检索结果-内蒙古大学图书馆

13th International Conference on Computer Engineering and Networks, CENet 2023

作者： Zhao, Yikun Zhang, Jinli Zhou, Fanqin Li, Wenjing Feng, Lei State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China

ISBN: (纸本)9789819992386

Millimeter wave (mmWave) aerial base stations offer distinct advantages such as enhanced flexibility and mobility compared to ground base station communication. Additionally, the utilization of mmWave technology enables the provision of high-speed, directional services to ground users while mitigating interference with existing ground networks. To establish a good communication link between mmWave aerial base stations and terrestrial mobile users, it is essential for base stations to form directional beam to cover terrestrial users. However, the movement of users and the change of environmental factors will lead to the misalignment between the mmWave beam formed by the base stations and the users, resulting in the instability of the communication link, which will seriously affect the quality of communication service. Therefore, this paper proposes a deep learning-based beam tracking scheme for mmWave aerial base stations. Simulation results show that our proposed schemes achieve higher beamforming gain and require less beam training overhead compared to existing conventional and deep learning-based approaches. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd 2024.

关键词： Base stations

来源：评论

学校读者我要写书评

暂无评论

An optimized RDMA QP communication mechanism for hyperscale AI infrastructure

引用

CLUSTER COMPUTING-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS 2025年第1期28卷 1-18页

作者： Wang, Junliang Lin, Baohong Zhang, Jiao Sun, Mengyu Pan, Yongchen China Telecom Co Ltd China Telecom Res Inst Guangzhou 510660 Peoples R China BUPT State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China

The current artificial intelligence (AI) infrastructure widely employs remote direct memory access (RDMA) protocol for high-performance communication in networks, utilizing Reliable Connection (RC)-based Queue Pairs (QP) to ensure end-to-end correct and ordered data transmission. However, as the scale of AI infrastructure continues to expand, this RC-based QP communication mechanism faces deficiencies in scalability and is prone to congestion, resulting in degraded network transfer performance. In this paper, we propose an optimized RDMA QP communication mechanism to address scalability and congestion issues in hyper-scale AI infrastructure networks. Firstly, we replace RC-based QPs with Reliable Datagram (RD)-based QPs and propose a new reliable mechanism to address scalability problems, eliminating the need for repetitive QP establishment by AI processes during external communication. Additionally, to mitigate congestion caused by a single path, we implement multipath data transmission by introducing a new unordered reception method in the network software stack. Through experiments and simulation tests, the optimized RDMA QP communication in large-scale AI infrastructure exhibits exceptional scalability and significantly reduces the occurrence of congestion, resulting in an overall network performance improvement of over 15%.

关键词： Queue pair (QP) Hyperscale communication Remote direct memory access (RDMA) Multipath transmission

来源：评论

学校读者我要写书评

暂无评论

Coordinating Communication and Computing for Wireless VR in Open Radio Access Networks

引用

IEEE Transactions on Mobile Computing 2025年第7期24卷 6075-6089页

作者： Li, Hongtao Chen, Ziqi Guo, Fengxian Li, Nan Sun, Yaohua Peng, Mugen Liu, Yuanwei Beijing University of Posts and Telecommunications State Key Laboratory of Networking and Switching Technology Beijing100876 China China Mobile Research Institute Beijing100053 China Hong Kong

Driven by diverse applications, radio access networks (RAN) are expected to embrace built-in computing and intelligence, forming a versatile wireless computing platform that closely integrates communication and computing. To fully unleash the potential of such a synergistic system, it is essential to coordinate communication and computing with intelligence unlocked by the radio intelligent controllers (RICs) in O-RAN. Building on the groundwork established by existing theoretical studies and simulations, we develop a platform that can emulate the events in the real-world system in more detail, bringing theoretical works closer to practical implementation. In this paper, we first introduce ns-GP-O-RAN, a software simulation platform developed over ns-3, enabling communication, computation task processing, large-scale data collection, and testing of system-level orchestration policies through user-level control. Taking virtual reality (VR) as an example, we formulate the computation offloading problem and develop a prediction-based computation offloading xAPP, which contains a prediction phase to predict users’ end-to-end (E2E) performance with the deep neural network and a system-level decision-making phase for global orchestration with the differential evolution algorithm. We evaluate the system capacity and E2E latency over the developed ns-GP-O-RAN, which is more effective than existing approaches. © 2002-2012 IEEE.

关键词： Radio access networks

来源：评论

学校读者我要写书评

暂无评论

Optimizing Dual-Mode UAV-Assisted Remote IoT Data Collection with Decentralized DRL in 6G-Enabled Space-Air-Ground Networks

Optimizing Dual-Mode UAV-Assisted Remote IoT Data Collection...

引用

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

作者： Jin, Ru Lin, Rongheng State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications China

ISBN: (纸本)9798350368741

As sixth-generation (6G) networks extend to remote regions, efficient data collection from Internet of Things (IoT) devices faces challenges such as limited infrastructure, diverse latency requirements, and energy constraints. To address these, we propose a dual-mode data transmission strategy: (1) a UAV-satellite network for real-time delay-sensitive data, where wireless power transfer (WPT) is introduced to alleviate energy depletion caused by frequent data transmission, and a deadline τmax is set in each timeslot to maintain data freshness. (2) a carry-store mode for delay-tolerant data. Both aim to minimize energy consumption while maximizing data collection. Specifically, we develop a time scheduling policy to allocate data and energy transmission time for delay-sensitive data within τmax, and propose a novel centralized control, distributed execution framework using decentralized deep reinforcement learning (DRL), termed DRL-Schedule, designed to optimize UAV navigation and ensure energy-efficient, timely data collection. Simulation results demonstrate that DRL-Schedule achieves an average improvement of approximately 11.2% and 12.8% over the best baseline PPO in energy efficiency, when varying different number of UAVs and IoTs, respectively, and notably surpassing GA, Greedy and Random. © 2025 IEEE.

关键词： 6G Deep reinforcement learning Energy-efficiency UAV-LEO integrated data collection WPT

来源：评论

学校读者我要写书评

暂无评论

SwinJSCC: Taming Swin Transformer for Deep Joint Source-Channel Coding

引用

IEEE TRANSACTIONS ON COGNITIVE COMMUNICATIONS AND networking 2025年第1期11卷 90-104页

作者： Yang, Ke Wang, Sixian Dai, Jincheng Qin, Xiaoqi Niu, Kai Zhang, Ping Beijing Univ Posts & Telecommun Key Lab Universal Wireless Commun Minist Educ Beijing 100876 Peoples R China Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China

As one of the key techniques to realize semantic communications, end-to-end optimized neural joint source-channel coding (JSCC) has made great progress over the past few years. A general trend in many recent works pushing the model adaptability or the application diversity of neural JSCC is based on the convolutional neural network (CNN) backbone, whose model capacity is yet limited, inherently leading to inferior system coding gain against traditional coded transmission systems. In this paper, we establish a new neural JSCC backbone that can also adapt flexibly to diverse channel conditions and transmission rates within a single model, our open-source project aims to promote the research in this field. Specifically, we show that with elaborate design, neural JSCC codec built on the emerging Swin Transformer backbone achieves superior performance than conventional neural JSCC codecs built upon CNN, while also requiring lower end-to-end processing latency. Paired with two spatial modulation modules that scale latent representations based on the channel state information and target transmission rate, our baseline SwinJSCC can further upgrade to a versatile version, which increases its capability to adapt to diverse channel conditions and rate configurations. Extensive experimental results show that our SwinJSCC achieves better or comparable performance versus the state-of-the-art engineered BPG + 5G LDPC coded transmission system with much faster end-to-end coding speed, especially for high-resolution images, in which case traditional CNN-based JSCC yet falls behind due to its limited model capacity.

关键词： Joint source-channel coding swin transformer attention mechanism image communications

来源：评论

学校读者我要写书评

暂无评论

Breaking the Inertial Thinking: Non-Blocking Multipath Congestion Control Based on the Single-Subflow Reinforcement Learning Model

引用

IEEE TRANSACTIONS ON NETWORK AND SERVICE MANAGEMENT 2024年第3期21卷 2876-2887页

作者： Wei, Dehui Zhang, Jiao Li, Haozhe Liu, Yuanjie Zhang, Xuan Pan, Tian Huang, Tao Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100088 Peoples R China Purple Mt Labs Nanjing 211111 Peoples R China

The Multipath TCP (MPTCP) protocol has received more attention due to the increasing number of terminals with multiple network interfaces. To meet the higher network performance demand of terminal services, many researches leverage reinforcement learning (RL) for MPTCP congestion control (CC) algorithms to improve the performance of MPTCP. However, we observe two limitations of existing RL-based mechanisms that make them impractical: 1) Fail to break the restriction of the input and output dimensions of RL, making the mechanisms unadaptable to the varying number of subflows. 2) Frequent model decisions block packet transmission, leading to under-utilization of bandwidth. This paper breaks the inertial thinking By "inertial thinking" here, we are referring to the initial reaction of others when dealing with CC in MPTCP. Given the interdependence between MPTCP subflows, scholars have traditionally opted for coupled CC. However, we have challenged this conventional thinking by independently handling the CC of different subflows in a single MPTCP flow and ensuring fairness. to overcome the above limitations and proposes Maggey, a non-blocking CC mechanism that applies the single-subflow model to multipath transmission. To this end, Maggey employs loosely coupled design principles and a unique reward function to ensure the fairness of the algorithm. Additionally, Maggey introduces iterative training to ensure the accuracy of training of the single-subflow model. Furthermore, a mode transition framework is artfully designed to avoid blocking, preserving the flexibility of RL-based CCs. These two features enhance the practicability of Maggey and the paper analyze the stability of Maggey. We implement Maggey in the Linux kernel and evaluate the performance of Maggey through extensive emulation and live experiments. The evaluation results show that Maggey boosts 26% throughput over DRL-CC at high bandwidth and improves 2%-60% throughput over traditional algorithms under diff

关键词： Training Bandwidth Adaptation models Throughput Packet loss Linux Kernel Multipath TCP congestion control reinforcement learning non-blocking

来源：评论

学校读者我要写书评

暂无评论

LEP: Leveraging Local Entropy Pruning for Sparsity in Large Language Models

LEP: Leveraging Local Entropy Pruning for Sparsity in Large ...

引用

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

作者： Chen, Yuli Cheng, Bo Zhang, Yingying Zhang, Shuhao Wu, Zhixuan Meng, Fanshen State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications China

ISBN: (纸本)9798350368741

The application of Large Language Models (LLMs) is rapidly expanding in fields such as natural language processing and computer vision. However, due to the enormous number of model parameters, while their emergent capabilities enhance performance, they also incur significant computational and storage costs, particularly during inference. This makes it particularly challenging to deploy large models in resource-constrained environments. To address this issue, we propose an innovative pruning method based on local entropy, aimed at improving model performance while reducing computational costs. Our approach leverages a small amount of calibration data and calculates the importance of weights based on the local entropy of weights and input activations, applying it to both unstructured and semi-structured pruning without the need for complex weight updates or retraining. Experimental results demonstrate that our method shows superior performance across models of various scales. Specifically, at a sparsity rate of 50%, our method significantly reduces model perplexity without the need for weight updates and outperforms existing pruning techniques such as SparseGPT and Wanda in zero-shot tasks. The code is available at: this https URL. © 2025 IEEE.

关键词： entropy large language models model compression unstructed pruning

来源：评论

学校读者我要写书评

暂无评论

AdaptLight: Toward Cross-Space-Time Collaboration for Adaptive Traffic Signal Control 20th

AdaptLight: Toward Cross-Space-Time Collaboration for Adapti...

引用

20th Pacific Rim International Conference on Artificial Intelligence

作者： Cai, Xintian Liu, Yilin Yuan, Quan Luo, Guiyang Li, Jinglin Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China

ISBN: (纸本)9789819970186;9789819970193

Recent multi-agent deep reinforcement learning (MADRL) approaches have shown notable benefits in traffic signal control. However, the spatial-temporal coupling, hysteresis, and heterogeneity of collaborative agents are usually ignored. states and actions among multiple intersections induce complex coupling and hysteresis in both space and time dimensions, while the actions also present spatial-temporal heterogeneity due to fluctuated traffic. These characteristics impose a critical impact on the efficiency and flexibility of coordinated control. In this paper, we propose AdaptLight, an MADRL-based model to achieve cross-space-time collaboration. It captures the interactions among spatial-temporal traffic components and exploits action repetition to adaptively adjust decision granularity for heterogeneous traffic. For the spatial-temporal coupling and hysteresis issue, AdaptLight first establishes a feature extraction network based on spatial-temporal graph Transformer. To tackle the spatial-temporal action heterogeneity problem, an action-repetition enabled MADRL module is designed, which can decide asynchronous-cooperative actions spanning multiple timesteps. Experiments present that AdaptLight shows competitive performance on different datasets.

关键词： Multi-agent deep reinforcement learning Traffic signal control Graph transformer Action repetition

来源：评论

学校读者我要写书评

暂无评论

Residual-Network Enabled Deep soft Interference Cancellation for MIMO Detection Without Channel state Information 99

Residual-Network Enabled Deep soft Interference Cancellation...

引用

IEEE 99th Vehicular technology Conference (VTC-Spring)

作者： Li, Lihua Li, Tong Fan, Siyi Ji, Xuan Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China

ISBN: (纸本)9798350387414

In order to meet the requirement of high spectrum efficiency of MIMO systems, the number of antennas has been greatly increased to support simultaneous communication of multiple users, and higher order modulation is expected to be employed in the system at the same time, which make signal detection a difficult task. However, most traditional algorithms reply on the channel state information (CSI) accuracy get by channel estimation at receivers. DeepSIC, developed in the existing literature, is a promising framework that introduces data-driven deep learning methods into iterative soft interference cancellation (SIC) algorithms that can address challenges of unknown CSIs and improved performance. In this paper, a signal detection algorithm is designed for MIMO systems under high-order modulation when CSI is unknown, called ResLight-DeepSIC. ResLight-DeepSIC is based on DeepSIC, introducing the idea of residual network (ResNet), and achieves betterdetection performance and lower complexity with lightweight network. Link-level simulation shows that the improved algorithm has a gain of about 1 dB on the symbol error rate (SER), and reduces complexity by about 37% and 63% under 16 QAM and 64 QAM modulation, respectively. Therefore, the ResLight-DeepSIC proposed in this article has advantages in both performance and complexity with large multiple stream transmission and high order modulation.

关键词： Interference cancellation Residual network deep learning Signal detection

来源：评论

学校读者我要写书评

暂无评论

Emotional boundaries and intensity aware model for incomplete multimodal sentiment analysis

引用

DIGITAL SIGNAL PROCESSING 2025年 160卷

作者： Zhang, Yuqing Xie, Dongliang Luo, Dawei Sun, Baosheng Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China Beijing Jinghang Res Inst Comp & Commun Beijing Peoples R China

Multimodal sentiment analysis aims to obtain comprehensive emotional features from multiple data sources when all modalities are accessible. However, in real-world scenarios, it is often impossible for all modalities to be available all the time. This issue leads to a significant degradation in the performance of multimodal sentiment analysis. There are two major challenges in this field: accurately identifying samples near the classification boundaries is difficult, and the recognition performance varies significantly among different modality combinations. In this article, we propose an Emotional Boundaries and Emotional Intensity Aware (EBIA) model to enhance the robustness of incomplete multimodal sentiment analysis. Specifically, we design a Boundary Fuzzy Aware (BFA) module to learn the intra-class and inter-class consistency of samples, transferring full modalities integrity information to the missing modality environment, and prompting the model to focus on samples near the class boundaries. Additionally, we introduce a Weak Modality Aware (WMA) module that calculates additional predictions for each modality, guiding the model to focus on weak modality combinations. Extensive experiments and analyses conducted on three popular benchmark datasets demonstrate the effectiveness of our proposed method compared with several baseline methods.

关键词： Multimodal sentiment analysis Missing modality Classification boundary fuzzy Weak modality aware

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：