检索结果-内蒙古大学图书馆

arXiv 2022年

作者： Liu, Chuanhong Guo, Caili Yang, Yang Jiang, Nan The Beijing Key Laboratory of Network System Architecture and Convergence School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing100876 China The Beijing Laboratory of Advanced Information Networks School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing100876 China

Task-oriented communication is a new paradigm that aims at providing efficient connectivity for accomplishing intelligent tasks rather than the reception of every transmitted bit. In this paper, a deep learning-based task-oriented communication architecture is proposed where the user extracts, compresses and transmits semantics in an end-to-end (E2E) manner. Furthermore, an approach is proposed to compress the semantics according to their importance relevant to the task, namely, adaptable semantic compression (ASC). Assuming a delay-intolerant system, supporting multiple users indicates a problem that executing with the higher compression ratio requires fewer channel resources but leads to the distortion of semantics, while executing with the lower compression ratio requires more channel resources and thus may lead to a transmission failure due to delay constraint. To solve the problem, both compression ratio and resource allocation are optimized for the task-oriented communication system to maximize the success probability of tasks. Specifically, due to the nonconvexity of the problem, we propose a compression ratio and resource allocation (CRRA) algorithm by separating the problem into two subproblems and solving iteratively to obtain the convergent solution. Furthermore, considering the scenarios where users have various service levels, a compression ratio, resource allocation, and user selection (CRRAUS) algorithm is proposed to deal with the problem. In CRRAUS, users are adaptively selected to complete the corresponding intelligent tasks based on branch and bound method at the expense of higher algorithm complexity compared with CRRA. Simulation results show that the proposed ASC approach can reduce the size of transmitted data by up to 80% without reducing task success, and the proposed CRRA and CRRAUS algorithms can obtain at least 15% and 10% success gains over baseline algorithms, respectively. Copyright © 2022, The Authors. All rights reserved.

关键词： Resource allocation

来源：评论

学校读者我要写书评

暂无评论

Hard-sample Guided Hybrid Contrast Learning for Unsupervised Person Re-Identification

arXiv

引用

arXiv 2021年

作者： Hu, Zheng Zhu, Chuang He, Gang Beijing Laboratory of Advanced Information Networks Beijing Key Laboratory of Network System Architecture and Convergence Beijing University of Posts and Telecommunications Beijing100876 China

Unsupervised person re-identification (Re-ID) is a promising and very challenging research problem in computer vision. Learning robust and discriminative features with unlabeled data is of central importance to Re-ID. Recently, more attention has been paid to unsupervised ReID algorithms based on clustered pseudo-label. However, the previous approaches did not fully exploit information of hard samples, simply using cluster centroid or all instances for contrastive learning. In this paper, we propose a Hard-sample Guided Hybrid Contrast Learning (HHCL) approach combining cluster-level loss with instance-level loss for unsupervised person Re-ID. Our approach applies cluster centroid contrastive loss to ensure that the network is updated in a more stable way. Meanwhile, introduction of a hard instance contrastive loss further mines the discriminative information. Extensive experiments on two popular large-scale Re-ID benchmarks demonstrate that our HHCL outperforms previous state-of-the-art methods and significantly improves the performance of unsupervised person Re-ID. The code of our work and dataset are available soon at https://***/bupt-ai-cz/HHCL-ReID. Copyright © 2021, The Authors. All rights reserved.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

Task-Oriented Image Semantic Communication Based on Rate-Distortion Theory

arXiv

引用

arXiv 2022年

作者： Liu, Fangfang Tong, Wanjie Yang, Yang Sun, Zhengfen Guo, Caili The Beijing Key Laboratory of Network System Architecture and Convergence School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing100876 China The Beijing Laboratory of Advanced Information Net Works School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing100876 China

Task-oriented image semantic communication is a new communication paradigm, which aims to transmit semantics for artificial intelligent (AI) tasks while ignoring the reconstruction quality of the images. However, in some applications, such as autonomous driving, both image reconstruction quality and the performance of the followed AI tasks must be simultaneously considered. To tackle this challenge, this paper proposes a task-oriented semantic communication scheme with semantic reconstruction (TOSC-SR). Its main goal is to simultaneously minimize pixel-level and task-relevant semantic-level distortion during communications under a certain rate, which formulates a new rate-distortion optimization problem. To successfully measure the loss at the semantic level, a new form of semantic distortion measured by the mutual information between the semantic-reconstructed images and the task labels is proposed. Then, we derive an analytical solution for the formulated problem, where the self-consistent equations of the problem are obtained to determine the optimal mapping of the source and the semantic-reconstructed images. To implement TOSC-SR, we further obtain an extended form of rate-distortion form based on the variational approximation of mutual information, which is applicable to multiple AI tasks. Experimental results show that the proposed approach outperforms the traditional JPEG, JPEG2000, BPG, VVC-based image communication systems and deep learning based benchmarks in terms of image reconstruction quality, AI task performance, and multi-task generalization ability. © 2022, CC BY.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Mobility-Aware Proactive Video Segment Caching Based on Deep Reinforcement Learning

Mobility-Aware Proactive Video Segment Caching Based on Deep...

引用

IEEE International Conference on network Infrastructure and Digital Content (IC-NIDC)

作者： Xuefei Li Jiawei Wang Zhilong Zhang Danpu Liu Beijing Laboratory of Advanced Information Network Beijing University of Posts and Telecommunications Beijing China Beijing Key Laboratory of Network System Architecture and Convergence Beijing University of Posts and Telecommunications Beijing China

ISBN: (纸本)9781665405836

Maintaining efficient and successive video streaming services in cellular networks is challenging due to user mobility and ever-increasing volume of data traffic. A promising solution is to cache popular contents at the edge of wireless networks. Although caching schemes have been widely discussed, few of them jointly considered the characteristics of streamed video data and user mobility. In this paper, we construct a dynamic caching decision framework based on Long Short-Term Memory (LSTM) and Deep Q-network (DQN). Based on this framework, a mobility-aware segment-level caching strategy is proposed to maximize the cache hit rate. Simulation results show that our proposed method can achieve 20% performance improvement by comparing with baseline caching algorithms.

关键词： Cellular networks Simulation Wireless networks Heuristic algorithms Conferences Reinforcement learning Bandwidth

来源：评论

学校读者我要写书评

暂无评论

Dual path nets with a new multi-scale backbone architecture for pulmonary nodule detection

Dual path nets with a new multi-scale backbone architecture ...

引用

2020 IEEE International Conference on Information Technology, Big Data and Artificial Intelligence, ICIBA 2020

作者： Maozhi, Gong Muqing, Wu Beijing Laboratory of Advanced Information Networks Beijing Key Laboratory of Network System Architecture and Convergence Acknowledgment Beijing University of Posts and Telecommunications School of Information and Communication Engineering Beijing China

ISBN: (纸本)9781728152240

This article mainly completed the detection of lung nodules using deep neural network and built a complete end-to-end detection system. U-net structure has significant advantages in image segmentation. We propose a dual-path network with a U-shaped structure that combines the advantages of residual learning and intensive learning for the detection of lung nodules. And on this basis, the residual module of Resnet was modified. The feature image fed into the residual module is divided into multiple blocks for convolution. Smaller residual blocks are added to each part to broaden the receptive field of each layer. So, it has the ability of extracting multi-scale feature. This network structure is called Res2net. The processed feature images are fused. We also added the SE module to focus on the relationship between feature information channels. It can strengthen the role of useful channels and suppress the output of useless channels. These changes significantly improve the detection effect of lung nodules. We used the position regression of lung nodules to return the predicted position of lung nodules and achieved good results. The detection indicator FROC is significantly better than other detection projects. © 2020 IEEE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

DPCA-Net: Dual Path with 3D Channel Attention for Pulmonary Nodule Detection 6

DPCA-Net: Dual Path with 3D Channel Attention for Pulmonary ...

引用

6th IEEE International Conference on Computer and Communications, ICCC 2020

作者： Wang, Ruinan Wu, Muqing Beijing Laboratory of Advanced Information Networks Beijing Key Laboratory of Network System Architecture and Convergence Acknowledgment Beijing University of Posts and Telecommunications School of Information and Communication Engineering Beijing China

ISBN: (数字)9781728186351

ISBN: (纸本)9781728186351

Lung cancer affects human health severely and pulmonary nodule is an important manifestation of lung cancer. The diagnosis of lung cancer requires detection of each suspicious nodule. Early detection of lung cancer can greatly improve the survival rate of patients. It is very important to accurately find the location of pulmonary nodules in CT images and label them. Currently, CT is the most commonly used method to detect pulmonary nodules. In this paper, inspired by convolutional neural networks and the mechanisms of human attention, we propose a novel architectural block, which we term the Dual Path-Channel Attention (DPCA) block. This block can effectively prevent vanishing gradient phenomenon, enhance the transmission of context information and can weight related features according to the importance of different channels. We further propose a simple, highly efficient 3D pulmonary nodule detection model named DPCA-Net, which is constructed by DPCA block as a key component. Compared with other excellent pulmonary nodule detection models, our model achieves effective improvement on the public large-scale LUNA16 dataset. © 2020 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning for Task Offloading in Mobile Edge Computing for SDN based Wireless networks 7

Reinforcement Learning for Task Offloading in Mobile Edge Co...

引用

7th International Conference on Software Defined systems, SDS 2020

作者： Kiran, Nahida Pan, Chunyu Changchuan, Yin University of Posts and Telecommunications Beijing Key Laboratory of Network Systeiii Architecture and Convergence Beijing Beijing China

ISBN: (纸本)9781728172194

The explosive growth of the distributed computing resources in mobile edge computing (MEC) create a necessity to have a reasonable controller to ensure efficient utilization of distributed computing resources. The software defined networks (SDNs) capabilities can be used in MEC environment to reduce the energy consumption of user device without service disruptions. SDN, which is not proposed specifically for edge computing, can in fact serve as an enabler to let the real potential of edge computing be achieved and to lower the complexity barriers involved. This paper investigates resource allocation problem in wireless MEC by aiming to save the battery power of user device. Learning from experience plays an important role in a SDN based MEC infrastructure where reinforcement learning (RL) takes a long term goal into consideration besides immediate reward, which is very important for a dynamic environment. A novel software defined edge cloudlet (SDEC) based RL optimization framework is proposed in this paper to tackle the energy minimization problem in wireless MEC. Specifically, Q-learning and cooperative Q-learning based RL schemes are proposed for the intractable problem. Simulation results reveal that the proposed scheme achieves superior performance in saving the battery power of a user device compared to other benchmark methods such as Q-learning with a random algorithm and Q-learning with epsilon greedy. © 2020 IEEE.

关键词： Wireless networks

来源：评论

学校读者我要写书评

暂无评论

Time Error Compensation with Neural network for Time Synchronization in TSN

Time Error Compensation with Neural Network for Time Synchro...

引用

International Conference on Computer and Communications (ICCC)

作者： Haoze Yu Guochu Shou Xiaodong Zhang Ledong Hu Yaqiong Liu Yihong Hu Beijing Key Laboratory of Network System Architecture and Convergence School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing China

ISBN: (纸本)9781665409513

Time-sensitive networking featured with precise time synchronization, is significant in modern industrial systems for real-time guarantee. However, some industrial scenarios, such as smart grids and 5G fronthaul networks, have much higher requirements on time synchronization. This paper proposes a method to improve the synchronization performance of TSN with a neural network model. A time error compensation algorithm including timestamp data processing, neural network training, and error compensation is designed. Comparative experiment results show that this algorithm improves the stability and accuracy of time synchronization.

关键词： Training Time-frequency analysis Neural networks Error compensation Power system stability Stability analysis Real-time systems

来源：评论

学校读者我要写书评

暂无评论

Multi-domain Time Synchronization Model and Performance Evaluation in TSN

Multi-domain Time Synchronization Model and Performance Eval...

引用

International Conference on Computer and Communications (ICCC)

作者： Ledong Hu Guochu Shou Xiaodong Zhang Yaqiong Liu Yihong Hu Beijing Key Laboratory of Network System Architecture and Convergence School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing China

ISBN: (纸本)9781665409513

Time-Sensitive networking (TSN) has precise time synchronization capabilities and can provide reliable time-deterministic services for industrial communications and other fields. IEEE 802.1AS is used as the time synchronization protocol in TSN and has support for multiple gPTP domains, which further expands the application range of TSN and enhances the synchronization capabilities of TSN. Industrial automation and other use cases also have multi-domain time synchronization requirements of TSN, but there is still a lack of research in this field. Therefore, this paper studies the multi-domain synchronization mechanism in TSN and proposes the multi-domain time synchronization model according to actual application requirements. Based on the multi-domain model, this paper analyzes the time errors in multi-domain time synchronization, introduces the SDN paradigm, and proposes a software-defined time compensation method to compensate the time errors centrally. Further, the time synchronization experiment environment is built to analyze synchronization performance. The results verify the effectiveness of the multi-domain time synchronization model, demonstrating that the multi-domain TSN network can provide high-precision time synchronization services.

关键词： Performance evaluation Analytical models Automation Protocols Industrial communication Synchronization Reliability

来源：评论

学校读者我要写书评

暂无评论

An Effective Controller Placement Algorithm Based on Clustering in SDN 6

An Effective Controller Placement Algorithm Based on Cluster...

引用

6th IEEE International Conference on Computer and Communications, ICCC 2020

作者： Yujie, Rong Muqing, Wu Yiming, Chen Beijing University of Posts and Telecommunications School of Information and Communication Engineering Beijing Laboratory of Advanced Information Networks Beijing Key Laboratory of Network System Architecture and Convergence Acknowledgment Beijing China

ISBN: (纸本)9781728186351

Different from the traditional TCP/IP network, Software Defined network (SDN) uses a layered architecture to decouple the control plane and data plane. The controller is the core device of the control plane, and the control plane is mainly responsible for the decision-making control of the routing strategy. Because the processing power of a single controller is limited, the multi-controller network architecture has become an inevitable choice in the WAN. The layout of the controllers in SDN has an impact on the performance of the entire network. The controller placement problem (CPP) is an NP-hard problem. The challenges of CPP solution are as follows. 1) the number of controllers;2) the position of the controllers and the position of the switches under their management in the network topology. The clustering by fast search and find of density peaks (CFSFDP) algorithm is a density-based clustering algorithm. It has few initial parameters and does not require iterative solution. However, the important parameter dc of the CFSFDP algorithm is usually selected based on experience, and the number of cluster centers cannot be determined automatically. This paper proposes a new controller placement algorithm by improving the CFSFDP algorithm. We use the idea of information entropy to automatically determine the dc value, and automatically determine the number and location of the controllers based on the jumping point in the decision graph. It can be seen from the experimental results that compared with random placement, k-means and CFSFDP algorithm, our proposed algorithm can effectively reduce network propagation delay and improve network performance. © 2020 IEEE.

关键词： Controllers

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：