检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Zhang, Guoxin Song, Ziying Liu, Lin Ou, Zhonghong State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications China School of Computer Science and Technology Beijing Jiaotong University China

Multimodal 3D object detection has garnered considerable interest in autonomous driving. However, multimodal detectors suffer from dimension mismatches that derive from fusing 3D points with 2D pixels coarsely, which leads to suboptimal fusion performance. In this paper, we propose a multimodal framework FGU3R to tackle the issue mentioned above via unified 3D representation and fine-grained fusion, which consists of two important components. First, we propose an efficient feature extractor for raw and pseudo points, termed Pseudo-Raw Convolution (PRConv), which modulates multimodal features synchronously and aggregates the features from different types of points on key points based on multimodal interaction. Second, a Cross-Attention Adaptive Fusion (CAAF) is designed to fuse homogeneous 3D RoI (Region of Interest) features adaptively via a cross-attention variant in a fine-grained manner. Together they make fine-grained fusion on unified 3D representation. The experiments conducted on the KITTI and nuScenes show the effectiveness of our proposed method. Copyright © 2025, The Authors. All rights reserved.

关键词： Object recognition

来源：评论

学校读者我要写书评

暂无评论

Community-oriented Duplex Privacy Amplification and Active Poisoning Resistance for Heterogeneous Federated Learning

引用

IEEE Transactions on Dependable and Secure Computing 2025年

作者： Zhou, Zan Zhao, Jun Yang, Shujie Li, Hongjing Ma, Tengchao Xu, Changqiao Beijing University of Posts and Telecommunications State Key Laboratory of Networking and Switching Technology Beijing100876 China Nanyang Technological University School of Computer Science and Engineering China

Privacy protection and poisoning resilience are important concerns for federated learning (FL). During a relatively long period, the corresponding solutions are regarded as orthogonal and investigated separately. Unfortunately, due to the increasingly complex structure and ever-growing parameter dimensions of the models to be trained, the forthright coupling of existing differential privacy and Byzantine resilience techniques has been proved incompatible with FL. This emerging problem prompts us to give serious thought to jointly guaranteeing data privacy and model integrity. Besides, worse still, the multi-task characteristic and data imbalance of heterogeneous FL inevitably introduce huge variances, which make privacy-preserving under acceptable accuracy loss even more complicated, not to mention efficient and agile poisoning resistance. Against this bothersome situation, we propose a community-oriented secure heterogeneous FL (CoS-HFL) framework to provide guaranteed privacy protection and significant model robustness simultaneously. CoS-HFL includes two parts: community-oriented duplex privacy amplification and credit-based poisoning resistance. The former copes with potential leakage threats with both uplink and downlink obfuscations. The latter further actively thwarts poisoning attacks based on credibility evaluation. Furthermore, we conduct experiments on benchmark datasets to highlight the performance of CoS-HFL in terms of privacy amplification, poisoning resistance, and learning accuracy under adversarial and heterogeneous environments. © 2004-2012 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Graph Attentional Based Agglomerative Cluster for UAV Swarm Networks

引用

IEEE Transactions on Network Science and Engineering 2025年

作者： Huang, Shan Yao, Haipeng Wang, Xiaoman Mai, Tianle Wang, Zunliang Guo, Song Beijing University of Posts and Telecommunications State Key Laboratory of Networking and Switching Technology Beijing100876 China Hong Kong Polytechnic University Department of Computing Hong Kong

In recent years, UAVs have been widely employed in the construction of post-disaster emergency communication networks due to their flexible mobility and self-organizing capabilities, enabling them to quickly and spontaneously establish communication links. Post-disaster rescue efforts, restoring connectivity and communication, as well as providing disaster information, all require strong support from network applications. However, an essential element in the development of effective UAV swarm-based applications is the network system. Compared to the fixed networks, UAV swarm networks present unique challenges in design and implementation due to its characteristics of high mobility nodes, unstable links, and dynamic topology. Recently, clustering technology has gained recognition as an effective approach to constructing stable UAV swarm networks. In this paper, for UAV swarm networks, we propose a graph attention-based agglomerative clustering algorithm. This algorithm allows UAV nodes to learn similarity relationships through a graph attention network. By considering the mobility similarity between UAV nodes, each UAV node can merge with its adjacent nodes. Furthermore, we also design a cluster head selection algorithm based on mixed strategy games. The algorithm's effectiveness and accuracy were demonstrated by simulation results, which showed a 80.04% increase in network lifetime compared to baseline algorithms. © 2013 IEEE.

关键词： Disasters

来源：评论

学校读者我要写书评

暂无评论

Anchor-Aware Position Enhancement for Weakly Supervised Knowledge Graph Alignment

IEEE Transactions on Audio, Speech and Language Processing

引用

IEEE Transactions on Audio, Speech and Language Processing 2025年 33卷 1156-1169页

作者： Wei Tang Haifeng Sun Jingyu Wang Qi Qi Shimin Tao Hao Yang Bo Wu LianYuan Li State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China *** Beijing China Huawei Beijing China China Mobile Research Institute Beijing China

Entity alignment (EA) models currently rely heavily on large-scale labeled anchor links to succeed. However, these models often face challenges due to the limited availability of anchor links. To enhance the alignment performance of existing models under weakly supervised settings, some researchers have turned to active learning to search for more informative anchor links. Despite these efforts, the fundamental issue remains that existing neural EA models tend to overfit the limited anchor links, resulting in poor alignment performance. To tackle this problem, this paper addresses the challenge of low-resource EA by focusing on anchor entities and proposes a novel approach called Anchor-aware Position-Enhanced Low-resource Entity Alignment (APLEA). In the encoding process, we introduce a simple yet effective position encoding method that captures the long-term relationships between anchor entities and other distant entities using random walk with restart (RWR). This method allows us to extract more advanced positional features through a position attention layer. By incorporating structural information, relational information, and positional information from these three aspects, we obtain the final entity embeddings for alignment. In the decoding process, we address the issue of existing EA models disregarding the matched neighborhood constraints when inferring alignment results. Specifically, neighboring entities of a specific entity $e$ may not be strictly mapped to those of $e$ ’s counterpart. To overcome this limitation, we propose a simple refinement approach to the final similarity matrix without the need for additional training. Extensive experiments conducted on publicly available datasets demonstrate that APLEA significantly outperforms the state-of-the-art approaches by a large margin in weakly supervised EA tasks.

关键词： Knowledge graphs Feature extraction Encoding Vectors Training Graph neural networks Data mining Speech processing Electronic mail Decoding

来源：评论

学校读者我要写书评

暂无评论

Site-Specific Ray-Tracing and Experimental Validation for THz Channel Characterization

引用

IEEE Transactions on Terahertz Science and technology 2025年第3期15卷 319-331页

作者： Yuan, Zhiqiang Yu, Guangcheng Lyu, Yejian Zhang, Jianhua Zhang, Fengchun Fan, Wei Southeast University National Mobile Communications Research Laboratory School of Information Science and Engineering Nanjing210096 China Joint Institute Shanghai Jiao Tong University Terahertz Wireless Communications Lab Shanghai200240 China Beijing University of Posts and Telecommunications State Key Lab of Networking and Switching Technology Beijing100876 China Section Aalborg9220 Denmark Purple Mountain Laboratories Nanjing210096 China

Terahertz (THz) technology is seen as a key component of the upcoming 6G networks due to the huge untapped spectrum. Accurate channel models are crucial for the design and development of such THz communication and sensing systems. As the THz channel exhibits unique features including high propagation loss, sparsity, and near-field properties (with ultramassive multiple-input multiple-output systems), deterministic ray-tracing (RT), which can simulate wireless channels in a site-specific manner based on the high-frequency approximation of Maxwell's equation, is considered suitable for THz channel modeling. However, a thorough analysis of RT-based THz channel modeling and its extensive experimental validation (especially at 300 GHz) is still missing in the literature. In this article, we develop an RT modeling framework tailored for THz channels, which incorporates processes of environment reconstruction and electromagnetic calculation, material calibration, and acceleration. Then, three measurements with 14 locations ranging from 3 to 58 m covering frequency bands at 300 and 100 GHz are used to validate the developed RT framework. It is observed that there is an excellent match for the dominant paths between the real-world measurements and the RT results, and the near-field characteristics are realistically and accurately captured in the modeling. Furthermore, the validated RT framework is employed to enable an in-depth THz channel characterization at 300 GHz in an indoor hall scenario. This analysis is rooted in extensive channel data across 30 locations, consisting of experimental data at 12 locations in the first measurement and RT-simulated data for another 18 locations. The developed RT framework and the channel characterization aim to provide insights for future THz channel modeling and standardization studies. © 2011-2012 IEEE.

关键词： Maxwell equations

来源：评论

学校读者我要写书评

暂无评论

Asynchronous PID Control for T-S Fuzzy Systems Over Gilbert-Elliott Channels Utilizing Detected Channel Modes

引用

IEEE Transactions on Fuzzy Systems 2025年第5期33卷 1555-1567页

作者： Wang, Yezheng Wang, Zidong Zou, Lei Ge, Quanbo Dong, Hongli Nanjing University of Information Science and Technology School of Automation Nanjing210044 China Shandong University of Science and Technology College of Electrical Engineering and Automation Qingdao266590 China Brunel University London Department of Computer Science UxbridgeUB8 3PH United Kingdom Donghua University College of Information Science and Technology Shanghai201620 China Ministry of Education Engineering Research Center of Digitalized Textile and Fashion Technology Shanghai201620 China Nanjing University of Information Science and Technology Jiangsu Provincial University Key Laboratory of Big Data Analysis and Intelligent Systems Nanjing210044 China Nanjing210044 China Northeast Petroleum University State Key Laboratory of Continental Shale Oil Daqing163318 China Northeast Petroleum University Artificial Intelligence Energy Research Institute Daqing163318 China Northeast Petroleum University Heilongjiang Provincial Key Laboratory of Networking and Intelligent Control Daqing163318 China

This article is concerned with the H∞ proportional-integral-derivative (PID) control problem for Takagi-Sugeno fuzzy systems over lossy networks that are characterized by the Gilbert-Eillott model. The communication quality is reflected by the presence of two channel modes (i.e., "bad" mode and "good" mode), which switch randomly according to a Markov process. In the "bad" mode, packet dropouts are governed by a stochastic variable sequence. Considering the inaccessibility of channel modes, a mode detector is utilized to estimate the communication situation. The relationship between the actual channel mode and the estimated mode is depicted in terms of certain conditional probabilities. Moreover, a comprehensive model is constructed to represent the probability uncertainties arising from statistical errors in channel mode switching, packet dropouts, and mode detection processes. Subsequently, a robust asynchronous PID controller, based on the detected channel mode, is proposed. Sufficient conditions are then derived to ensure the mean-square stability of the closed-loop system while maintaining the desired H∞performance. Finally, the efficacy of the proposed design approach is demonstrated through a simulation example. © 1993-2012 IEEE.

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

Lattice Security Analysis Algorithms for the Quadratic Congruence Outsourcing Scheme in IoT

引用

IEEE Internet of Things Journal 2025年

作者： Zhang, Ran Bi, Jingguo Li, Lixiang Peng, Haipeng Beijing University of Posts and Telecommunications School of Cyberspace Security Beijing100876 China State Key Laboratory of Cryptology Beijing100878 China Beijing University of Posts and Telecommunications Information Security Center State Key Laboratory of Networking and Switching Technology Beijing100876 China

Designing secure outsourcing schemes enables resource-constrained IoT devices to perform highly complex computational tasks. Solving the quadratic congruence problem is one of the core components in the construction of cryptographic algorithms for IoT. Recently, Rangasamy designed an outsourcing scheme for solving quadratic congruence equations. Interestingly, we find that the scheme has the risk of secret information being cracked. We propose two lattice attack algorithms in this paper. In the first attack algorithm, we prove that there is a possibility of leaking secret information in the public parameters of the outsourcing scheme. Specifically, by intercepting the parameters transmitted between the client and the server, and combining the ideas from Fermat's Little Theorem and the Euclidean algorithm, the attacker can obtain multiples of the secret parameter p. Furthermore, after recover the value of p, our attack algorithm is capable of recovering all secret parameters in the quadratic congruence equation. In the second attack algorithm, we note that the authors recommend the randomly chosen value of k to be small in the original outsourcing scheme. However, in this paper, we point out that the random value k should not be too small, otherwise, the outsourcing scheme would be wrecked. More precisely, we use the Coppersmith method to provide an approach that can recover all the secret information. Our attack algorithm will succeed once k satisfied |k| © 2014 IEEE.

关键词： Information leakage

来源：评论

学校读者我要写书评

暂无评论

A Transformer-Block-Wise Collaborative Training Mechanism with Hybrid Parallelism Over Heterogeneous Networks

A Transformer-Block-Wise Collaborative Training Mechanism wi...

引用

IEEE Conference on Wireless Communications and networking

作者： Jiewei Chen Jingrong Wang Shaoyong Guo Jiakai Hao Xuesong Qiu Zehui Xiong State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications State Grid Beijing Information & Communication Company Pillar of Information Systems Technology and Design Singapore University of Technology and Design

ISBN: (数字)9798350368369

ISBN: (纸本)9798350368376

With the rise of AI-Generated Content (AIGC) services in wireless networks, efficient and high-quality distributed training of Large Language Models (LLMs) has become essential for enabling the large-scale application of next generation AI technologies. However, the extensive parameters of LLMs impose significant demands on memory, computing power and communication resources in heterogeneous networks. To efficiently utilize the dispersed network resources, this paper presents a First-Pipeline- Then-Federated Learning (FPTFL) approach with a hybrid parallel scheduling strategy to facilitate the training of Transformer-based LLMs. We propose a block-wise splitting mechanism to partition the Transformer's encoder into distinct segments, which are deployed cross individual devices. The encoder parameters and intermediate smashed data are uploaded to the edge server, where the whole model is updated through federated aggregation. Particularly, we develop a fine-grained computation-efficient method based on pipeline parallelism, enabling the segments to cooperatively train the entire encoder. An optimization problem is formulated to determine the LLM segments and the number of micro-batches under network resource constraints, with the goal of minimizing the total latency of LLM training services. Simulation results demonstrate that our approach enables Transformer-based model training on resource-constrained devices, preserves model performance, and reduces waiting time.

关键词： Training Performance evaluation Large language models Wireless networks Pipelines Parallel processing Transformers Heterogeneous networks Classification algorithms Optimization

来源：评论

学校读者我要写书评

暂无评论

A Beyond-Line-of-Sight Long-Distance Reflection Communication System via LEO Satellites

A Beyond-Line-of-Sight Long-Distance Reflection Communicatio...

引用

IEEE Conference on Wireless Communications and networking

作者： Zihan Zhang Xiaoling Hu Xiaowei Qian Mugen Peng State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China National Mobile Communications Research Laboratory Southeast University Nanjing China

ISBN: (数字)9798350368369

ISBN: (纸本)9798350368376

Reconfigurable Intelligent Surface (RIS) is a promising low-cost technology for enhancing wireless communication coverage. It consists of passive reflecting elements that can adjust the phase shifts of incident signals. RIS is not only deployed on ground structures but also integrated with unmanned aerial vehicles, high-altitude platforms, and satellites. This enables the establishment of virtual line-of-sight (VLoS) links, thereby over-coming obstacles, extending communication range, and reducing long-distance path loss. However, satellite-mounted RIS faces challenges such as increased launch costs and the need for complex dynamic beamforming optimization. Fortunately, the deployment of large plane array antennas is becoming mainstream in low Earth orbit (LEO) satellite constellations. These large arrays, acting as high-reflectivity meta-surfaces, enable reflection communication via satellites. In this paper, we propose a beyond-line-of-sight long-distance reflection communication system using LEO satellites. Specifically, we introduce a new paradigm where LEO satellite serves as a passive reflector to establish a VLoS path between ground stations. We also propose a satellite selection method to choose suitable satellites as passive reflectors and a communication window prediction method to determine optimal communication opportunities. Simulation results validate the feasibility and effectiveness of the proposed system.

关键词： Wireless communication Satellites Simulation Low earth orbit satellites Prediction methods Reconfigurable intelligent surfaces Reflection Transceivers Vehicle dynamics Antenna arrays

来源：评论

学校读者我要写书评

暂无评论

Deep Joint Source-Channel Coding Based on Feedback-Driven Codebook Optimization

Deep Joint Source-Channel Coding Based on Feedback-Driven Co...

引用

IEEE Conference on Wireless Communications and networking

作者： Weijie Zheng Haotai Liang Chen Dong Xiaodong Xu State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China Department of Broadband Communication Peng Cheng Laboratory Shenzhen Guangdong China

ISBN: (数字)9798350368369

ISBN: (纸本)9798350368376

The rapid growth of wireless communication technologies has made semantic communication an increasingly important approach for efficient data transmission. In this paper, a novel deep joint source-channel coding (DeepJSCC) framework based on Feedback-Driven Codebook Optimization (FDCO) for wireless image transmission is proposed. The framework dynamically optimizes the codebook based on channel feedback to improve image reconstruction performance under varying channel conditions. Specifically, a FDCO network is introduced to adjust the balance between common and individual information in the codebook based on the input signal-to-noise ratio (SNR). In addition, the residual between the original and quantized images is encoded to obtain semantic details, which are transmitted to reduce semantic quantization loss. Experimental results demonstrate that the proposed framework improves image quality and compression efficiency, especially under low SNR, and validates FDCO's dynamic adjustment of the codebook's information balance, leading to enhanced performance.

关键词： Wireless communication Image quality Image coding Quantization (signal) Image communication Semantic communication Robustness Optimization Image reconstruction Signal to noise ratio

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：