检索结果-内蒙古大学图书馆

International Conference on Applied Machine Learning (ICAML)

作者： Yue Zhang Lihong Cao State Key Laboratory of Media Convergence and Communication Communication University of China Beijing China Neuroscience and Intelligent Media Institute Communication University of China Beijing China State Key Laboratory of Mathematical Engineering and Advanced Computing Wuxi China

ISBN: (纸本)9781665421263

Sign languages are the primary medium of communication for the deaf and hearing-impaired community, and the number of human beings with hearing impairment is increasing at a rapid rate. There is a significant communication barrier between the hearing-impaired and normal-hearing communities. One solution to this issue is to offload the effort of text understanding for the deaf and hearing-impaired community in the form of Sign Language Generation (SLG). In this paper, first of all, we systematically review the mainstream Neural Machine Translation (NMT)-based SLG approaches. Specifically, according to their model architectures, we divide them into two categories: the Recurrent Neural Networks (RNNs)-based and Transformer-based models, and discuss their advantages and limitations. Next, we introduce the existing publicly available datasets and the data preprossessing methods in the field of SLG. Then, we summarize the performance evaluation methods for NMT-based SLG models. Finally, we offer two possible future research directions for better development in the field of SLG.

关键词： Performance evaluation Recurrent neural networks Pipelines Data preprocessing Gesture recognition Machine learning Auditory system

来源：评论

学校读者我要写书评

暂无评论

Unknown Traffic Identification Based on Deep Adaptation Networks

Unknown Traffic Identification Based on Deep Adaptation Netw...

引用

IEEE Conference on ALocal Computer Networks Workshops (LCN Workshops)

作者： Ziji Yang Wei Lin State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou China

ISBN: (数字)9781728183145

ISBN: (纸本)9781728183152

Network traffic classification has become an important basis for computer networks. However, the emergence of new applications, which generate unknown traffic constantly, has brought new challenges. The most critical challenge is how to divide the mixed unknown traffic into clusters containing only one category. In this paper, we propose a transfer learning approach using Deep Adaptation Network (DAN). This approach utilizes a few labeled samples from known traffic to improve the clustering purity of unknown traffic. We first trained a Convolutional Neural Network (CNN) model on unlabeled dataset with sampled time-series features. Then the model was extended to an adaptation model, co-trained on labeled and unlabeled samples. We evaluated our model using two publicly available datasets, achieving a purity of 98.23%. Our results demonstrate the effectiveness of DAN model in unknown traffic clustering. Moreover, we studied three sampling techniques and five clustering algorithms in our model for better clustering performance.

关键词： Adaptation models Protocols Transfer learning Telecommunication traffic Labeling Convolutional neural networks Knowledge transfer

来源：评论

学校读者我要写书评

暂无评论

Topology-Aware Software Assignment in Networked Systems 2

Topology-Aware Software Assignment in Networked Systems

引用

2019 2nd International Conference on advanced Algorithms and Control engineering, ICAACE 2019

作者： Wang, Junchao Liu, Hao Liu, Zhenwu Zhang, Zheng Pang, Jianmin State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou China

Cyberspace mimic defense has been proven to be a revolutionary defense technology that "changes the rules of the game" to ensure the security of cyberspace. However, the software diversity inherent in mimic defense technology may increase the difficulty in managing software executable binaries, especially when updating or debugging a software. In this paper, we study the problem of software assignment in a networked system to minimise the number of binaries generated by mimic compilers. Mimic compilers can help to generate functional equivelant software executables which exhibits diverse characteristics such as binary size etc. Theoretically the software assignment problem is equal to the traditional graph coloring problem. To guarantee network severity, we apply a Welsh-Powell-based Software Assignment (WPSA) algorithm to determine the number of binaries needed and the assignment of these binaries to hosts in a network. We conduct experiments on real world network topologies. Experimental results show that our algorithm can effectively reduce the number of binaries needed in networked systems. © Published under licence by IOP Publishing Ltd.

关键词： Network topology

来源：评论

学校读者我要写书评

暂无评论

OneFlow: Redesign the Distributed Deep Learning Framework from Scratch

arXiv

引用

arXiv 2021年

作者： Yuan, Jinhui Li, Xinqi Cheng, Cheng Liu, Juncheng Guo, Ran Cai, Shenghang Yao, Chi Yang, Fei Yi, Xiaodong Wu, Chuan Zhang, Haoran Zhao, Jie OneFlow Research Sweden Zhejiang Laboratory China The University of Hong Kong Hong Kong University of Pennsylvania United States State Key Laboratory of Mathematical Engineering and Advanced Computing China

Deep learning frameworks such as TensorFlow and PyTorch provide a productive interface for expressing and training a deep neural network (DNN) model on a single device or using data parallelism. Still, they may not be flexible or efficient enough in training emerging large models on distributed devices, which require more sophisticated parallelism beyond data parallelism. Plugins or wrappers have been developed to strengthen these frameworks for model or pipeline parallelism, but they complicate the usage and implementation of distributed deep learning. Aiming at a simple, neat redesign of distributed deep learning frameworks for various parallelism paradigms, we present OneFlow, a novel distributed training framework based on an SBP (split, broadcast and partial-value) abstraction and the actor model. SBP enables much easier programming of data parallelism and model parallelism than existing frameworks, and the actor model provides a succinct runtime mechanism to manage the complex dependencies imposed by resource constraints, data movement and computation in distributed deep learning. We demonstrate the general applicability and efficiency of OneFlow for training various large DNN models with case studies and extensive experiments. The results show that OneFlow outperforms many well-known customized libraries built on top of the state-of-the-art frameworks. The code of OneFlow is available at: https://***/Oneflow-Inc/oneflow. © 2021, CC BY.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Clustering and Routing Optimization Algorithm for Heterogeneous Wireless Sensor Networks

Clustering and Routing Optimization Algorithm for Heterogene...

引用

International Wireless Communications and Mobile computing Conference, IWCMC

作者： Ling Chen Wenwen Liu Daofu Gong Yan Chen State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou China

ISBN: (数字)9781728131290

ISBN: (纸本)9781728131306

Wireless sensor network (WSN) is the cutting-edge technology of modern monitoring technology. There is a direct relationship between the length of its lifetime and its performance. Efficient and reasonable clustering and routing algorithm can reduce the energy loss in the network, which is of great significance for prolonging the network lifetime. The traditional WSN protocol has been difficult to meet the needs of the contemporary society. This paper focuses on the advantages and disadvantages of the traditional WSN protocol, and puts forward relevant improvement measures to optimize the traditional WSN protocol and extend the network life time.

关键词： Wireless sensor networks Clustering algorithms Routing Energy consumption Head Wireless communication Clustering methods

来源：评论

学校读者我要写书评

暂无评论

Binary software vulnerability detection method based on attention mechanism

Binary software vulnerability detection method based on atte...

引用

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

作者： Wenjie Han Jianmin Pang Xin Zhou Di Zhu State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou China

Aiming at the stack overflow vulnerability in binary software, this paper proposes a binary vulnerability detection method based on the attention mechanism. First, this paper analyze the basic characteristics of stack overflow vulnerabilities, and perform data preprocessing on the decompiled files to make the neural network better adapt to the characteristics of stack overflow vulnerabilities, then formulate instruction specifications at the assembly language level, and finally input the data into the fusion attention mechanism Learning in the neural network. This paper compares and analyzes three kinds of neural networks on the CWE121 data set. The experimental results show that after neural network training, the detection method based on the attention mechanism can be effective and accurately discover whether the target area has stack overflow vulnerabilities, thereby greatly improving the detection efficiency.

关键词： Training Computational modeling Neural networks Force Data preprocessing Software Data models

来源：评论

学校读者我要写书评

暂无评论

A Framework for Multi-Variant Execution Environment

A Framework for Multi-Variant Execution Environment

引用

2019 International Conference on Artificial Intelligence Technologies and Applications, ICAITA 2019

作者： Junchao, Wang Zhenwu, Liu Hao, Liu Jianmin, Pang State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou China

Software diversity has been proven to be an effective approach to enhance system security. To make the best of the advantage brought by software diversity, a multi-variant execution environment is needed. However, although some MVEEs have been proposed, most of them are either too simple or only focusing one aspect which limits their widely adoption in industry. In this paper we propose a framework for multi-variant execution environment to enhance the security of software systems. The framework addresses different aspects when implementing a MVEE and can help to make the best of software diversity to enhance the system's security. © 2019 IOP Publishing Ltd. All rights reserved.

关键词： Artificial intelligence

来源：评论

学校读者我要写书评

暂无评论

Median Filtering Detection Based on Quaternion Convolutional Neural Network

引用

Computers, Materials & Continua 2020年第10期65卷 929-943页

作者： Jinwei Wang Qiye Ni Yang Zhang Xiangyang Luo Yunqing Shi Jiangtao Zhai Sunil Kr Jha Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology Nanjing University of Information Science&TechnologyNanjing210044China State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou450001China Department of Computer and Software Nanjing University of Information Science&TechnologyNanjing210044China State Key Laboratory of Information Security Institute of Information EngineeringChinese Academy of SciencesBeijing100093China Department of Electrical Computer Engineering New Jersey Institute of TechnologyNew JerseyNJ07102USA

Median filtering is a nonlinear signal processing technique and has an advantage in the field of image ***,more attention has been paid to the forensics research of median *** this paper,a median filtering forensics method based on quaternion convolutional neural network(QCNN)is *** median filtering residuals(MFR)are used to preprocess the *** the output of MFR is expanded to four channels and used as the input of *** QCNN,quaternion convolution is designed that can better mix the information of different channels than traditional *** quaternion pooling layer is designed to evaluate the result of quaternion *** is proposed to features well combine the three-channel information of color image and fully extract forensics *** show that the proposed method has higher accuracy and shorter training time than the traditional convolutional neural network with the same convolution depth.

关键词： Median filtering forensics quaternion convolution layer quaternion pooling layer color image

来源：评论

学校读者我要写书评

暂无评论

DSA-NIE: Double Static Attributes Based Node Importance Evaluation of Inter-Domain Routing System 5

DSA-NIE: Double Static Attributes Based Node Importance Eval...

引用

5th IEEE International Conference on Computer and Communications, ICCC 2019

作者： Zhu, Huihu Qiu, Han Zhu, Junhu Wang, Qingxian State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou China

ISBN: (纸本)9781728147437

Assessing the importance of AS nodes is important for the protection of inter-domain routing system security. However, the existing literature on node importance evaluation methods based on static attributes are difficult to accurately characterize nodes. And which based on dynamic attributes have high time complexity. To address these problems, we analyze the relationship between static structural attributes and dynamic attributes after node failure. It is found that the degree can indirectly reflect the impact of the UPDATE packet propagation and the number of routing paths can reflect the impact caused by load redistribution. And then, we build static attribute characterization model of node failure effect. After that an evaluation method DSA-NIE is put forward. Given the two static attributes, DSA-NIE could effectively assessment of the importance of nodes. When compared with SD-KNI algorithm, the time complexity has been reduced to O(lmn). And the experiment result shows that the failure effect of the key nodes assessed by the DSA-NIE is 7.4% higher than the average, and up to 12.9% higher. © 2019 IEEE.

关键词： evaluation Inter-domain Routing System static attribute

来源：评论

学校读者我要写书评

暂无评论

Computer evidence analysis technology based on weighted frequent pattern growth algorithm 5th

Computer evidence analysis technology based on weighted freq...

引用

5th International Conference on Artificial Intelligence and Security, ICAIS 2019

作者： Xue, Tianxiao Li, Qingbao Zhang, Ping Chen, Zhifeng Feng, Peijun Luo, Nan State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou China

ISBN: (纸本)9783030242701

The current analysis of computer forensics is still dependent on the investigation personnel, leading to the problem of heavy workload and low efficiency. At the same time, computer evidences have the characteristics of complex structure and large amount of data, which is prone to the problem of association rules redundancy. In order to improve the efficiency of computer evidence analysis, this paper uses association analysis technology to analyze the user's behavior habits from the obtained computer system log, browser records, file operation traces and other information;combined with the idea of weighted features, the weight set was generated according to the difference of attribute importance, and the Frequent Pattern Growth (FP-Growth) algorithm is improved. The experimental results show that the improved FP-Growth algorithm can effectively remove the redundant rules in evidence analysis, and is more practical for user behavior analysis. © Springer Nature Switzerland AG 2019.

关键词： Efficiency

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：