检索结果-内蒙古大学图书馆

Joint Class-Balanced Client Selection and Bandwidth Allocation for Cost-Efficient Federated Learning in Mobile Edge Computing Networks

引用

IEEE Transactions on Mobile Computing 2025年第7期24卷 5681-5698页

作者： Tang, Jian Li, Xiuhua Li, Hui Li, Penghua Wang, Xiaofei Leung, Victor C. M. Chongqing University School of Big Data & Software Engineering Chongqing400000 China Chongqing University of Posts and Telecommunications School of Automation Chongqing400065 China Tianjin University College of Intelligence and Computing Tianjin300072 China Shenzhen MSU-BIT University Artificial Intelligence Research Institute Shenzhen518172 China Shenzhen University College of Computer Science and Software Engineering Shenzhen518060 China University of British Columbia Department of Electrical and Computer Engineering VancouverBCV6T1Z4 Canada

Federated Learning (FL) has significant potential to protect data privacy and mitigate network burden in mobile edge computing (MEC) networks. However, due to the system and data heterogeneity of mobile clients (MCs), client selection and bandwidth allocation is key for achieving cost-efficient FL in MEC networks with limited bandwidth. To address these challenges, we investigate the issue of joint client selection and bandwidth allocation for reducing the cost (i.e., latency and energy consumption) of FL training. We formulate the problem and decompose it into a holistic subproblem to reduce the number of rounds and a partial subproblem to reduce the costs of FL each round. We propose a joint class-balanced client selection and bandwidth allocation (CBCSBA) framework to address the whole problem. Specifically, for the holistic subproblem, CBCSBA combines MCs into groups, each having data distribution as close as possible to class-balanced distribution;For the partial subproblem, CBCSBA reduces costs by exploratively selecting a group and sequentially optimizing the latency and energy consumption of MCs within the group. Experimental results show that CBCSBA outperforms the baseline frameworks in reducing latency by 28.2% and energy consumption by 25.3% on average in the considered four datasets. © 2002-2012 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

A Novel Gooseneck Barnacle Optimizer Based Machine Learning Approach for Rainfall Prediction 5

A Novel Gooseneck Barnacle Optimizer Based Machine Learning ...

引用

5th International Conference on Sustainable Technologies for Industry 5.0, STI 2023

作者： Ahmed, Marzia Sulaiman, Mohd Herwan Rahman, Mostafijur Mohamad, Ahmad Johari Faculty of Electrical & Electronics Engineering Technology Pahang Malaysia Daffodil International University Department of Software Engineering Dhaka Bangladesh Green University of Bangladesh Department of Computer Science and Engineering Bangladesh

ISBN: (纸本)9798350394290

In dynamic meteorological prediction, accurate rainfall forecasting is a mystery. In a complex and dynamic natural environment with unpredictable sky movements, we propose an innovative methodology that forecasts weekly average rainfall patterns using the Gooseneck Barnacle Optimizer and the Least Squares Support Vector Machine (LSSVM). The significance of rainfall prediction is shown by its widespread implications, including public health. However, conventional single-model techniques and machine-learning methods must represent rainfall pattern changes, limiting our preparation. With a 2014-2018 dataset from reliable meteorological libraries, our research explores the possibility of this novel combination. The Gooseneck Barnacle Optimizer (GBO) serves as the bedrock of our methodology, introducing a novel evolutionary algorithm inspired by the intricate mating behaviors of gooseneck barnacles. GBO aptly captures the dynamic interplay of factors such as navigational sperm casting properties, food availability, food attractiveness, wind direction, and intertidal zone wave movement during mating, culminating in two crucial optimization stages: exploration and exploitation. In contrast to previous algorithms like the Barnacle Mating Optimizer (BMO), GBO stands out by more accurately emulating the unique mating behaviors of gooseneck barnacles. The prediction job is performed by integrating the Least Squares Support Vector Machines (LSSVM) algorithm with GBO as an objective function, using the optimized hyper-parameter values. The results suggest that the GBO algorithm exhibits superior performance compared to existing methodologies. This is accomplished by efficiently augmenting the original random population for a particular problem, resulting in a convergence towards the global optimum and producing significantly enhanced optimization outcomes. © 2023 IEEE.

关键词： Rain

来源：评论

学校读者我要写书评

暂无评论

Faithfulness Measurable Masked Language Models 41

Faithfulness Measurable Masked Language Models

引用

41st International Conference on Machine Learning, ICML 2024

作者： Madsen, Andreas Reddy, Siva Chandar, Sarath Mila Montreal Canada Computer Engineering and Software Engineering Department Polytechnique Montreal Montreal Canada Computer Science and Linguistics McGill University Montreal Canada Facebook CIFAR AI Canada Canada CIFAR AI Canada

A common approach to explaining NLP models is to use importance measures that express which tokens are important for a prediction. Unfortunately, such explanations are often wrong despite being persuasive. Therefore, it is essential to measure their faithfulness. One such metric is if tokens are truly important, then masking them should result in worse model performance. However, token masking introduces out-of-distribution issues, and existing solutions that address this are computationally expensive and employ proxy models. Furthermore, other metrics are very limited in scope. This work proposes an inherently faithfulness measurable model that addresses these challenges. This is achieved using a novel fine-tuning method that incorporates masking, such that masking tokens become in-distribution by design. This differs from existing approaches, which are completely model-agnostic but are inapplicable in practice. We demonstrate the generality of our approach by applying it to 16 different datasets and validate it using statistical in-distribution tests. The faithfulness is then measured with 9 different importance measures. Because masking is in-distribution, importance measures that themselves use masking become consistently more faithful. Additionally, because the model makes faithfulness cheap to measure, we can optimize explanations towards maximal faithfulness;thus, our model becomes indirectly inherently explainable. Copyright 2024 by the author(s)

关键词：

来源：评论

学校读者我要写书评

暂无评论

Lightweight Key Management Scheme Using Fuzzy Extractor for Wireless Mobile Sensor Network

引用

computers, Materials & Continua 2022年第4期71卷 1695-1712页

作者： Eid Rehman Ibrahima Kalil Toure Kashif Sultan Muhammad Asif Muhammad Habib Najam Ul Hasan Oh-Young Song Aaqif Afzaal Abbasi Department of Software Engineering Foundation UniversityIslamabad44000Pakistan Centre Informatique Universite Gamal Abdel Nasser deConakryGuinea Department of Software Engineering Bahria University Islamabad CampusIslamabadPakistan GDC Ahmad Karak Higher Education DepartmentKhyber PakhtunkhwaPakistan Department of Electrical and Computer Engineering Dhofar UniversitySalalahOman Department of Software Sejong UniversitySeoul05006Korea

The mature design of wireless mobile sensor network makes it to be used in vast verities of applications including from home used to the security *** such types of applications based on wireless mobile sensor network are generally using real time data,most of them are interested in real time communication directly from cluster head of cluster instead of a base station in cluster *** would be possible if an external user allows to directly access real time data from the cluster head in cluster wireless mobile sensor network instead of accessing data from base *** this leads to a serious security breach and degrades the performance of any security protocol available in this *** existing schemes for authentication and cluster key management for external users,exchange a number of messages between cluster head and base station to allow external to access real time data from the base station instead of cluster *** increase communication cost and delay in such real time access *** handle this critical issue in cluster wireless mobile sensor network,we propose a lightweight authentication and key management scheme using a fuzzy *** this scheme,any external user can access data directly from the cluster head of any cluster without the involvement of the base *** proposed scheme only uses the one-way hash functions and bitwise XOR operations,apart from the fuzzy extractor method for the user biometric verification at the login *** presented scheme supports scalability for an increasing number of nodes using *** proposed scheme increases the life-time of the network by decreasing the key pool size.

关键词： Fuzzy extractor user authentication key management cluster session key wireless mobile sensor network

来源：评论

学校读者我要写书评

暂无评论

Class and Data-Incremental Learning Framework for Baggage Threat Segmentation via Knowledge Distillation

引用

IEEE Access 2025年 13卷 95977-96000页

作者： Ammara Nasim Saad Mazhar Khan Anum Abdul Salam Arslan Shaukat Taimur Hassan Adeel M. Syed Muhammad Usman Akram Department of Computer and Software Engineering National University of Sciences and Technology Islamabad Pakistan Department of Electrical and Computer Engineering Abu Dhabi University Abu Dhabi United Arab Emirates Department of Software Engineering Bahria University Islamabad Pakistan

With air travel growing rapidly worldwide, airports are busier than ever, making efficient security screening a top priority. To meet this growing demand, various advanced X-ray baggage scanners have been deployed at airports worldwide. Researchers have proposed multiple automated threat detection systems to enhance security screening efficiency; however, automated baggage threat segmentation remains a complex task, especially in the context of multi-class threat detection across diverse datasets. Traditional deep learning models struggle with differentiating between multiple threat types and suffer from catastrophic forgetting when exposed to new classes. To address these limitations, we propose an incremental learning framework that enables the model to progressively learn new threat categories while retaining previously acquired knowledge. Our approach utilizes SegFormer as the backbone and introduces a custom loss function, combining mutual distillation loss, KL divergence, and cross-entropy loss, to enhance knowledge retention and adaptability. The model is trained sequentially on three publicly available datasets, SIXRAY, GDXRAY, and PIDRAY, enabling it to generalize effectively across diverse baggage imagery. Through extensive experiments, we demonstrate that our method outperforms state-of-the-art incremental learning techniques, achieving superior segmentation accuracy and knowledge retention. Furthermore, Grad-CAM visualizations and t-SNE plots provide interpretability, offering insights into the model’s learning behavior and class separability. The proposed framework establishes a scalable and adaptable solution for real-world security screening applications, enabling efficient threat detection without requiring model retraining from scratch.

关键词： Incremental learning Transformers Threat assessment Image segmentation Security Decoding Airports Adaptation models Semantic segmentation Training

来源：评论

学校读者我要写书评

暂无评论

Movie Recommendation System Using Community Detection Based on the Girvan–Newman Algorithm 14th

Movie Recommendation System Using Community Detection Based ...

引用

14th International Conference on computer Science and its Applications, CSA 2022 and the 16th KIPS International Conference on Ubiquitous Information Technologies and Applications, CUTE 2022

作者： Ugli, Sadriddinov Ilkhomjon Rovshan Park, Doo-Soon Kim, Daeyoung Yang, Yixuan Peng, Sony Siet, Sophort Department of Computer Software Engineering Soonchunhyang University Asan Korea Republic of Departmen of Software Convergence Soonchunhyang University Asan Korea Republic of

ISBN: (纸本)9789819912513

Recently, a lot of research on the movie recommendation system has been conducted. The most important element in the movie recommendation system is to recommend the most suitable movie for an individual. To increase this accuracy, this paper proposes a movie recommendation system using community detection based on boundaries. In particular, the demographic characteristics of users are used to establish implicit bonds between users. After building an implicit social network, we applied community detection based on the Girvan–Newman algorithm to discover potential communities. Mean Average Error (MAE) was applied to evaluate the accuracy of the proposed model. The method presented in this paper improves the performance of the recommendation system by considering the user's profile-based similarity. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Recommender systems

来源：评论

学校读者我要写书评

暂无评论

An LLM-Based Framework for Synthetic Data Generation 15

An LLM-Based Framework for Synthetic Data Generation

引用

15th IEEE Annual Computing and Communication Workshop and Conference, CCWC 2025

作者： Goyal, Mandeep Mahmoud, Qusay H. Ontario Tech University Department of Electrical Computer and Software Engineering OshawaONL1G 0C5 Canada

ISBN: (纸本)9798331507695

The demand for high-quality datasets is rapidly increasing across sectors such as healthcare, finance, and cybersecurity, yet challenges like data scarcity and privacy concerns persist. To address this, we introduce a framework for synthetic data generation that empowers users to create realistic datasets while maintaining privacy. The framework leverages fine-tuned Large Language Models (LLMs) and differential privacy techniques, including IBM's diffprivlib, to generate synthetic data that replicates real-world patterns without exposing sensitive information. A proof-of-concept platform has been constructed to facilitate seamless data generation and augmentation, making it particularly useful in scenarios where original datasets are inaccessible, scarce, or privacy-restricted. The platform supports the creation of datasets across five key categories, employing advanced methods to preserve data integrity while ensuring compliance with stringent privacy standards. By combining cutting-edge AI technologies with robust privacy-preserving techniques, this framework offers a practical solution for researchers and professionals seeking reliable synthetic data to drive innovation in data-sensitive fields. © 2025 IEEE.

关键词： Data integrity

来源：评论

学校读者我要写书评

暂无评论

Comparative Study of Iteration and Coefficient Influence on Volume Preservation in Position-based Dynamics Simulation 6

Comparative Study of Iteration and Coefficient Influence on ...

引用

6th IEEE International Conference on Knowledge Innovation and Invention, ICKII 2023

作者： Va, Hongly Ma, Jun Hong, Min Soonchunhyang University Department of Software Convergence Asan-si Korea Republic of Soonchunhyang University Department of Computer Software Engineering Asan-si Korea Republic of

ISBN: (纸本)9798350323535

We proposed a practical approach with position-based dynamics (PBD) to simulate the volume of a deformable body. A variety of materials were simulated using the general constraints of PBD such as stretch, bending, and volume constraints. The method for the constraint system based on PBD was fast, unconditionally stable, and controllable. However, it also had limitations and challenges that need to be addressed. The animation of the simulator required high-performance speed, therefore we used the parallelization technique of PBD since the multi-core system and massively parallel GPUs were more efficient than the naïve approach (serialize method). Since we chose Unity3D and OpenGL as the main platforms, we used the compute shader as the kernel to execute the program to solve the system of constraints. Even though the PBD-based approach was fast, the coefficients were inaccurate to maintain the object's volume. Therefore, we demonstrated the complexity in the calculation of GPU-based PBD compared to CPU-based PBD in each different case of constraint type correlated with the coefficients. We also evaluated the effectiveness of various constraint methods within the PBD approach by comparing the error in the volume of the deformable object. © 2023 IEEE.

关键词： Graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

Survey on digital twins for Internet of Vehicles:Fundamentals,challenges,and opportunities

引用

Digital Communications and Networks 2024年第2期10卷 237-247页

作者： Jiajie Guo Muhammad Bilal Yuying Qiu Cheng Qian Xiaolong Xu Kim-Kwang Raymond Choo School of Computer and Software Nanjing University of Information Science and TechnologyNanjing210044China Department of Computer Engineering Hankuk University of Foreign StudiesGyeonggi-do17035South Korea Jiangsu Hydraulic Research Institute Nanjing210044China Department of Information Systems and Cyber Security University of Texas at San AntonioSan AntonioTX78249-0631USA

As autonomous vehicles and the other supporting infrastructures(e.g.,smart cities and intelligent transportation systems)become more commonplace,the Internet of Vehicles(IoV)is getting increasingly *** have been attempts to utilize Digital Twins(DTs)to facilitate the design,evaluation,and deployment of IoV-based systems,for example by supporting high-fidelity modeling,real-time monitoring,and advanced predictive ***,the literature review undertaken in this paper suggests that integrating DTs into IoV-based system design and deployment remains an understudied *** addition,this paper explains how DTs can benefit IoV system designers and implementers,as well as describes several challenges and opportunities for future researchers.

关键词： Internet of vehicles Digital twin Simulation Traffic systems

来源：评论

学校读者我要写书评

暂无评论

Staging of Different Tumor by Utilizing Laser Guidance (@405 nm) in CT Scan Images Along with Statistical Analysis

引用

Lasers in engineering 2025年第4-6期59卷 309-325页

作者： Amjad, M. Ullah, H. Siddique, M.A. Majid, M.A. Aleem, A. Akhtar, M. Gilanie, G. Abdullah-Al-Wadud, M. Farhan, M. Akram, A. Batool, Z. Nazir, A. Biophotonics Imaging Techniques Laboratory Institute of Physics The Islamia University of Bahawalpur 63100 Pakistan Department of Computer Science Faculty of Computing The Islamia University of Bahawalpur Bahawalpur Pakistan Dept. of Software Engineering College of Computer and Information Sciences King Saud University Riyadh11543 Saudi Arabia School of Earth Sciences and Engineering Hohai University Nanjing China

Early detection of lung cancer, liver cancer (HCC), and pancreatitis is crucial for effective treatment. This study assessed the use of laser guidance CT scans for diagnosis and staging. 26 patients of CT scan which were collected from BVH and civil hospital Bahawalpur, Pakistan. Visible light with a 405 nm wavelength is used. Results showed staging of lung cancer (2 at stage 1, 2 at stage 2, 3 at stage 4) and HCC (stages 1 to 4a). Pancreatic diseases detected included acute pancreatitis (44%), chronic pancreatitis (22%), gallstones (22%), and pancreatic tumors (11%). Alkaline phosphatase, SGOT, and blood urea levels were notably elevated in all three cancers compared to controls. CT scans show promise in identifying cancer stages and diagnosing pancreatitis, aiding in tailored treatment. Statistical analysis shows that alkaline phosphatase emerges as a potential biomarker for these conditions. ©2025 Old City Publishing, Inc.

关键词： Image analysis

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：