NOMA scheme provides comparatively better data sum rates than the baseline OMA scheme. © 2022, CC BY.

关键词： Energy efficiency

来源：评论

学校读者我要写书评

暂无评论

Understanding Dynamic Associations: Gait Recognition via Cross-view Spatiotemporal Aggregation Network

引用

IEEE Transactions on Circuits and Systems for Video Technology 2022年 1-1页

作者： Chen, Jianyu Wang, Zhongyuan Zheng, Caixia Zeng, Kangli Zou, Qin Xiong, Zixiang School of Computer Science National Engineering Research Center for Multimedia Software Wuhan University Wuhan China Key Laboratory of Applied Statistics of MOE Northeast Normal University Changchun China School of Computer Science Wuhan University Wuhan China Department of Electrical and Computer Engineering Texas A&M University College Station TX USA

Gait recognition is a technique to recognize subjects under long-distance conditions. Existing gait recognition methods generally rely on extracting different motion representations of body parts to capture the gait information. However, the differences in appearance silhouettes between subjects are not discriminative enough to identify the distinct gaits. Therefore, exploring the representations of unique motion differences of subjects from the temporal domain is an effective way to boost gait recognition performance. In this paper, we propose a cross-view spatiotemporal aggregation network (CSTAN) to learn the most discriminative motion representations of subjects from multi-scale features. Particularly, CSTAN uses temporal contextual feature aggregation (TCFA) block to interact multi-scale features, increasing the diversity of temporal representations in feature sequences and producing suitable motion representations. Besides, a potential spatial feature assessment (PSFA) block is proposed to address the problem of missing silhouettes due to occlusion. Considering the multi-view context, we propose a cross-view prediction learning (CVPL) block to mitigate the impact of the view itself. By aggregating features in the spatiotemporal and view domains, CSTAN can learn a more appropriate representation of human motion differences. Extensive experiments on two public gait recognition datasets demonstrate the state-of-the-art performance of our proposed CSTAN. It is worth mentioning that the performance of CSTAN achieves a breakthrough in some complex environments, e.g., the subjects wearing different coats or carrying a bag. IEEE

关键词： Data mining

来源：评论

学校读者我要写书评

暂无评论

Online Joint Power Allocation and Task Scheduling for LEO Satellite Networks

Online Joint Power Allocation and Task Scheduling for LEO Sa...

引用

IEEE Conference on Wireless Communications and Networking

作者： Zheyuan Li Lijun He Juncheng Wang Ziye Jia Yanting Wang Chau Yuen School of Software Northwestern Polytechnical University China School of Information and Control Engineering China University of Mining and Technology China Department of Computer Science Hong Kong Baptist University China Ministry of Industry and Information Technology Nanjing University of Aeronautics and Astronautics China School of Electrical and Electronics Engineering Nanyang Technological University Singapore

ISBN: (数字)9798350368369

ISBN: (纸本)9798350368376

The excessive proliferation of Low Earth Orbit (LEO) satellites inescapably bring the explosive growth of space data in LEO Satellite Networks (LSNs). Meanwhile, the stochastic arrivals of space data together with the time-varying satellite-ground links in LSNs pose significant challenges for offloading a large volume of space data from LSNs to ground stations. To circumvent these challenges, we systematically study the energy-constrained online data offloading problem to jointly optimize power allocation and task scheduling for LSNs. First, we leverage Lyapunov optimization to decouple our formulated long-term stochastic joint optimization problem into a set of per-time-slot subproblems. Then, each subproblem is decoupled into a task scheduling problem and a power allocation problem. Next, we derive the optimal solution to the power allocation problem and propose a multi-armed bandit based quasi-optimal solution to the task scheduling problem. Finally, extensive simulation results show that our proposed algorithm has superior performance over the state-of-the-art solutions.

关键词： Space vehicles Satellites Simulation Low earth orbit satellites Scheduling Explosives Resource management Optimization Space stations

来源：评论

学校读者我要写书评

暂无评论

Correction: Framework for Agent-Based Multistage Application Partitioning Algorithm in Mobile Cloud Computing

引用

SN computer science 2024年第5期5卷 1-1页

作者： Kanwal, Asia Amjad, Tehmina Ashraf, Humaira Department of Computer Science and Software Engineering International Islamic University Islamabad Pakistan

来源：评论

学校读者我要写书评

暂无评论

Adaptive Student Inference Network for Efficient Single Image Super-Resolution

Adaptive Student Inference Network for Efficient Single Imag...

引用

IEEE International Conference on Data Mining (ICDM)

作者： Kang Miao Zhao Zhang Jiahuan Ren Mingbo Zhao Haijun Zhang Richang Hong School of Computer Science and Information Engineering Hefei University of Technology Hefei China Key Laboratory of Knowledge Engineering with Big Data (Ministry of Education) & Intelligent Interconnected Systems Laboratory of Anhui Province Hefei University of Technology Hefei China Yunnan Key Laboratory of Software Engineering Kunming China School of Information Science and Technology Donghua University Shanghai China Department of Computer Science Harbin Institute of Technology (Shenzhen) Shenzhen China

Recent advances in single image super-resolution (SISR) have achieved remarkable performance through deep learning. However, the high computational cost hinders the deployment of SISR models on edge devices. Instead of proposing new SISR models, a new trend is emerging to improve network efficiency by reducing parameters, FLOPs, and inference time through slight modifications to the original models. However, recent methods usually focus on reducing only one of three metrics, i.e., FLOPs, parameters and inference time, which inevitably increases the other two metrics. In this paper, we propose a novel Adaptive Student Inference Network (ASIN) on popular SISR models, which aims at reducing FLOPs and inference time while maintaining the number of parameters and restoring clearer high-resolution images. Specifically, our ASIN divides a SISR model into three components (head, body and tail) and adopts various strategies for each part. For head and tail parts, to ensure the restored images contain more detailed information, a novel auxiliary Enhanced Teacher Network (ETNet) is designed, which is trained with the ground-truth images to obtain more prior knowledge to guide student network to extract more accurate textures using a new knowledge distillation method. For the body part, owing to the varying difficulties of the reconstructions in different regions, we propose an Adaptive Depth Predicted Module (ADPM) to dynamically shorten average depth of network to reduce the computational cost of overall network. Extensive experiments on two datasets demonstrate the effectiveness and state-of-the-art performance of our ASIN compared to its counterparts.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Design of an energy-efficient IOT device-assisted wearable sensor platform for healthcare data management

引用

Measurement: Sensors 2023年 30卷

作者： Ahamed, Bazeer Sellamuthu, Suseela Karri, Purushotam Naidu Srinivas, Inumarthi V. Mohammed Zabeeulla, A.N. Ashok Kumar, M. Department of IT University of Technology And Applied Sciences Al Musanna Oman Senior Grade School of Computing Sciences and Engineering Vellore Institute of Technology TamilNadu Chennai India Department of CSE Gayatri Vidya Parishad College of Engineering for Women Andhra Pradesh Visakhapatnam535546 India Mumbai400077 India Bangalore India Department of Computer Science and Software Engineering Skyline University Nigeria.Kano Nigeria

Advances in health-related behaviors, technical breakthroughs, and the spread of healthy living activities have expanded personal health evaluation on a broad scale. In medical applications, networks, data, and communication systems are extensively utilized and further expanding. Real-time applications for remote health monitoring generate massive amounts of data in a short period. Internet of Things (IoT) and cloud systems, resulting in unprocessed and inadequate data for end-users. Since cloud technology is a dispersed intermediary layer between the edge network and the cloud infrastructure, the delay is significantly decreased, and reliable communications can be established using cloud computing on an IoT-enabled healthcare device. This study presents an Energy Efficient Healthcare Data Management Method (EE-HDMM), which enhances health monitoring using an IoT-assisted wearable sensor infrastructure. The difficulty of constrained sources and the energy consumption is minimized by proposing the optimum energy resource allocation method. Wearable sensor technology supported by cloud computing is the ideal and dependable platform for powerful life-critical systems that are unlikely to have communication delays. Compared to conventional techniques, the suggested EE-HDMM has a high precision rate of 97%, a sensitivity rate of 94%, and a system efficiency of 96%. © 2023

关键词： Internet of things

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：