检索结果-内蒙古大学图书馆

Feasibility study of cooperative sensing: radar cross section, synchronization, cooperative cluster,performance and prototype

引用

Science China(Information Sciences) 2025年第5期68卷 194-212页

作者： Guangyi LIU Rongyan XI Tao JIANG Xiaoqian WANG Lincong HAN Jing JIN Jing DONG Yajuan WANG Yuxiang ZHANG Qixing WANG Jiangzhou WANG Future Research Laboratory China Mobile Research Institute State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications School of Engineering University of Kent

This paper investigates the potential of integrated sensing and communication(ISAC) for the upcoming sixth generation(6G) mobile communications system, which is expected to go beyond communication and incorporate sensing capabilities. A major challenge with conventional ISAC systems is the self-interference caused by using a base station(BS)as both the sensing transmitter and receiver, which requires complex and costly hardware upgrades to enable full *** address this issue, we propose a cooperative sensing approach, where node A acts as the transmitter and node B as the receiver. In this study, we analyze the scattering characteristics of targets in a cooperative sensing scenario and find that the radar cross section(RCS) of unmanned aerial vehicle(UAV) follows a lognormal distribution. We then propose a time-frequency synchronization method, a cooperative sensing cluster construction method, and a multi-node signal-level fusion method to accurately determine the target's position. Simulation results, incorporating non-ideal factors, show that the proposed system can achieve precise target positioning. The accuracy is improved by more than five times compared to mono-static sensing and over 25% compared to result-level fusion methods. Additionally, prototype experimental results validate the feasibility of the proposed system.

关键词： 6G integrated sensing and communication synchronization bistatic RCS multi-node collaboration cooperative cluster prototype

来源：评论

学校读者我要写书评

暂无评论

Knowledge Distillation for Image Restoration: Simultaneous Learning from Degraded and Clean Images

Knowledge Distillation for Image Restoration: Simultaneous L...

引用

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

作者： Zhang, Yongheng Yan, Danfeng State Key Laboratory of Networking and Switching Technology BUPT Beijing China

ISBN: (纸本)9798350368741

Model compression through knowledge distillation has seen extensive application in classification and segmentation tasks. However, its potential in image-to-image translation, particularly in image restoration, remains underexplored. To address this gap, we propose a Simultaneous Learning Knowledge Distillation (SLKD) framework tailored for model compression in image restoration tasks. SLKD employs a dual-teacher, single-student architecture with two distinct learning strategies: Degradation Removal Learning (DRL) and Image Reconstruction Learning (IRL), simultaneously. In DRL, the student encoder learns from Teacher A to focus on removing degradation factors, guided by a novel BRISQUE extractor. In IRL, the student decoder learns from Teacher B to reconstruct clean images, with the assistance of a proposed PIQE extractor. These strategies enable the student to learn from degraded and clean images simultaneously, ensuring high-quality compression of image restoration models. Experimental results across five datasets and three tasks demonstrate that SLKD achieves substantial reductions in FLOPs and parameters, exceeding 80%, while maintaining strong image restoration performance. © 2025 IEEE.

关键词： feature extractor image restoration Knowledge distillation simultaneous learning

来源：评论

学校读者我要写书评

暂无评论

Soft Knowledge Distillation with Multi-Dimensional Cross-Net Attention for Image Restoration Models Compression

Soft Knowledge Distillation with Multi-Dimensional Cross-Net...

引用

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

作者： Zhang, Yongheng Yan, Danfeng State Key Laboratory of Networking and Switching Technology BUPT Beijing China

ISBN: (纸本)9798350368741

Transformer-based encoder-decoder models have achieved remarkable success in image-to-image transfer tasks, particularly in image restoration. However, their high computational complexity-manifested in elevated FLOPs and parameter counts-limits their application in real-world scenarios. Existing knowledge distillation methods in image restoration typically employ lightweight student models that directly mimic the intermediate features and reconstruction results of the teacher, overlooking the implicit attention relationships between them. To address this, we propose a Soft Knowledge Distillation (SKD) strategy that incorporates a Multi-dimensional Cross-net Attention (MCA) mechanism for compressing image restoration models. This mechanism facilitates interaction between the student and teacher across both channel and spatial dimensions, enabling the student to implicitly learn the attention matrices. Additionally, we employ a Gaussian kernel function to measure the distance between student and teacher features in kernel space, ensuring stable and efficient feature learning. To further enhance the quality of reconstructed images, we replace the commonly used L1 or KL divergence loss with a contrastive learning loss at the image level. Experiments on three tasks-image deraining, deblurring, and denoising-demonstrate that our SKD strategy significantly reduces computational complexity while maintaining strong image restoration capabilities. © 2025 IEEE.

关键词： contrastive learning image restoration Knowledge distillation multi-dimensional cross-net attention

来源：评论

学校读者我要写书评

暂无评论

RecStream: Graph-aware Stream Management for Concurrent Recommendation Model Online Serving 31

RecStream: Graph-aware Stream Management for Concurrent Reco...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Guo, Shuxi Qi, Qi Sun, Haifeng Liao, Jianxin Wang, Jingyu State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China

ISBN: (纸本)9798891761971

Recommendation Models (RMs) are crucial for predicting user preferences and enhancing personalized experiences on large-scale platforms. As the application of recommendation models grows, optimizing their online serving performance has become a significant challenge. However, current serving systems perform poorly under highly concurrent scenarios. To address this, we introduce RecStream, a system designed to optimize stream configurations based on model characteristics for handling high concurrency requests. We employ a hybrid Graph Neural Network architecture to determine the best configurations for various RMs. Experimental results demonstrate that RecStream achieves significant performance improvements, reducing latency by up to 74%. ©2025 Association for Computational Linguistics.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Flight Trajectory Control With Network-Oriented Hierarchical Reinforcement Learning for UAVs-Assisted Data Time-Sensitive IoT

引用

IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS 2025年第5期26卷 6332-6345页

作者： Wang, Jingyu Li, Songyang Chen, Dezhi Fu, Qianlong Qi, Qi Sun, Haifeng He, Bo Liao, Jianxin Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China Pengcheng Lab Shenzhen 518000 Peoples R China

Within the Internet of Things (IoT) for traffic monitoring, the employment of autonomous aerial vehicles (AAVs) as relays for collecting and transmitting real-time data from traffic sensors to base stations has proven a promising approach. In UAV-assisted Data Time-Sensitive IoT (DTIoT), the Age of Information is a crucial metric assessing data freshness, measuring the elapsed time from traffic sensors to the base station. Optimizing flight trajectories of multiple UAVs to minimize AoI while adhering to energy constraints poses a significant challenge. Current research often employs deep reinforcement learning for UAV trajectory control. Nevertheless, managing multi-agent continuous trajectories in intricate DTIoT network environments faces obstacles due to sparse rewards, thus impeding the training of deep neural network-based control policies using traditional DRL techniques. In this paper, we propose a network-oriented hierarchical reinforcement learning (NO-HRL) to control the UAVs' flight trajectory in DTIoT networks for minimizing the AoI. We devise a control policy leveraging a two-tier hierarchical DRL framework, with the upper tier determining the target and the lower tier executing it. We also introduce a decoupled sequential training approach to efficiently train the mutually dependent two-tier DRL network of NO-HRL. Experimental results demonstrate that our method excels in optimizing AoI for DTIoT compared to other baselines.

关键词： Trajectory Autonomous aerial vehicles Sensors Training Internet of Things Aerospace electronics Monitoring Data collection Base stations Information age Autonomous aerial vehicle (AAV) sensor data collection age of information hierarchical reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Differential Privacy Federated Learning Based on Adaptive Adjustment

引用

Computers, Materials & Continua 2025年第3期82卷 4777-4795页

作者： Yanjin Cheng Wenmin Li Sujuan Qin Tengfei Tu The State Key Laboratory of Networking and Switching Technology Beijing University of Posts and TelecommunicationsBeijing100876China The School of Cyberspace Security Beijing University of Posts and TelecommunicationsBeijing100876China

Federated learning effectively alleviates privacy and security issues raised by the development of artificial intelligence through a distributed training *** research has shown that attackers can compromise user privacy and security by stealing model ***,differential privacy is applied in federated learning to further address malicious ***,the addition of noise and the update clipping mechanism in differential privacy jointly limit the further development of federated learning in privacy protection and performance ***,we propose an adaptive adjusted differential privacy federated learning ***,a dynamic adaptive privacy budget allocation strategy is proposed,which flexibly adjusts the privacy budget within a given range based on the client’s data volume and training requirements,thereby alleviating the loss of privacy budget and the magnitude of model ***,a longitudinal clipping differential privacy strategy is proposed,which based on the differences in factors that affect parameter updates,uses sparse methods to trim local updates,thereby reducing the impact of privacy pruning steps on model *** two strategies work together to ensure user privacy while the effect of differential privacy on model accuracy is *** evaluate the effectiveness of our method,we conducted extensive experiments on benchmark datasets,and the results showed that our proposed method performed well in terms of performance and privacy protection.

关键词： Federated learning privacy protection differential privacy deep learning

来源：评论

学校读者我要写书评

暂无评论

Exploiting Structured Sparsity in Near Field: From the Perspective of Decomposition

引用

IEEE COMMUNICATIONS MAGAZINE 2025年第1期63卷 37-43页

作者： Guo, Xufeng Chen, Yuanbin Wang, Ying Yuen, Chau Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing Peoples R China Nanyang Technol Univ Sch Elect & Elect Engn Nanyang Singapore

The structured sparsity can be leveraged in traditional far-field channels, greatly facilitating efficient sparse channel recovery by compressing the complexity of overheads to the level of the scatterer number. However, when experiencing a fundamental shift from planar-wave-based far-field modeling to spherical-wave-based near-field modeling, whether these benefits persist in the near-field regime remains an open issue. To answer this question, this article delves into structured sparsity in the near-field realm, examining its peculiarities and challenges. In particular, we present the key features of near-field structured sparsity in contrast to the far-field counterpart, drawing from both physical and mathematical perspectives. Upon unmasking the theoretical bottlenecks, we resort to bypassing them by decoupling the geometric parameters of the scatterers, termed the triple parametric decomposition (TPD) framework. It is demonstrated that our novel TPD framework can achieve robust recovery of near-field sparse channels by applying the potential structured sparsity and avoiding the curse of complexity and overhead.

关键词： Three-dimensional displays Complexity theory Near field communication Telecommunication network management Channel estimation Wireless sensor networks Linear antenna arrays 6G mobile communication switching circuits Spectral efficiency Space division multiplexing MIMO

来源：评论

学校读者我要写书评

暂无评论

FEDNPAIT: Federated Learning with NADAM and PADAM for Instruction Tuning 24th

FEDNPAIT: Federated Learning with NADAM and PADAM for Ins...

引用

24th International Conference on Algorithms and Architectures for Parallel Processing, ICA3PP 2024

作者： Gao, Zhipeng Li, Yichen Yu, Xinlei The State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Communications Beijing China

ISBN: (纸本)9789819615278

Large language models (LLMs) trained on massive publicly accessible data excel in general domain tasks. However, the performance of LLMs in specialized domains such as healthcare and mathematics frequently fails to meet established benchmarks. The primary obstacle to improving this performance is the scarcity of high-quality and domain-specific data due to stringent data privacy regulations. Federated learning, which facilitates the training of LLMs across distributed devices through a decentralized approach that precludes centralized data aggregation presents a potential solution. Nevertheless, federated learning for instruction tuning encounters challenges including data heterogeneity and substantial computational resource requirements. Consequently, this paper proposes FEDNPAIT, an innovative federated learning framework for instruction tuning that incorporates Nesterov-accelerated momentum and a new parameter for controlling the adaptive strength of gradient updates during optimization to improve the performance of LLMs in heterogeneous federated learning environments. Furthermore, FEDNPAIT effectively addresses computational resource limitations through the incorporation of Low-Rank Adaptation. Experimental results indicate that FEDNPAIT effectively mitigates the adverse effects of data heterogeneity in federated environments and outperforms existing federated learning algorithms. Specifically, the results from tuned Llama3-8B demonstrate enhanced generalization and robustness, illustrating how to incorporate complex optimization algorithms into federated training of LLMs, and opening new possibilities for federated instruction tuning in sensitive industries. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

FANT: Flexible Attention-Shifting Network Telemetry

引用

APPLIED SCIENCES-BASEL 2025年第2期15卷 892-892页

作者： Cui, Mingwei Peng, Yufan Yang, Fan Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China Purple Mt Labs Nanjing 211111 Peoples R China

As data center networks grow in scale and complexity, the active inband network telemetry (AINT) system collects a broader range of network status metrics to provide comprehensive visibility for AINT-related network applications, but it also leads to higher measurement costs. To address this issue, we introduce the Flexible Attention-shifting Network Telemetry (FANT), which dynamically focuses on critical links in each measurement cycle. Specifically, FANT employs a metric categorization strategy that divides all measurement metrics into two categories: basic measurements, which are lightweight but cover fewer metrics, and detailed measurements, which are comprehensive but incur higher overhead. Based on the analysis of the previous cycle's measurements, FANT identifies which links are suspicious and then activates certain probe traces through an attention-shifting mechanism to collect detailed measurements of these links in the current cycle. To further save bandwidth, we model the attention-shifting process and apply heuristic algorithms to solve it. Our experiments show that FANT effectively supports the operation of ANT network applications. In a fat-tree topology with 30 pods, FANT significantly reduces bandwidth usage to 42.6% of the state-of-the-art solution. For scenarios requiring rapid computation, FANT can accelerate algorithm execution 105x by setting acceleration factors, with only a 6.4% performance loss.

关键词： active inband network telemetry monitoring programmable data plane

来源：评论

学校读者我要写书评

暂无评论

A Semantic Communication System for Point Cloud

引用

IEEE TRANSACTIONS ON VEHICULAR technology 2025年第1期74卷 894-910页

作者： Liu, Xiaoyi Liang, Haotai Bao, Zhicheng Dong, Chen Xu, Xiaodong Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China Peng Cheng Lab Dept Broadband Commun Shenzhen 518066 Peoples R China

Point cloud, as a 3D representation, finds wide applications in domains such as autonomous driving, virtual reality (VR), and augmented reality (AR). However, traditional communication systems are not well-suited for large-scale point cloud data transmissions, as such systems operate at the bit level without leveraging the inherent semantic information of the point cloud. This paper introduces a point cloud-based semantic communication system (PCSC) that leverages AI-based encoding techniques to extract semantic information from the point cloud. Furthermore, joint source-channel coding (JSCC) technology is employed to overcome noise channel distortion and address the "cliff effect" prevalent in traditional communication methods. Additionally, the proposed system achieves controllable coding rates without requiring extensive network fine-tuning. By analyzing the significance of the encoded semantic vector, the method discards semantically-unimportant information, enhancing transmission efficiency. Moreover, the PCSC is integrated with the recently proposed model division multiple access (MDMA) technology, resulting in a multi-user point cloud MDMA transmission system (M-PCSC). Experimental results demonstrate that the proposed method surpasses traditional approaches by 10 dB under the PSNR D1 and PSNR D2 metrics while effectively mitigating the "cliff effect" in transmission.

关键词： Semantics Point cloud compression Communication systems Physical layer Feature extraction Data mining Wireless communication Semantic communications point cloud transmission controllable coding rate model division multiple access

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：