检索结果-内蒙古大学图书馆

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Zeyu Jiang Xiaohong Liu Guoxing Yang Weizhi Li Aini Li Guangyu Wang State Key Lab of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China UCL Cancer Institute University College London London United Kingdom Department of Mathematics and Theories Peng Cheng Laboratory Shenzhen China

In communication systems, the challenge of ensuring accurate data transmission across noisy channels remains paramount. While semantic communication shows potential in improving image transmission and reconstruction, existing methods still suffer from perceptual quality degradation in high-noise environments. To address these issues, we introduce DiffSC, a novel semantic communication framework that integrates the Diffusion Probabilistic Model (DPM). Within DPM, Gaussian noise modeling is leveraged to facilitate enhanced image generation. DiffSC is trained to recover semantic information compromised during transmission, amplifying its capabilities in both image reconstruction and denoising. Additionally, we propose a Multi-Dimensional Feature Extraction Module (MFM) that employs multiple convolution kernels along with channel and spatial attention mechanisms to enrich encoding and decoding. Experimental results demonstrate that DiffSC significantly outperforms existing systems, improving SSIM by more than 19% and PSNR by more than 11% compared to Deep JSCC. Further ablation studies demonstrate the effectiveness of our proposed denoiser and feature extraction modules.

关键词：

来源：评论

学校读者我要写书评

暂无评论

TAMM: A Task-Adaptive Multi-Modal Fusion Network for Facial-Related Health Assessments on 3D Facial Images

TAMM: A Task-Adaptive Multi-Modal Fusion Network for Facial-...

引用

IEEE International Conference on Image Processing

作者： Kai Wang Xiaohong Liu Zhengchao Luo Fajin Feng Guangyu Wang College of Future Technology Peking University Beijing China UCL Cancer Institute University College London London UK State Key Lab of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China

Previous studies showed that facial appearance is an important phenotypic indicator of human diseases or biological conditions. Recent advancements in deep learning have shown great potential in facial image analysis, including health status assessments. However, prior methods mainly focused on single modality analysis of either the 2D texture images or 3D facial meshes, which are limited in their ability to fully capture the relationships between biometric measurements and diseases. To address these issues, we propose a task-adaptive multi-modal fusion network, TAMM, for facial-related health assessments. Our model leverages both the geometric and texture features of 3D facial images by a task-adaptive Transformer (TAFormer), which can dynamically extract features from different modalities and scales for various tasks via spatial attention and cross modal multi-scale attention, effectively capture intra- and inter-modal relationships between features. Experimental results on a dataset of 19,775 patients demonstrate that TAMM achieves the state-of-the-art performance on various regression and classification tasks including age, BMI, and fatty liver disease predictions. Ablation studies shows the importance of multi-modal fusion and task-specific adaptability of our model in achieving optimal performance.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multi-Domain Dialog state Tracking based on Machine Reading Comprehension 2021

Multi-Domain Dialog State Tracking based on Machine Reading ...

引用

4th International Conference on Signal Processing and Machine Learning, SPML 2021

作者： Wang, Guanqun Liang, Zhengyu Zhang, Zheng Xiong, Yongping Consulting Center of State Grid Science and Technology Projects State Grid Energy Research Institute Co. Ltd Beijing102209 China State Key Lab. of Switching and Networking Technology Beijing University of Posts and Telecommunications China

ISBN: (纸本)9781450390170

With the development of big data and deep learning technology, the goal of creating an automatic human-machine dialog system is no longer an illusion. However, obtaining correct dialog label is very challenging due to expensive purchase cost and time cost. Through the transfer of knowledge, we proposed RC_DST which enables the dialog state tracking model to infer the state of dialog between different domains accurately and improves the accuracy in a new domain without or with a few of labeled data. In this paper, we utilizes XLNet which performs well in processing long texts to encode the dependency between the dialog context and slot semantics. Meanwhile, we expands machine reading comprehension to non-categorical and categorical slots in different ways. Extensive experiment results show that our method achieves competitive results in a new domain with zero samples compared with exclusive training data in this domain. © 2021 ACM.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Measurement-based optical path loss model for indoor visible light communication

Measurement-based optical path loss model for indoor visible...

引用

IEEE Globecom Workshops

作者： Yu Tong Pan Tang Jianhua Zhang Yue Yin Shuo Liu Baobao Liu Baoling Liu Guangyi Liu Liang Xia State Key Lab of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China China Mobile Research Institute Beijing China

ISBN: (纸本)9781665459761

In this paper, we first present the optical path loss (OPL) measurements for visible light communication (VLC) in two typical indoor scenarios, i.e., the corridor and the conference room. Through analyzing the Floating-intercept (FI) model results of measured OPL, the path loss exponent (PLE) for the FI model in the conference room (1.83) is larger than the PLE in the corridor (1.77) at 625 nm case. Moreover, the maximum difference in OPL between variable light wavelengths for two typical scenarios can reach about 2.26 dB. This indicates that the OPL is highly dependent on the scenario size, the propagation distance, and the light wavelength. Thus, an OPL model is further proposed as a function of the propagation distance and the light wavelength. Furthermore, the model parameters for those two indoor scenarios are extracted from the measured data. The small standard derivation of the derived parameters shows that our proposed model has the goodness of fit to the measured OPL.

关键词： Optical losses Analytical models Wavelength measurement Optical propagation Optical variables measurement Loss measurement Data models

来源：评论

学校读者我要写书评

暂无评论

Self-adaptive and Efficient Training Node Selection for Federated Learning in B5G/6G Edge Network

Self-adaptive and Efficient Training Node Selection for Fede...

引用

IEEE Symposium on Network Operations and Management

作者： Can Tan Peng Yu Wenjing Li Fanqin Zhou Ying Wang Siya Xu Xuesong Qiu Qingbi Zheng Pei Xiao State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China Future Research Lab China Mobile Research Institute Beijing China University of Surrey United Kingdom

In the upcoming B5G/6G era, devices will generate a amount of heterogeneous data at the network edge. As a paradigm for implementing distributed and privacy-preserving machine learning (ML), Federated Learning (FL) has drawn great attention to secure data sharing in edge networks. However, FL takes too much time and communication resources to train and transmit model parameters, which is unaffordable for edge devices with limited capabilities. To achieve a trade-off between resource and efficiency, it is crucial to select appropriate training nodes. While existing works about node selection focus on the resources allocation and pay less attention to the node mobility and seamless service. In this paper, we considering mobility, computation capability, and transmission power of training nodes to minimize the FL system cost. We propose an algorithm and mechanism respectively for different scenarios of node speed. An algorithm based on Deep Reinforcement Learning (DRL) matches with stationary and low-speed training nodes. A heuristic mechanism is used for nodes with high mobility. Simulation results show that the proposed schemes select appropriate training nodes effectively, and reduce the system cost by up to 20%.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Impact of Device Specifications on VLC Path Loss and System Performance: An Empirical Study

Impact of Device Specifications on VLC Path Loss and System ...

引用

IEEE International Conference on Communications in China (ICCC)

作者： Shuo Liu Pan Tang Yu Tong Yue Yin Linchao Li Anrui Niu Liang Xia Jianhua Zhang State Key Lab of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China State Key Laboratory of Information Photonics and Optical Communications Beijing University of Posts and Telecommunications Beijing China China Mobile Research Institute Beijing China

Visible light communication (VLC) is a promising technology for sixth-generation (6G) networks due to its unlicensed bands and high transmission rates. In recent years, several studies have explored the physical propagation characteristics of VLC channels. Nevertheless, the existing models are limited by reliance on customized optical devices. In addition, only sporadic studies have investigated the influence of device specifications, yet neglecting the joint effects of the transmitter and receiver specifications. In this paper, we experimentally measure the optical channel taking into account the specifications of light-emitting diode (LED) and photo-diode (PD). Measurement results show that the path loss exponent (PLE) is 2.47 and 1.78 when the angle of departure (AOD) is larger and smaller than the half-power angle (HPA), respectively. Moreover, changing the PD size from 9.7 mm to 2 mm results in a 15.05 dB difference in path loss (PL). We further derive a PL model that can capture the joint effects of LED and PD specifications based on empirical data. The root mean square error (RMSE) of the measured and simulated data is 2.52, thereby demonstrating the accuracy and practicality of the proposed model. Based on the model, we particularly investigate the impact of the device specifications on the bit error rate (BER) performance. Results indicate that the specifications of the device need to be carefully considered for developing VLC systems. Overall, the proposed model accurately characterizes PL under different device specifications and provides a feasible and universal device selection paradigm for the design and optimization of VLC systems.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Channel Measurements and Modeling in an Alpine Course Scenario at 4.9 and 26 GHz

Channel Measurements and Modeling in an Alpine Course Scenar...

引用

International Conference on Computer and Communications (ICCC)

作者： Jun Men Lei Tian Pan Tang Qidu Song Xinyu Zhao Baobao Liu Jianhua Zhang Liang Xia State Key Lab of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China China Mobile Research Institute Beijing China

The skiing sports scenario is considered to be one of the application scenarios of the fifth-generation (5G) mobile communication technology. Field channel measurement campaigns are conducted at 4.9 GHz and 26 GHz on an alpine course without snow covered. Based on the measurements, path loss and statistical parameters of delay are analyzed and compared between the two frequency bands. It is found that, in line-of-sight (LOS) case, the path loss exponents (PLEs) are close to those of free space whereas the mean values of the root-mean-square (RMS) delay spread (DS) are nearly the same for both frequency bands. In non-line-of-sight (NLOS) case where the LOS path is blocked by terrain, the PLEs are 9.42 and 13.19 for 4.9 GHz and 26 GHz, respectively. RMS DS is 86.22 ns at 4.9 GHz, which is twice larger than that at 26 GHz. In the case of the receiver is near-ground which emulates the situation of skier falling down, the PLEs are quite large which are 4.9 at 4.9 GHz and 5.9 at 26 GHz. The RMS DS are 14.53 ns and 2.99 ns at 4.9 and 26 GHz, which are extremely smaller than those of the LOS and NLOS case, especially in high frequency band.

关键词： Antenna measurements Microwave antennas Microwave measurement Snow Millimeter wave measurements Loss measurement Frequency measurement

来源：评论

学校读者我要写书评

暂无评论

Measurement-based Analysis of Atmospheric Attenuation by Considering Different Weather Types for Visible Light Communications

Measurement-based Analysis of Atmospheric Attenuation by Con...

引用

European Conference on Antennas and Propagation, EuCAP

作者： Baobao Liu Pan Tang Yue Yin Jianhua Zhang Liang Xia State Key Lab of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China China Mobile Research Institute Beijing China

Facing the current shortage of radio frequency spectrum resources, visible light communications (VLC) technology has been proposed to be a candidate technology in the sixth generation (6G), especially in the application of Intelligent Transportation Systems (ITS). And the propagation characteristics of VLC channel are critical for VLC system design and optimization. Furthermore, the impact of weather conditions on VLC channel cannot be ignored considering the vehicular VLC in ITS. In this paper, we experimentally investigate the impact of different weather types including fog and relative humidity (RH) on VLC in a laboratory atmospheric chamber. To simulate adverse weather conditions, we change the levels of fog density and water vapor introduced into the chamber, respectively. Then we measure the power attenuation and the bit error rate (BER) performance of VLC system under fog and RH conditions. We find that the VLC system will exhibit worse BER performance in the fog impaired channel than RH conditions. Besides, the results show that the bit errors will not happen until the attenuation reaches the threshold, i.e. attenuated by about 60% for both fog and RH conditions in our experiments. After that, the BER will deteriorate badly even up to 20%.

关键词： Power measurement Atmospheric measurements Bit error rate Measurement uncertainty Attenuation measurement Attenuation System analysis and design

来源：评论

学校读者我要写书评

暂无评论

Empirical Analysis of Sensing Channel Characteristics and Environment Effects at 28 GHz

Empirical Analysis of Sensing Channel Characteristics and En...

引用

IEEE Globecom Workshops

作者： Jialin Wang Jianhua Zhang Yuxiang Zhang Tao Jiang Li Yu Guangyi Liu State Key Lab of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China China Mobile Research Institution Beijing China

ISBN: (纸本)9781665459761

Joint communication and sensing (JCAS), in which the sensing and communication share the same frequency band, resources and hardware, is envisioned as a promising technology beyond 5G and 6G communication. For the JCAS system design and performance evaluation, accurate and efficient channel models are crucial. Since the sense is realized through information interaction in the environment, JCAS channel modeling needs to consider the effect of the environment based on traditional methods. However, most of the existing studies on the JCAS channel regard the environment as a clean background and ignore the effect of the environment on the channel. To research this issue, we analyze the channel characteristics based on the collected data by conducting two sets of measurements at 28 GHz in an outdoor scenario. We define the power ratio (PR) to quantity the variation of the environment around the sensed target. Further, we introduce a metal board to study the effect of the interference environment on the sensing channel. Results show that whether the target can be sensed depends on the ratio of the target power in the sum of environment multipath powers. Based on this discovery, we give the conditions that the target can be sensed. Finally, we propose an evaluation method of sensing recognition based on the variation of the sensed target power.

关键词： Performance evaluation Correlation coefficient Power measurement Target recognition Conferences Metals Interference

来源：评论

学校读者我要写书评

暂无评论

Coordination of networking and computing: toward new information infrastructure and new services mode

引用

Frontiers of Information technology & Electronic Engineering 2024年第5期25卷 629-632页

作者： Xiaoyun WANG Tao SUN Yong CUI Rajkumar BUYYA Deke GUO Qun HUANG Hassnaa MOUSTAFA Chen TIAN Shangguang WANG China Mobile Communications Corporation Beijing 100032China China Mobile Research Institute Beijing 100053China Department of Computer Science and Technology Tsinghua UniversityBeijing 100084China CLOUDS Lab School of Computing and Information SystemsUniversity of MelbourneMelbourneVIC 3010Australia National Key Laboratory of Information Systems Engineering National University of Defense TechnologyChangsha 410073China School of Computer Science Peking UniversityBeijing 100871China Intel Corporation Santa ClaraCA 95054USA State Key Laboratory for Novel Sofiware Technology Nanjing UniversityNanjing 210023China State Key Laboratory of Networking and Switching Technology Bejing University of Posts and TelecommunicationsBeijing 100876China

Computing,serving as the cornerstone of information processing,plays a pivotal role in the digital service ***"network"and"computing,"responsible for information transmission and processing respectively,traditionally belong to different stakeholders and have evolved ***,the recent trend toward the coordination and integration of computing and networks has garnered significant attention from both industry and *** such as"computility network"and"computing force network"have emerged,and the International Telecommunication Union-Telecommunication Standardization(ITU-T)has initiated efforts to develop standards for the coordination of networking and computing(CNC),focusing on the architecture and framework.

关键词： computing services network

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：