检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Tang, Anke Yang, Enneng Shen, Li Luo, Yong Hu, Han Du, Bo Tao, Dacheng School of Computer Science Wuhan University China Northeastern University China School of Cyber Science and Technology Shenzhen Campus of Sun Yat-sen University China School of Information and Electronics Beijing Institute of Technology China College of Computing & Data Science Nanyang Technological University Singapore

Deep model merging represents an emerging research direction that combines multiple fine-tuned models to harness their specialized capabilities across different tasks and domains. Current model merging techniques focus on merging all available models simultaneously, with weight interpolation-based methods being the predominant approaches. However, these conventional approaches are not well-suited for scenarios where models become available sequentially, and they often suffer from high memory requirements and potential interference between tasks. In this study, we propose a training-free projection-based continual merging method that processes models sequentially through orthogonal projections of weight matrices and adaptive scaling mechanisms. Our method operates by projecting new parameter updates onto subspaces orthogonal to existing merged parameter updates while using an adaptive scaling mechanism to maintain stable parameter distances, enabling efficient sequential integration of task-specific knowledge. Our approach maintains constant memory complexity to the number of models, minimizes interference between tasks through orthogonal projections, and retains the performance of previously merged models through adaptive task vector scaling. Extensive experiments on CLIP-ViT models demonstrate that our method achieves a 5-8% average accuracy improvement while maintaining robust performance in different task orderings. © 2025, CC BY.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Deep Learning-Driven Throughput Maximization in Covert Communication for UAV-RIS Cognitive Systems

IEEE Open Journal of the Communications Society

引用

IEEE Open Journal of the Communications Society 2025年 6卷 4140-4155页

作者： Nhan Vo, Van Quoc Long, Nguyen Dang, Viet-Hung Dac Ho, Tu Tran, Hung Chatzinotas, Symeon Tran, Dinh-Hieu Sanguanpong, Surasak So-In, Chakchai Duy Tan University Faculty of Information Technology Da Nang550000 Viet Nam Duy Tan University Institute of Research and Development Da Nang550000 Viet Nam Norwegian University of Science and Technology Faculty of Information Technology and Electrical Engineering Department of Information Security and Communication Technology Trondheim7034 Norway National Economics University DATCOM Lab Faculty of Data Science and Artificial Intelligence College of Technology Hanoi12116 Viet Nam University of Luxembourg Interdisciplinary Centre for Security Reliability and Trust Esch-sur-Alzette4365 Luxembourg Kasetsart University Faculty of Engineering Department of Computer Engineering Bangkok10900 Thailand Khon Kaen University Applied Network Technology College of Computing Khon Kaen40002 Thailand

This paper examines a cognitive radio (CR) nonorthogonal multiple access (NOMA) system in which an unmanned aerial vehicle equipped with a reconfigurable intelligent surface (UAV-RIS) plays two roles: relaying and friendly jamming. The communication protocol has two phases. The first is an energy harvesting phase in which the UAV harvests radio frequency energy from a power beacon. In the second phase, a secondary transmitter (ST) simultaneously sends superimposed signals to secondary receivers (SRs) (a public SR and a covert SR) via NOMA with the assistance of the UAV-RIS. Then, a UAV warden and a UAV jammer launch a cooperative attack, in which the first adversary wiretaps the signals from the ST and UAV-RIS, whereas the second interferes with the SRs to force the ST to increase its transmit power. For improved secrecy, the UAV-RIS uses its harvested energy to combat the UAV warden. For this system, the secrecy performance is evaluated on the basis of the concept of covert communication. In particular, optimization algorithms are employed to maximize the covert SR throughput under outage probability and security constraints. A deep neural network model is subsequently trained to discover the relationships between the environmental parameters and optimized parameters to enable rapid adaptation to environmental conditions. © 2020 IEEE.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

High-Fidelity Editable Portrait Synthesis with 3D GAN Inversion

High-Fidelity Editable Portrait Synthesis with 3D GAN Invers...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Jindong Xie Jiachen Liu Yupei Lin Jinbao Wang Xianxu Hou Linlin Shen Computer Vision Institute School of Computer Science and Software Engineering Shenzhen University National Engineering Laboratory for Big Data System Computing Technology Shenzhen University School of Information Engineering Guangdong University of Technology Guangzhou China Guangdong Provincial Key Laboratory of Intelligent Information Processing School of AI and Advanced Computing Xi’an Jiaotong-Liverpool University

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

The 3D generative adversarial network (GAN) inversion converts an image into 3D representation to attain high-fidelity reconstruction and facilitate realistic image manipulation within the 3D latent space. However, previous approaches face challenges regarding the trade-off between the reconstruction ability and editability. That is, reversing a real-world image to a low-dimensional latent code would inevitably lead to information loss, and achieving a near-perfect reconstruction using high-rate triplane representation often limits the ability to manipulate the image freely in the latent space. To address these issues, we propose a novel latent conditioning encoder-based framework with the alignment between the low-dimensional latent and high-dimensional triplane. A non-semantic guided editing strategy bridges the intrinsic relation between the latent condition and triplane generation, making it possible to edit the high-dimensional representation by latent manipulation. As a result, our method can achieve high-fidelity reconstruction and editing simultaneously by directly controlling the latent code. Experimental results demonstrate that our approach excels in reconstruction and editing quality compared to previous 3D inversion methods. Furthermore, our method can also edit even real faces with large poses and out-of-domain cases.

关键词： Bridges Three-dimensional displays Codes Signal processing Generative adversarial networks Vectors Acoustics Speech processing Image reconstruction Faces

来源：评论

学校读者我要写书评

暂无评论

Leveraging Instruction-based Learning for Effective UAV Traffic Management in Dynamic Environment

Leveraging Instruction-based Learning for Effective UAV Traf...

引用

Multi-Agent Systems for Collaborative Intelligence (ICMSCI), International Conference on

作者： R.S. Ramya Keren Lois Daniel S. Rajkumar Sruthi S Madhavan Akshya. J Mani Deepak Choudhry Department of Computer Science and Engineering Sri Ramakrishna Engineering College Coimbatore Department of Computer Science and Business Systems KGiSL Institute of Technology Coimbatore Department of Information Technology Sona College of Technology Salem Department of Artificial Intelligence and Data Science Nehru Institute of Engineering and Technology Coimbatore Department of Computational Intelligence SRM Institute of Science and Technology Chennai Department of Computing Technologies SRM Institute of Science and Technology Chennai

ISBN: (数字)9798331509828

ISBN: (纸本)9798331509835

With the popularization of UAVs (Unmanned Aerial Vehicles) in surveillance, delivery services, environmental monitoring, and increasingly other operational aspects, besides the need to take off and land, efficient traffic management becomes integral for assuring and securing further development and exploitation. Most of the existing traffic management systems for UAVs depend on traditional rule-based algorithms or simpler machine learning techniques, such as DQN or PPO. However, these systems also have enormous challenges: dynamic environments, high-density traffic, and robust decision-making over unpredictable weather and network conditions. In addition, the above traditional models usually experience difficulties in balancing exploration and exploitation, leading to suboptimal routing and enlarged collision risks in complicated urban airspaces. These challenges have driven us to develop a new proposal: a UA V traffic management framework based on Double Deep Q-Network (DDQN) combined with Instruction-Based Learning. In this approach, advanced learning for DDQN optimizes processes at decision-making points for UAVs. The model proposed herein, along with comprehensive simulations and quantitative analyses, demonstrated higher performance in terms of leadership and cumulative rewards while keeping lower training loss compared to the classical algorithms such as DQN, A3C, and PPO. At the same time, the DDQN model presented a better success rate (95%), as also reflected in the balanced precision-recall curve showing an area under the curve of 0.90, over others in the on-point prediction of UAVstate and safety margin of collision. Our experimental results demonstrate that the proposed DDQN-based framework significantly enhances UA V traffic management by optimizing energy consumption in complex airspaces and reducing the risk of collision.

关键词： Training Analytical models Adaptation models Statistical analysis Atmospheric modeling Computational modeling Decision making Predictive models Autonomous aerial vehicles Safety

来源：评论

学校读者我要写书评

暂无评论

Empowering autonomous IoT devices in blockchain through gasless transactions

引用

Blockchain: Research and Applications 2025年第2期6卷

作者： Madhwal, Yash Yanovich, Yury Korotkevich, Aleksandra Parshina, Daria Seropian, Nshteh Gavrilov, Stepan Nikolaev, Alex Balachander, S. Murugan, A. Skolkovo Institute of Science and Technology Moscow121205 Russia Faculty of Computer Science HSE University Moscow109028 Russia Faculty of Information and Communication Technologies ITMO University St. Petersburg197101 Russia Moscow Institute of Physics and Technology Moscow141701 Russia Department of Data Science and Business Systems School of Computing SRM Institute of Science and Technology Kattankulathur603203 India

The article introduces a proof-of-concept (PoC) that demonstrates the management of Internet of Things (IoT) devices' infrastructure via smart contracts, facilitating their interaction with the blockchain through gasless transactions. The focus is empowering IoT devices to autonomously sign transactions using their verified private keys, eliminating the necessity for external wallets and enabling blockchain interaction using Biconomy without incurring gas fees. In this PoC, managers can validate IoT devices, permitting them to transmit transactions securely without being able to manipulate measurements or risking losing crypto assets in case of hardware malfunctions. This innovative method ensures that devices with minimal funds can access sensor data and communicate with a smart contract on the blockchain to update information utilizing account abstraction. Detailed workflow and simulation results are provided to showcase the practicality and advantages of this approach in scenarios demanding seamless automated blockchain engagement through IoT devices. The PoC code is openly accessible on GitHub, enhancing the transparency and accessibility of our research outcomes. © 2025 The Author(s)

关键词： Account abstraction Blockchain Gasless transaction Internet of Things Mobile application Smart contract

来源：评论

学校读者我要写书评

暂无评论

Joint Task Partitioning and Resource Allocation in UAV-Enabled Vehicular Edge computing Based on Deep Reinforcement Learning

引用

IEEE Internet of Things Journal 2025年第11期12卷 15453-15466页

作者： Liang, Hongbin Zhang, Han Ale, Laha Hong, Xintao Wang, Lei Jia, Qiong Zhao, Dongmei Southwest Jiaotong University School of Transportation and Logistics National United Engineering Laboratory of Integrated and Intelligent Transportation National Engineering Laboratory of Integrated Transportation Big Data Application Technology Chengdu611756 China Southwest Jiaotong University School of Computing and Artificial Intelligence Chengdu611756 China Southwest Jiaotong University School of Information Science and Technology Chengdu611756 China Chinese Academy of Sciences Laboratory of Noise and Vibration Research Institute of Acoustics Beijing100190 China McMaster University Department of Electrical and Computer Engineering HamiltonONL8S 4K1 Canada

Vehicle Edge computing (VEC) leverages compact cloud computing at the mobile network edge to meet the processing and latency needs of vehicles. By bringing computation closer to the vehicles, VEC reduces data transmission, minimizes latency, and boosts performance for compute-intensive applications. However, during peak hours of urban road traffic, the scarce computational resources available at edge servers could pose challenges in fulfilling the processing needs of vehicles. Introducing Unmanned Aerial Vehicles (UAVs) as supplementary edge computing nodes could significantly mitigate the aforementioned issue. In this paper, we propose a flexible edge computing framework in which a fleet of UAVs function as mobile computational service providers, offering computation offloading services to multiple vehicles. We design and optimize a computation offloading model for the UAV-enabled vehicle edge computing environment. The proposed model tackles the task offloading challenge, aiming to optimize UAV revenue and task processing efficiency while considering the constraints of UAVs' restricted computational power and energy resources. Towards this end, our model jointly considers two key factors: task partitioning and computational resource allocation. To tackle the challenges posed by the aforementioned non-convex optimization problem, we construct a Markov Decision Process (MDP) model for the multi-UAV-enabled mobile edge computing system and introduce an innovative Multi-Agent Deep Reinforcement Learning (MADRL) framework addressing the decision-making challenge represented by MDP model. Comprehensive simulation outcomes illustrate that our devised task offloading technique outperforms other optimization methods. © 2014 IEEE.

关键词： Convex optimization

来源：评论

学校读者我要写书评

暂无评论

Provably Secure Authenticated Key-Management Mechanism for e-Healthcare Environment

引用

IEEE Internet of Things Journal 2025年

作者： Saleem, Muhammad Asad Li, Xiong Mahmood, Khalid Ghaffar, Zahid Xie, Yong Wang, Guijuan University of Electronic Science and Technology of China School of Computer Science and Engineering Sichuan Chengdu611731 China National Yunlin University of Science and Technology Graduate School of Intelligent Data Science Yunlin Douliu64002 Taiwan National Yunlin University of Science and Technology Graduate School of Engineering Science and Technology Douliu64002 Taiwan Qinghai Institute of Technology School of Computer and Information Science China Qinghai Provincial Laboratory for Intelligent Computing and Application China Key Laboratory of Computing Power Network and Information Security Jinan250353 China Shandong Provincial Key Laboratory of Industrial Network and Information System Security China Shandong Fundamental Research Center for Computer Science Jinan250353 China

The Internet of Things (IoT) is rapidly permeating all aspects of human life, involving a network of devices that share sensitive data. A notable application is the e-healthcare systems, which employ connected sensors, medical servers, and wearable devices. However, the public nature of communication in e-healthcare systems poses challenges such as security, privacy, and authentication of participating entities. Recently, many authentication protocols have been introduced to address these challenges. However, most of these protocols remain vulnerable to various security attacks, including device or medical server impersonation, denial of service, physical or cloning, and de-synchronization attacks. Therefore, we introduce an authenticated key-management protocol utilizing hash functions and Cipher-Block Chaining-Advanced Encryption Standard encryption (CBC-AES) encryption. The proposed protocol also employs the Physical Unclonable Function (PUF), which makes it more robust and efficient in resisting physical or cloning attacks. Additionally, the proposed scheme resists various security threats, including impersonation, session key leakage, ephemeral secret leakage, and de-synchronization attacks. We analyze the scheme’s security and reliability through formal and informal analysis. The informal analysis demonstrates that the scheme encompasses crucial security features, while the formal analysis substantiates. Moreover, performance analysis of the proposed protocol with various competing results indicates that our protocol achieves an average reduction in communication and computation overheads by 36.03.% and 41.79%, respectively. © 2014 IEEE.

关键词： Electronic health record

来源：评论

学校读者我要写书评

暂无评论

A Bi-selection Method Based on Consistent Matrix for Large-Scale datasets

引用

IEEE Transactions on Fuzzy Systems 2025年

作者： Quan, Jinsheng Qiao, Fengcai Yang, Tian Shen, Shuo Qian, Yuhua Hunan Normal University Hunan Provincial Key Laboratory of Intelligent Computing and Language Information Processing Institute of Interdisciplinary Studies Changsha410081 China National University of Defense Technology College of Advanced Interdisciplinary Studies Changsha410081 China Shanxi University Institute of Big Data Science and Industry Taiyuan030006 China

Biselection (feature and sample selection) enhances the efficiency and accuracy of machine learning models when handling large-scale data. Fuzzy rough sets, an uncertainty mathematics model known for its excellent interpretability, are widely used in machine learning, particularly for feature selection. While the consistent matrix has significantly improved the computational efficiency and scalability of feature selection, like most fuzzy rough set-based methods, it focuses only on feature selection and seldom incorporates sample selection. This feature-centric approach can limit classification performance, particularly in noisy and large-scale datasets where both features and samples require judicious selection. To overcome these limitations, this article explores the integration of sample selection with feature selection. First, we introduce a β-consistent granulation method to generate more accurate and concise fuzzy information granules, reducing the impact of noisy samples. In addition, a novel membership function is employed to distinguish noise while simultaneously evaluating both features and samples. As a result, a biselection algorithm with linear computation complexity is proposed to select high-quality features and samples. Numerical experiments demonstrate that, compared to eleven representative algorithms, our proposed method achieves an average accuracy improvement of 9.66% and a 933-fold increase in efficiency. © 2025 IEEE.

关键词： Biselection consistent matrix (CM) fuzzy sets granular computing rough sets

来源：评论

学校读者我要写书评

暂无评论

Optimizing AIGC Services Using Learning-based Stackelberg Game in Vehicular Metaverses

引用

IEEE Transactions on Vehicular Technology 2025年

作者： Lai, Bingkun Luo, Xiaofeng Kang, Jiawen Gao, Xiaozheng Yang, Zuyuan Niyato, Dusit Mao, Shiwen Guangdong University of Technology School of Automation Guangzhou510006 China Beijing Institute of Technology School of Information and Electronics Beijing10081 China Nanyang Technological University College of Computing and Data Science Singapore Auburn University Department of Electrical and Computer Engineering AuburnAL United States

The emerging vehicular metaverse embodies the next-generation vehicular networking paradigm. In the vehicular metaverses, Artificial Intelligence-Generated Content (AIGC) technology as a powerful content generation tool, is capable of providing an immersive experience for Vehicular Metaverse Users (VMUs). Due to limited computational resources within vehicles, VMUs rely on AIGC Service Providers (ASPs) to execute resource-intensive AIGC tasks within vehicular metaverses. However, large-scale AIGC service requests can lead to resource scarcity within the ASP, ultimately leading to declining service quality for VMUs. To tackle this challenge, we introduce a novel Stackelberg game framework utilizing the Generative Diffusion Model (GDM) for AIGC services, in which we experimentally reveal a relationship between image quality and diffusion steps. A Transformer-based Deep Reinforcement Learning (TDRL) algorithm is employed to find the optimal Stackelberg equilibrium under incomplete information. Numerical results indicate that our method converges to equilibrium efficiently, with superior utilities compared to baseline approaches. © 1967-2012 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Continual Reinforcement Learning for Digital Twin Synchronization Optimization

引用

IEEE Transactions on Mobile computing 2025年

作者： Tong, Haonan Chen, Mingzhe Zhao, Jun Hu, Ye Yang, Zhaohui Liu, Yuchen Yin, Changchuan Beijing University of Posts and Telecommunications Beijing Key Laboratory of Network System Architecture and Convergence Beijing Advanced Information Network Laboratory Beijing China College of Computing and Data Science Singapore University of Miami Department of Electrical and Computer Engineering Frost Institute for Data Science and Computing Coral Gables33146 United States University of Miami Department of Industrial and Systems Engineering Coral Gables33146 United States Zhejiang University College of Information Science and Electronic Engineering Hangzhou310027 China North Carolina State University Department of Computer Science RaleighNC27695 United States Beijing Key Laboratory of Network System Architecture and Convergence China Beijing University of Posts and Telecommunications Beijing Advanced Information Network Laboratory Beijing100876 China

This article investigates the adaptive resource allocation scheme for digital twin (DT) synchronization optimization over dynamic wireless networks. In our considered model, a base station (BS) continuously collects factory physical object state data from wireless devices to build a real-time virtual DT system for factory event analysis. Due to continuous data transmission, maintaining DT synchronization must use extensive wireless resources. To address this issue, a subset of devices is selected to transmit their sensing data, and resource block (RB) allocation is optimized. This problem is formulated as a constrained Markov process (CMDP) problem that minimizes the long-term mismatch between the physical and virtual systems. To solve this CMDP, we first transform the problem into a dual problem that refines RB constraint impacts on device scheduling strategies. We then propose a continual reinforcement learning (CRL) algorithm to solve the dual problem. The CRL algorithm learns a stable policy across historical experiences for quick adaptation to dynamics in physical states and network capacity. Simulation results show that the CRL can adapt quickly to network capacity changes and reduce normalized root mean square error (NRMSE) between physical and virtual states by up to 55.2%, using the same RB number as traditional methods. © 2025 IEEE. All rights reserved.

关键词： Markov processes

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：