检索结果-内蒙古大学图书馆

8th International Conference on Data Science in Cyberspace, DSC 2023

作者： Shen, Yueyue Song, Feifan Wu, Jingchen Wu, Xu School of Cyberspace Security Beijing University of Posts and Telecommunications Key Lab of Trustworthy Distributed Computing and Service Ministry of Education Beijing China School of Economics and Finance Changchun Finance College Key Lab of Trustworthy Distributed Computing and Service Ministry of Education Jilin China School of Computer Science Beijing University of Posts and Telecommunications Key Lab of Trustworthy Distributed Computing and Service Ministry of Education Beijing China

ISBN: (纸本)9798350331035

Although the general event extraction model has achieved good results in the standard data sets, there is a gap in the extraction results in low-resource environments. The general event extraction model still has room for improvement in the performance of special fields. Therefore, this paper proposes a semi-supervised reading comprehension event extraction model adapted to transfer learning, based on improved UDA for data enhancement, called SSDAEE, which can increase the accuracy of reading comprehension event type recognition. SSDAEE has a good performance on the public opinion data set. It can not only enhance a small number of supervised data in low-resource fields with unsupervised data but also improve the generalization ability of the model. Experiments show that when only 5% unsupervised data is used, the F1 score of SSDAEE on the public opinion data set is increased by 9% compared with the fully supervised model, and by 2% compared with other semi-supervised models. © 2023 IEEE.

关键词： data augmentation event extraction semi-supervised transfer learning

来源：评论

学校读者我要写书评

暂无评论

A new routing and connectivity availability metric of low duty cycle random sleep scheduled multi-hop wireless sensor networks

引用

Journal of China Universities of Posts and Telecommunications 2011年第SUPPL.2期18卷 101-108页

作者： Zhang, Tian-Le Yuan, Yu-Yu Wu, Xu Luo, Zong-Wei Wang, Chun-Lu Key Laboratory of Trustworthy Distributed Computing and Service Ministry of Education Beijing University of Posts and Telecommunications Beijing 100876 China E-Business Technology Institute University of Hong Kong Hong Kong Hong Kong

Wireless sensor network (WSN) is attracting intensive interests in many application areas. In some scenarios such as monitoring of the coal-mining area or remote area, nodes seldom move and may switch from sleeping to waking mode to minimize power consumption. Energy conservation design for the network lifetime extension is the primary issue. By adopting sleep scheduling to keep low duty cycle can help reduce the energy consumption. Random sleep scheduling is a desirable mechanism for its simplicity and guaranteed stable low duty cycle. Moreover, the low duty cycle sleeping of nodes may destroy the connectivity of network. Most of existing WSN routing techniques are designed for fully connected networks assuming that the nodes of the routing path between source and destination are all available at the same time. Maintaining fully connected backbone involves high duty cycle and costly clock synchronization which consumes more energy. If certain delay is tolerant, partially connected routing can achieve successful packet forwarding in networks with the deficiency of connectivity. In this paper a new partially connected routing and simple random sleep scheduling scheme is proposed for WSN with low duty cycle. A continuous time Markov chains (CTMC) connectivity availability model is introduced to evaluate the end to end packet delivery ratio of multi-hop networks subjected to the impact of both frequent disconnection and random sleep latency. The proposed connectivity availability model can also be used as the key metric for route path selection to achieve stable and high performance. Deployment guidelines are given for WSN with random sleep scheduling policy satisfying reliability, timeliness and duty cycle requirements. By modeling and simulation, we find that as compared to the fully connected routing protocol, the partially connected routing design can achieve much higher packet delivery ratio and lower duty cycle within the tolerant delay. The computation and simul

关键词： Energy utilization

来源：评论

学校读者我要写书评

暂无评论

An Unsupervised Graph Embedding Method Based on Dynamic Graph Attention Networks and Infomax for Link Prediction 8

An Unsupervised Graph Embedding Method Based on Dynamic Grap...

引用

8th International Conference on Data Science in Cyberspace, DSC 2023

作者： Lu, Jingzhu Song, Feifan Wu, Jingchen Wu, Xu School of Cyberspace Security Beijing University of Posts and Telecommunications Key Lab of Trustworthy Distributed Computing and Service Ministry of Education Beijing China School of Economics and Finance Changchun Finance College Key Lab of Trustworthy Distributed Computing and Service Ministry of Education Jilin China School of Computer Science Beijing University of Posts and Telecommunications Key Lab of Trustworthy Distributed Computing and Service Ministry of Education Beijing China

ISBN: (纸本)9798350331035

Due to the lack of labeled data in many real-world scenarios, Graph Neural Network models that require a large volume of labeled data are not effective, and research based on unsupervised learning frameworks has become essential. To solve the problem of sparsity in labeled graph data, this paper proposes an unsupervised graph representation learning method based on graph infomax and dynamic graph attention mechanism(IMDGA). This method conducts unsupervised learning based on the idea of contrastive learning and optimizes the encoder by calculating the maximum mutual information between the graph's local information and global information. At the same time, we use a dynamic Graph Attention Network as an encoder model to prevent the Graph Convolutional Neural network's inability to adapt to neighbor nodes with different weights. Finally, link prediction experiments on three public data sets of cora, citeseer, and pumbed show that the proposed model has achieved superior results compared with other graph representation learning, showing its effectiveness and competitiveness. © 2023 IEEE.

关键词： dynamic graph attention networks graph representation learning maximum mutual information unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

Word activation forces-based language modeling and smoothing

Word activation forces-based language modeling and smoothing

引用

2013 5th International Conference on Intelligent Human-Machine Systems and Cybernetics, IHMSC 2013

作者： Qin, Min Liu, Gang Li, Baoxiang Lu, Yueming School of Information and Communication Engineering BUPT Beijing China Key Laboratory of Trustworthy Distributed Computing and Service Ministry of Education BUPT China

ISBN: (纸本)9780769550114

N-gram language models are useful for modeling the local dependencies of word occurrences but not for capturing global word dependencies. When the window size n is limited, the n-gram is weak in terms of capturing long distance dependencies. Long-distance Dependency information has long been proven useful in language model. However, the improved performance of long-distance LMs over conventional n-gram models generally comes at the cost of increased decoding complexity and model size. Word Activation Forces has been proven a simple and human-comparable accurate measure to identify word closest associates. In this paper, Word Activation Forces-Based language model is proposed to capture the long distance dependency between words, but which is as fast for decoding as a conventional word n-gram. As shown by experiments on broadcast news, the proposed language modeling and smoothing can significantly reduce the perplexity of language models and word error rate with moderate computational cost. © 2013 IEEE.

关键词： Decoding

来源：评论

学校读者我要写书评

暂无评论

Noisy chaotic time series prediction based on wavelet echo state network

International Journal of Digital Content Technology and its ...

引用

International Journal of Digital Content Technology and its Applications 2012年第9期6卷 173-181页

作者： Yang, Fei Fang, Binxing Wang, Chunlu Zuo, Xingquan Zhong, Ruiming School of Computer Science Key Laboratory of Trustworthy Distributed Computing Beijing University of Post and Telecommunications Service of the Ministry of Education of China Beijing 100876 China Nanjing Panda Electronics Company limited Nanjing 210002 China

As a research focus of intelligence algorithm, the prediction of classic noiseless chaotic time series has a great development in recent years. However, the existing prediction models cannot get good performance for real-world chaotic time series because of the interference of noise components. In order to take full advantage of the property of real-world chaotic time series, the paper proposes a novel prediction model based on wavelet transform and echo state network (WESN). The basic idea of WESN is that firstly wavelet decomposition is used to separate the chaotic dynamics component and noise components, then the gotten components can be predicted by echo state network (ESN) independently, and finally the prediction results of time series are obtained by assembling the prediction values of all components. By using real-world sunspot time series for verification, the prediction results show that the proposed model has higher prediction accuracy by comparing with the models of direct echo state network (DESN), and direct echo state network (DSVM).

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Short-term traffic flow prediction based on echo state networks

引用

Advances in Information Sciences and service Sciences 2012年第9期4卷 269-277页

作者： Yang, Fei Wang, Chunlu Zuo, Xingquan Zhong, Ruiming Xiang, Fei School of Computer Science Beijing University of Post and Telecommunications Key Lab of Trustworthy Distributed Computing and Service of the Ministry of Education of China Services Sciences and Intelligent Transportation Research Center 10086 China

The traffic flow is critical for the traffic control and traffic guidance. As the short-term traffic flow has the characteristics of highly nonlinear, complexity and chaos, the traditional neural networks can not totally satisfy the performance in the aspects of prediction accuracy, training speed and stability. In order to solve those problems, a novel traffic flow prediction model based on echo state networks (ESN) is proposed. The model is established based on the chaotic property of traffic flow. Firstly, through phase space reconstruction of chaotic traffic flow data, the suitable input and output of ESN can be obtained. Then, by optimizing its key parameters of ESN, the ESN can make full use of the chaotic property of the input and output data to get more accurate prediction results. The propose model is applied to a real-world traffic flow prediction problem in the area of Xizhimen bridge of Beijing city, and compare with back-propagation neural networks (BPNN), Elman neural networks (Elman) and wavelet neural networks (WNN). The prediction results show that the model outperforms comparative other models in accuracy, training speed and stability.

关键词： Elman neural networks

来源：评论

学校读者我要写书评

暂无评论

A Non-Statistical Reinstatement Algorithm for Orientation Field of Incomplete Fingerprint

引用

电子学报(英文版) 2012年第3期21卷 477-480页

作者： JING Xiaojun ZHANG Bo LIU Xinjing WANG Dong School of Infornation and Communication Engineering Beijing University of Posts and Telecommunications Beijing 100876 China Key Laboratory of Trustworthy Distributed Computing and Service (BUPT) Ministry of Education Beijing University o.f Posts and Telecommunications Beijing 100876 China

Fingerprint orientation is the key informa- tion in fingerprint enhancement and matching. Accord- ing to the problem of incomplete fingerprint orientation, a novel algorithm of fingerprint orientation reinstatement is proposed based on Local binary pattern (LBP) and mutual information function. Firstly, mutual information of finger- print orientation is computed. Secondly, the feature vec- tor is defined for incomplete area classification by Support vector machines (SVM). Then, the fingerprint orientation field of the incomplete area is re-computed and measured by LBP and mutual information function. For orientation reinstatement, LBP is used to measure the similarity of incomplete and complete area, while the mutual informa- tion function is used to measure correlation and compe- tition of the neighborhood. Additionally, fuzzy criterion is proposed to assign different weights for neighborhood block and incomplete block for modifying orientation. As a result, the modified fingerprint orientation is used for post-processing. And the performance is shown in the ex- perimental and proves the efficiency and reliability of our algorithm.

关键词：指纹图像增强方向场算法互信息函数统计支持向量机 LBP 二值模式

来源：评论

学校读者我要写书评

暂无评论

Automatic detection of breast nodule in the ultrasound images using CNN

引用

The Journal of China Universities of Posts and Telecommunications 2019年第2期26卷 9-16页

作者： Pang Hao Bu Yunyun Wang Cong Xiao Hui School of Software Beijing University of Posts and TelecommunicationsBeijing 100876China Key Laboratory of Trustworthy Distributed Computing and Service (BUPT) Ministry of EducationBeijing 100876China Department of Ultrasound Beijing Friendship HospitalCapital Medical University China Electronic Data Service Company Limited

Breast cancer is the most common cancer among women *** is widely used as a harmless test for early breast cancer *** ultrasound network(USNet) model is *** is an improved object detection model specifically for breast nodule detection on ultrasound *** improved the backbone network,optimized the generation of feature maps,and adjusted the loss ***,USNet trained with real clinical *** evaluation results show that the trained model has strong nodule detection *** mean average precision(mAP) value can reach 0.734 *** nodule detection rate is 95.11%,and the in situ cancer detection rate is 79.65%.At the same time,detection speed can reach 27.3 frame per second(FPS),and the video data can be processed in real time.

关键词： convolutional neural network object detection ultrasound breast nodule

来源：评论

学校读者我要写书评

暂无评论

Exact analysis of transmit/receive zero-forcing based Mimo AF two-way relaying when relay has two antennas

Exact analysis of transmit/receive zero-forcing based Mimo A...

引用

International Conference on Cyberspace Technology, CCT 2013

作者： Li, Rongsheng Lv, Tiejun Lu, Yueming Key Laboratory of Trustworthy Distributed Computing and Service Ministry of Education School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing 100876 China

ISBN: (纸本)9781849198011

In this paper, the performance of joint transmit/receive zero-forcing (ZF) is analyzed for multiple-input multiple-out (IV11MO) amplify-and-forward (AT) two-way relay network when the relay has two antennas. The performance is studied by deriving exact expressions of the overall outage probability and the diversity order under the uncorrelated Rayleigh fading channel and relay-corr elated Rayleigh fading channel as well. Analysis and simulation results indicate that, compared to uncorrelated Rayleigh fading, spatial correlation at the relay deteriorates the performance of the system greatly, but does not affect the diversity order of the system.

关键词： Antennas

来源：评论

学校读者我要写书评

暂无评论

A decomposition based multiobjective evolutionary algorithm for dynamic overlapping community detection 13th

A decomposition based multiobjective evolutionary algorithm ...

引用

13th International Conference on Bio-Inspired computing: Theories and Applications, BIC-TA 2018

作者： Wan, Xing Zuo, Xingquan Song, Feng School of Computer Science Beijing University of Posts and Telecommunications Beijing China Key Laboratory of Trustworthy Distributed Computing and Service Ministry of Education Beijing China

ISBN: (纸本)9789811328282

Dynamic and overlapping are common features of community structures of many real world complex networks. There are few studies on detecting dynamic overlapping communities, but those studies consider only single optimization objective. In practice, it is necessary to evaluate the community detection by multiple metrics to reflect different aspects of a community structure. In this paper, we propose a multi-objective evolutionary algorithm based approach for the problem of dynamic overlapping community detection, with consideration of three optimization objectives: partition density (D), the extended modularity (EQ), and the community smoothing (NMILFK). The dynamic overlapping network is regarded as a set of network snapshots. The multi-objective evolutionary algorithm based on decomposition (MOEA/D) is used to detect the communities for each snapshot. Experiments show that our approach can find uniformly distributed Pareto solutions for the problem and outperforms those comparative approaches. © Springer Nature Singapore Pte Ltd. 2018.

关键词： Complex networks

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：