检索结果-内蒙古大学图书馆

International Conference on Computer and Communications (ICCC)

作者： Han Qiu Huihu Zhu Yufeng Li Ziyi Zeng National Digital Switching System Engineering and Technological R&D Center Zhengzhou China

ISBN: (纸本)9781538683408

Cascading failures are a severe threat to inter-domain routing system, while little has been done to detect and report impacts. To address the problem, we analyze the process of cascading failure in the inter-domain routing system, and find that it is constructed by two stages as key nodes/links failure and related nodes/links cascading failure. The first stage is shorter as the onset time is seconds' level, while the second stage is longer as the onset time is at least one hour. Then a two-phase detection method Fd-SP mechanism is put forward, the first stage of which adopts real-time detection methods to identify the initial failure nodes and links and the second stage uses dF 2 -CFM model to predict the damage range of cascade failure before true propagation. To validate Fd-SP, it is applied to security events containing cascading failures and the results of Fd-SP are compared with the datasets collected from rIPE during security events. Using the number of Update messages received by nodes as the metric, the maximum relative error is 1.69 and the average one is 0.92. As our prediction is a little large than reality, the false negative rate of Fd-SP is low.

关键词： Power system faults Power system protection routing Peer-to-peer computing Internet Kinetic theory Predictive models

来源：评论

学校读者我要写书评

暂无评论

引用

International Conference on Computational Intelligence and Applications (ICCIA)

作者： Xiaojuan Wang Ying Zeng Jun Shu Chi Zhang Bin Yan National Digital Switching System Engineering & Technological R&D Center Zhengzhou China

ISBN: (纸本)9781538695722;9781538695715

Eye movements obtained during eye tracking include both the fixation location and cognitive activities of the subject when viewing the remote sensing images. Aiming at the problem that the existing region of interest analysis methods based on the eye movement do not determine whether the area contains targets or not, in this paper a method was presented to predict the targets in the regions of interest using cognitive activities from eye movement. The method realized the detection and location of targets in remote-sensing images through the combination of fixation location and cognitive activity in eye movements. Experiments showed that the cognitive activities related to fixations make significant contribution to target prediction in the remote sensing image with complex backgrounds.

关键词： Feature extraction remote sensing Object detection Visualization Semantics Gaze tracking Image color analysis

来源：评论

学校读者我要写书评

暂无评论

Shortwave-Position-resource Scheduling Method Based on Signal Quality Estimation via Neural Network and dynamic Weight Adjustment 2017

Shortwave-Position-Resource Scheduling Method Based on Signa...

引用

Proceedings of the International Conference on Imaging, Signal Processing and Communication

作者： Lijia Zhang Jing Zhang Xiaoyi Zhang National Digital Switching System Engineering & Technique R&D Center

ISBN: (纸本)9781450352895

The manual expert scheduling, currently utilized in shortwave position system, is unable to satisfy the modern scenario of intensive tasks and deficient resources, thus resulting in the low position accuracy. To address the problem, we propose a shortwave-position-resource scheduling method based on Back Propagating Neural Network(BPNN) and dynamic Weight Adjustment(dWA). Firstly, we conceive a basic centralized optimization model for problem formulation, constructing an objective function consisted by Geometrical dilution of Precision(GdOP) and signal quality via linear weight summation. Furthermore, we study the dynamic adjustment strategy of weights. With weights updated according to the gradients of factors, the objective function inclines to the factor of better performance during iterations, thus leading to a better optimization searching ability. Moreover, by introducing the Bayesian regulation(Br) and Levenberg-Marquardt(LM) algorithm, we make an improvement on the generalization performance of BPNN, then achieving the signal quality estimation in the shortwave channel. Simulations indicate that the proposed method obtains an excellent robustness in scenes of different tasks number. When compared with other scheduling methods, the proposed method can effectively enhance the location accuracy. Experiments also show that the timeliness of the method is relatively low and needs further improvement.

关键词： Scheduling Shortwave position Neural network dynamic weight

来源：评论

学校读者我要写书评

暂无评论

Multi-label Extreme Learning Machine Based on Label Matrix Factorization 2

Multi-label Extreme Learning Machine Based on Label Matrix F...

引用

2017 IEEE 2nd International Conference on Big data Analysis（ICBdA 2017）

作者： Li Sihao Chen Fucai Huang ruiyang Xie Yixi National Digital Switching System Engineering & Technological R&D Center

ISBN: (纸本)9781509036202

Multi-label learning aims to predict the label sets an instance belongs *** Learning Machine(ELM),as a single-hidden layer feedforward neural network algorithm,has been extended to multi-label scenarios because of faster learning speed and less human *** at dealing with the problem of ignoring the inter-label dependencies,the proposed method,ELM-LMF,decomposes the label matrix into latent label matrix mapping the label space to the latent space and the k-label dependency matrix preserving the interlabel relationships,classifies in latent space with ELM,and maps the predicted labels back to the original *** on 3 benchmark multi-label data sets with 3 other state-of-art methods prove the feasibility of ELM-LMF.

关键词： extreme learning machine multi-label learning label matrix factorization classification machine learning

来源：评论

学校读者我要写书评

暂无评论

Collaborative Filtering Algorithm Based on denoising Auto-Encoder and Item Embedding 2

Collaborative Filtering Algorithm Based on Denoising Auto-En...

引用

2017 IEEE 第2届先进信息技术,电子与自动化控制国际会议(IAEAC 2017)

作者： Yudong Guo Yongwang Tang National Digital Switching System Engineering and Technological R&D Center

An collaborative filtering algorithm based on denoising Auto-Encoder and item embedding(CdAWE) was proposed to solve the absent analysis of item co-occurrence relation and the cold start of model para

ISBN: (纸本)9781467389808

关键词： Training Matrix decomposition Prediction algorithms Noise reduction Context modeling Sparse matrices Load modeling

来源：评论

学校读者我要写书评

暂无评论

An Approach for detecting Spammer Invasion Events with Poisson Process

An Approach for Detecting Spammer Invasion Events with Poiss...

引用

International Conference on Mechanical, Control and Computer engineering (ICMCCE)

作者： Qiang Qu Hongtao Yu ruiyang Huang National Digital Switching System Engineering & Technological R&D Center Zhengzhou China

The following topics are dealt with: feature extraction; control system synthesis; learning (artificial intelligence); production engineering computing; three-term control; design engineering; optimisation; pattern cl... 详细信息

关键词： Handheld computers Security Event detection Process control

来源：评论

学校读者我要写书评

暂无评论

Hardware Acceleration resource Allocation Mechanism for VNF

引用

Procedia Computer Science 2018年 131卷 746-755页

作者： Hongwei Fan Yuxiang Hu Shaojun Zhang Quan ren National Digital Switching System Engineering & Technological R&D Center Kim Street No. 7 Zhengzhou 450000 China

Improving the performance for VNF (Virtualized Network Function) by hardware acceleration is a hot research topic in SdN/NFV architecture. After introducing the acceleration resource for VNF, how to implement uniform deployment and allocation of the acceleration resources is an urgent problem to be solved. Thus in this paper, a unified deployment architecture of acceleration resource for VNF is proposed firstly. On the basis of architecture, the problem of acceleration resource allocation is modeled. By analyzing the influence of acceleration resource for VNF on deployment of service chains, we propose the evaluation index of acceleration resource allocation algorithm for VNF. Finally, we design a two-stage acceleration resource allocation algorithm to solve the problem. The experimental results show that the proposed algorithm can rationally allocate acceleration resources to nodes in the entire network and improve the efficiency of acceleration resource utilization.

关键词： SdN/NFV Virtualized Network Function Hardware Acceleration resource Allocation

来源：评论

学校读者我要写书评

暂无评论

Knowledge distillation for Machine Translation

Knowledge Distillation for Machine Translation

引用

2018 2nd International Conference on Computer Science and Intelligent Communication(CSIC 2018)

作者： Zhen Li dan Qu Chaojie Xie Xuejuan Wei National Digital Switching System Engineering & Technological R&D Center P.R.China

Encoder-to-decoder is a newly architecture for Neural Machine Translation(NMT). Convolutional Neural Network(CNN) based on this framework has gained significant success in NMT task. Challenges remain in the practical use of CNN model, which is in need of bilingual sentence pairs for training and each bilingual data is designed for CNN translation model needing retraining. Although some successful performance has been reported, it is an important research direction to avoid model overfitting caused by the scarcity of parallel corpus. The paper introduces a simple and efficient knowledge distillation method for regularization to solve CNN training overfitting problems by transferring the knowledge of source model to adapted model on low-resource languages in NMT task. The experiment on English-Czech dataset result shows that our model solve the over fitting problem, get better generalization, and improve the performance of a low-resource languages translation task.

关键词： Neural Machine Translation (NMT) Convolutional Neural Network (CNN) knowledge distillation encoder-decoder low-resource

来源：评论

学校读者我要写书评

暂无评论

An Equivalent Scheduling Strategy for Cyberspace Mimicry defense Based on Security Priority

An Equivalent Scheduling Strategy for Cyberspace Mimicry Def...

引用

International Conference on Advanced Mechatronic systems (ICAMechS)

作者： Ke Song SHuai Wei WenJian Zhang LiBo Tan NDSC National Digital Switching System Engineering & Technological R&D Center Zhengzhou China

With the continuous development and popularization of network equipment and services, people's dependence on cyberspace is becoming stronger and stronger, and the importance of network security is becoming more and more prominent. Mimic defense is based on the system architecture of dynamic heterogeneous redundancy. data is sent to multiple heterogeneous functional equivalents through the input agents, and then the response of heterogeneous functional equivalents is judged according to a certain arbitration policy by the output agent. The abnormal output of a functional equivalent can be effectively eliminated, enabling the system to achieve intrusion prevention against unknown system vulnerabilities or backdoors. This paper analyzes the effect of the isomorphism of functional equivalents on system security, and proposes a mimic method based on the equivalence of mimicry maximization. Experiments have shown that the effectiveness of security defense can be significantly improved.

关键词： Security Computer architecture Cyberspace redundancy Software dynamic scheduling Optimal scheduling

来源：评论

学校读者我要写书评

暂无评论

N-fold Superposition: Improving Neural Networks by reducing the Noise in Feature Maps

arXiv

引用

arXiv 2018年

作者： Liu, Yang Qu, Qiang Gao, Chao National Digital Switching System Engineering & Technological R&D Center Zhengzhou450002 China

Considering the use of Fully Connected (FC) layer limits the performance of Convolutional Neural Networks (CNNs), this paper develops a method to improve the coupling between the convolution layer and the FC layer by reducing the noise in Feature Maps (FMs). Our approach is divided into three steps. Firstly, we separate all the FMs into n blocks equally. Then, the weighted summation of FMs at the same position in all blocks constitutes a new block of FMs. Finally, we replicate this new block into n copies and concatenate them as the input to the FC layer. This sharing of FMs could reduce the noise in them apparently and avert the impact by a particular FM on the specific part weight of hidden layers, hence preventing the network from overfitting to some extent. Using the Fermat Lemma, we prove that this method could make the global minima value range of the loss function wider, by which makes it easier for neural networks to converge and accelerates the convergence process. This method does not significantly increase the amounts of network parameters (only a few more coefficients added), and the experiments demonstrate that this method could increase the convergence speed and improve the classification performance of neural networks. Copyright © 2018, The Authors. All rights reserved.

关键词： Image classification

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：