检索结果-内蒙古大学图书馆

36th IEEE/IFIP Network Operations and Management Symposium, NOMS 2023

作者： Zhou, Yanbo Feng, Lei Xie, Kunyi Zhou, Fanqin Li, Wenjing Yu, Peng Beijing University of Posts and Telecommunications The State Key Laboratory of Networking and Switching Technology Beijing China

ISBN: (纸本)9781665477161

The high-precision clock synchronization is pursued with the consideration of the balance for time-domain resource utilization, when 5G technologies are expected to carry the timing services. Firstly, a clock synchronization model is established in the case that the system exists the observed value loss. The probability distribution of the loss of the observed value is evaluated by a newly proposed delay deterministic confidence method. Then, the error boundness of the clock synchronization is investigated by the Kalman filter algorithm, and the optimization problem for 5G time-domain resource configuration is formulated with guaranteeing the precision of clock synchronization;Finally, the proposed optimization problem is solved by dueling double deep Q-learning-based Lyapunov optimization. The experimental results verify the effectiveness and superiority of the proposed method in terms of the joint optimization of clock synchronization error covariance and throughput. © 2023 IEEE.

关键词： Optimization

来源：评论

学校读者我要写书评

暂无评论

BlindSpotEliminator: Collaborative Point Cloud Perception in Cellular-V2X Networks

BlindSpotEliminator: Collaborative Point Cloud Perception in...

引用

2023 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2023

作者： Chen, Ziyue Luo, Guiyang Shao, Congzhang Yuan, Quan Li, Jinglin Beijing University of Posts and Telecommunications State Key Laboratory of Networking and Switching Technology Beijing China

ISBN: (纸本)9798350337020

Multi-agent collaborative perception depends on sharing sensory information to improve perception accuracy and robustness, as well as to extend coverage. However, most collaborative perception methods ignore the limitations of communication networks, such as limited bandwidth and the possibility of wireless conflicts. To fill this gap, this paper proposes BlindSpotEliminator, a conflict-free scheduler over the cellular-V2X networks for supporting practical collaborative point cloud perception to eliminate blind spots. BlindSpotEliminator first identifies the blind spots for each vehicle, then lists the corresponding conflict relationships based on the distribution of the blind spots and communication conflicts, and finally designs an optimized point cloud data transmission strategy to eliminate the blind spots of each vehicle. Extensive experiments show that compared with greedy algorithm and random methods, BlindSpotEliminator achieves better efficiency, i.e., transmitting 20% more point cloud data. © 2023 IEEE.

关键词： Cellular V2X collaborative perception optimization strategy point cloud data transmission

来源：评论

学校读者我要写书评

暂无评论

Intelligent Energy-Efficiency Trajectory Planning of Heterogeneous Air Base Stations for 6G Emergency Communication

Intelligent Energy-Efficiency Trajectory Planning of Heterog...

引用

2023 IEEE/CIC International Conference on Communications in China, ICCC Workshops 2023

作者： Tian, Jingyue Yu, Peng Li, Wenjing Feng, Lei Zhou, Fanqin Beijing University of Posts and Telecommunications State Key Laboratory of Networking and Switching Technology Beijing China

ISBN: (纸本)9798350345407

With the development of 6G, emergency communication services upgrade and the need for edge intelligence is increasing. However, today's 6G emergency communication scenarios are changeable and complex, and existing methods cannot achieve large-scale and QoS guarantees. So we proposed a heterogeneous coverage compensation mechanism with static complete deployment and dynamic enhancement deployment. To enhance the capacity of hotpots determined by the clustering algorithm, we jointly optimize the UAV trajectory and users' connectivity to maximize the UAV's energy efficiency(EE). The optimized problem is NP-hard and is solved by deep reinforcement learning(DRL) algorithm. Simulation results show our method can significantly improve EE, which is much higher than Q-learning, PSO, and GA. © 2023 IEEE.

关键词： Trajectories

来源：评论

学校读者我要写书评

暂无评论

A Federated Learning Approach for Net Load Forecasting in Microgrids. 24

A Federated Learning Approach for Net Load Forecasting in Mi...

引用

24th Asia-Pacific Network Operations and Management Symposium, APNOMS 2023

作者： Qin, Zihan Shao, Sujie Guo, Shaoyong Qiu, Xuesong State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China

ISBN: (纸本)9788995004395

Microgrid is a small-scale power grid composed of various distributed energy resources, and net load forecasting (NLF) plays significant role in its energy generation planning. Due to the dispersed distribution of nodes within the microgrid and disparate user data, completing the net load forecasting task becomes challenging. In order to efficiently forecast the net load while ensuring data security, this paper proposes an approach for NLF based on federated learning (FL). Firstly, a FL framework for NLF in microgrids is constructed to complete the NLF task with long short-term memory neural network (LSTM) model and node selection (NS) in aggregation. Secondly, considering factors such as the data precision and quantity, an accuracy optimization problem model is proposed. Then, the problem is formulated as a Markov decision process (MDP), and a NS algorithm based on Asynchronous Advantage Actor-Critic (A3C) is designed for selecting a node set to optimize model aggregation. The simulation demonstrate that the proposed NS method based on A3C improves the performance of FL, resulting in better convergence of LSTM network and ensuring the scalability of the cluster in microgrid. Copyright 2023 KICS.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Multi-Relational Knowledge Enhanced Contrastive Learning for Event Extraction 9

Multi-Relational Knowledge Enhanced Contrastive Learning for...

引用

9th International Conference on Computer and Communications, ICCC 2023

作者： Liu, Yixuan Shuang, Kai Beijing University of Posts and Telecommunications State Key Laboratory of Networking and Switching Technology Beijing China

ISBN: (纸本)9798350317251

Event extraction (EE) is a fundamental task under the field of natural language processing. Most of the existing EE methods perform supervised fine-tuning, training downstream with the help of the ability to pre-train language models (PLMs), limited labeled data is difficult to satisfy the training needs of PLMs and faces problems such as limited generalization. Contrastive learning, as a self-supervised training paradigm, improves the representational power of PLMs by acquiring a wider range of knowledge from large-scale unlabeled corpora than is implied by the data labels. However, existing contrastive learning methods mainly focus on the sentence-level, while the development at the token-level represented by EE tasks is still in its early stages, resulting in noisy representations of tokens and a lack of detailed use of event knowledge, which makes the performance bottleneck of the PLMs after contrastive learning when fine-tuning downstream EE tasks. To this end, we propose MRKE-CLEE, a multi-relational knowledge enhanced contrastive learning for event extraction. MRKE-CLEE first accomplishes the coreference resolution and AMR parsing as supported tasks to obtain multiple event-relationship knowledge pseudo-labels from a large-scale unlabeled corpus through the auxiliary model, then uses in-sentence knowledge guided contrast to filter noisy contrastive pairs to make the resulting token representations more distinct, and using event type knowledge guided contrast to aggregate token representations with the same event type. With the combination of multi-relational event knowledge, using contrastive learning to improve the performance of event detection (ED) and event argument extraction (EAE). Experiments demonstrate that our MRKE-CLEE achieves a new state-of-the-art performance on the ACE2005 and MAVEN datasets and introduces new attempts to use token-level contrastive learning in EE tasks. © 2023 IEEE.

关键词： Extraction

来源：评论

学校读者我要写书评

暂无评论

Accelerate Single Image Super-Resolution Using Object Detection Process

引用

Computers, Materials & Continua 2023年第8期76卷 1585-1597页

作者： Xiaolin Xing Shujie Yang Bohan Li State Key Laboratory of Networking and Switching Technology Beijing University of Posts and TelecommunicationsBeijing100876China Department of Applied Math and Statistics The Johns Hopkins UniversityBaltimore21211USA

Image Super-Resolution(SR)research has achieved great success with powerful neural *** deeper networks with more parameters improve the restoration quality but add the computation complexity,which means more inference time would be cost,hindering image SR from practical *** the spatial distribution of the objects or things in images,a twostage local objects SR system is proposed,which consists of two modules,the object detection module and the SR ***,You Only Look Once(YOLO),which is efficient in generic object detection tasks,is selected to detect the input images for obtaining objects of interest,then put them into the SR module and output corresponding High-Resolution(HR)*** computational power consumption of image SR is optimized by reducing the resolution of input *** addition,we establish a dataset,TrafficSign500,for our ***,the performance of the proposed system is evaluated under several state-Of-The-Art(SOTA)YOLOv5 and SISR *** show that our system can achieve a tremendous computation improvement in image SR.

关键词： Object detection super-resolution computation complexity YOLOv5 inference time objects of interest

来源：评论

学校读者我要写书评

暂无评论

Fine-Grained Flow Control Agent on Path MTU for IoT Software 23

Fine-Grained Flow Control Agent on Path MTU for IoT Software

引用

14th Asia-Pacific Symposium on Internetware, Internetware 2023

作者： He, Hongchuan Chen, Dezhi Zhuang, Zirui Qi, Qi Zhang, Lejian Xu, Tong Wang, Jingyu Beijing University of Posts and Telecommunications State Key Laboratory of Networking and Switching Technology Beijing China

ISBN: (纸本)9798400708947

Internet of Things (IoT) software is used to control the distributed hardware of the underlying network and provide a reliable operating platform for various services. In production system, diversity IoT software provides multiple services, flows of different software run in parallel on the same IoT platform. Thus, system-level network parameter configurations may not be suitable for all service needs. In this paper, we focus on the challenge of the personally parameterizing transmission unit size and congestion windows (CWND) in flow control. We propose deeper flow control software model (DeepFC) for finer-grained flow control than traditional algorithms. DeepFC consist of two parts: (i) Since system-level transmission unit size may degrade network performance due to frequent fragmentation, we combine path MTU (PMTU) and deep reinforcement learning (DRL) to predict fine-grained flow-level transmission unit size. (ii) Transmission unit size is related to CWND in flow control. The fine-grained transmission unit size needs fine-grained congestion control solution. In DeepFC, we consider the mutual coupling between transmission unit size and CWND parameter configuration to further improve network performance. Experimental results show that DeepFC can reduce fragmentation by 67.8% compared to the protocols with system-level transmission unit size, flow completion time can be reduced by 20.93%, and throughput can be increased by 19.96% compared to the average of benchmark algorithms. © 2023 ACM.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Long-tailed Detection Based on Multi-Expert Aggregation

Long-tailed Detection Based on Multi-Expert Aggregation

引用

2023 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2023

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

ISBN: (纸本)9798350337020

Long-tailed data distribution is a common phenomenon observed in nature. However, this distribution poses challenges for neural network training, as it performs well on samples with head categories, but not so well on tail categories - precisely what requires attention in certain scenes. Although existing methods involve re-sampling data or designing loss functions, they often fail to adapt well to changing data distributions. In this article, we propose a long-tailed detection approach based on multi-expert aggregation (MEA). This method employs multi-expert networks to learn about head, medium, and tail categories from a fixed long-tailed distribution. Different data are modelled using different loss designs, and the results of multi-expert networks are aggregated to achieve good accuracy for both head and tail categories. The effectiveness of this method has been verified on LVIS. © 2023 IEEE.

关键词： deep learning long-tailed recognition multi-expert learning object detection

来源：评论

学校读者我要写书评

暂无评论

Age-Aware Status Updating for Solar Energy Storage Systems 23

Age-Aware Status Updating for Solar Energy Storage Systems

引用

23rd IEEE International Conference on Communication technology, ICCT 2023

作者： Tian, YuJia Meng, Yue Qin, Xiaoqi Ma, Nan Beijing University of Posts and Telecommunications State Key Laboratory of Networking and Switching Technology Beijing China

ISBN: (纸本)9798350325959

Solar Energy Storage Systems (SESS) are deployed to address the unstable and intermittent nature of photovoltaic (PV) systems by allowing swift adjustments to system power. In such scenarios, the freshness of collected status updates is critical for controlling system performance. This timeliness is quantified through the Age of Information (AoI) metric. LoRa is introduced to the SESS as a Low-Power Wide-Area Network technology (LPWAN) to ensure reliable and energy-efficient data transmission. We explore a load monitoring system employing LoRa devices with periodic Slotted ALOHA (SA) access, where data updates with unfixed priority since the load process can be modeled as the time-variant Ornstein-Uhlenbeck (OU) process. The long-term expectation of AoI under fixed priorities is derived while the monitored process is time-invariant. Then the estimation error minimization problem is formulated while monitoring the time-variant OU process, which jointly considers delivered samples, estimated parameters, and AoI. To efficiently solve the problem, a greedy maximum scheduling (GMS) algorithm is proposed. Simulation results validate the AoI analysis with time-invariant processes and demonstrate the effectiveness of our data updating and estimation strategy using both generated data and real industry park data. © 2023 IEEE.

关键词： Age of Information Internet of Things Load monitoring Solar Energy Storage Systems

来源：评论

学校读者我要写书评

暂无评论

DT-PBFT: A Double-Layer Group Consensus Algorithm of Credibility for IoT Blockchain 2

DT-PBFT: A Double-Layer Group Consensus Algorithm of Credibi...

引用

2nd International Conference on Big Data, Information and Computer Network, BDICN 2023

作者： Chen, Yaxin Jia, Yongpu Beijing University of Posts and Telecommunications State Key Laboratory of Networking and Switching Technology Beijing China

ISBN: (纸本)9798350321364

Blockchain technology is becoming more and more popular, but performance problems have always troubled it, especially the scale of distributed networks is limited. Like Practical Byzantine consensus algorithm (PBFT), the main bottlenecks are concentrated in the scope of consensus too small, and the communication is too high. In order to implement PBFT in large-scale systems such as large-scale IoT ecosystems and blockchains, this paper proposes a reputation-based two-layer improved PBFT consensus computing model DT-PBFT. In order to better combine with the large -scale Internet of Things ecosystem, we give each node regional location, communication capacity, computing power and other related attributes to each involvement, so that Client nodes can input node attributes as parameters into the reputation calculation algorithm, to select the master node of each layer in the double-layer consensus model. Determining the master node through reputation improves the credibility of the consensus algorithm. At the same time, the DTPBFT cancels the Pre-Prepare stage, reducing communication complexity. Finally, this paper transforms the existing PBFT algorithm on the Fabric platform to implement the DT-PBFT consensus algorithm and compares its performance with traditional PBFT, proving that DT-PBFT has the best balance and excellent comprehensive performance in large-scale IoT application scenarios. © 2023 IEEE.

关键词： Blockchain

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：