检索结果-内蒙古大学图书馆

23rd IEEE Hot Chips Symposium, HCS 2011

作者： Fan, Dongrui Zhang, Hao Wang, Da Ye, Xiaochun Song, Fenglong Zhang, Junchao Fan, Lingjun Advanced Micro-System Group National Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences China

This article consists of a collection of slides from the author's conference presentation on the special features, system design, processing capabilities, and targeted markets for AMS's Godson-T microprocessor...

ISBN: (纸本)9781467388771

关键词：

来源：评论

学校读者我要写书评

暂无评论

Protein structure prediction based on a domain clustering database

Protein structure prediction based on a domain clustering da...

引用

2007 1st International Conference on Bioinformatics and Biomedical Engineering, ICBBE

作者： Ma, Zhaoyun Zhang, Fa Xu, Lin Feng, Shengzhong Liu, Zhiyong Key Laboratory of Computer System and Architecture Chinese Academy of Sciences Institute of Computing Technology Graduate School Chinese Academy of Sciences

ISBN: (纸本)1424411203

Homology modeling, as a successful protein structure prediction method, has two major deficiencies, i.e., the lack of the templates (known structures), and the accuracy of alignment between the query (unknown structure) and its templates. To solve these problems, we have constructed a conservative domain clustering template database, and proposed a profile-based alignment algorithm based on the profile extracted from each domain clustering in the database. The extracted profile can well represent the information of the sequence and structure as well. Compared with other alignment methods, such as T-coffee and Smith-Water man, our results show that with this method it's possible to obtain a higher hit rate for template searching and a more accurate and reliable query-template alignment. Therefore, the quality of protein structure prediction can be improved. © 2007 Crown Copyright.

关键词： Proteins

来源：评论

学校读者我要写书评

暂无评论

Towards Stricter Black-box Integrity Verification of Deep Neural Network Models 24

Towards Stricter Black-box Integrity Verification of Deep Ne...

引用

32nd ACM International Conference on Multimedia, MM 2024

作者： He, Chaoxiang Bai, Xiaofan Ma, Xiaojing Zhu, Bin B. Hu, Pingyi Fu, Jiayun Jin, Hai Zhang, Dongmei Huazhong University of Science and Technology Hubei Wuhan China Shanghai Jiao Tong University Shanghai China Microsoft Corporation Beijing China National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Hubei Engineering Research Center on Big Data Security Hubei Key Laboratory of Distributed System Security School of Cyber Science and Engineering Huazhong University of Science and Technology China National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology China

ISBN: (纸本)9798400706868

Cloud-based machine learning services offer significant advantages but also introduce the risk of tampering with cloud-deployed deep neural network (DNN) models. Black-box integrity verification (BIV) allows model owners and end-users to determine if a cloud-deployed DNN model has been tampered with by examining only the top-1 label responses. Fingerprinting generates fingerprint samples to query the model, achieving BIV with no impact on the model's accuracy. In this paper, we present BIVBench, the first comprehensive benchmark for BIV of DNN models. BIVBench covers 16 types of model modifications, providing extensive coverage of practical modification scenarios. Our analysis reveals that existing fingerprinting methods, which are typically focused on significant tampering, lack the sensitivity needed to effectively detect subtle yet common and potentially severe modifications. To address this limitation, we propose MiSentry (Model Integrity Sentry), a novel fingerprinting method that leverages meta-learning. MiSentry strategically incorporates a few subtly modified models into the meta-learning model zoo and maximizes the divergence of output predictions between the target model and the modified models in the model zoo to generate highly sensitive, generalizable, and effective fingerprint samples. Extensive evaluations using BIVBench demonstrate that MiSentry outperforms existing state-of-the-art methods overall and significantly surpasses them in detecting subtle modifications. The BIVBench and supplementary materials are available at: https://***/CGCL-codes/BIVBench. © 2024 ACM.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

BTMD: A framework of binary translation based malcode detector

BTMD: A framework of binary translation based malcode detect...

引用

4th International Conference on Cyber-Enabled Distributed computing and Knowledge Discovery, CyberC 2012

作者： Shan, Zheng Guo, Haoran Pang, Jianmin Key Laboratory of Computer System and Architecture Institute of Computing Technology Chinese Academy of Sciences Beijing China National Digital Switching System Engineering and Technology Research Center Zhengzhou China

ISBN: (纸本)9780769548104

Binary Translation technology is used to convert binary code of one Instruction Set Architecture (ISA) into another. This technology can solve the software-inheritance problem and ISA-compatibility between different computers architecture. In this paper, we describe BTMD (Binary Translation based Malcode Detector), a novel framework that exploits static and dynamic binary translation features to detect broad spectrum malware and prevent its execution. By operating directly on binary code with MD Rules on the availability of source code, BTMD is appropriate for translating low-level binary code to high-level proper representation, obtaining CFG (Control Flow Graph) and other high-level Control Structure by MD Parser. Then Critical API Graph based on CFG is generated to do sub graph matching with the defined Malware Behavior Template. MD Engine in BTMD is called to undertake the process to take on the remaining code analysis. Compared with other detection approaches, BTMD is found to be very efficient in terms of detection capability and false alarm rate. © 2012 IEEE.

关键词： Binary codes

来源：评论

学校读者我要写书评

暂无评论

A clustering algorithm use SOM and K-Means in Intrusion Detection

A clustering algorithm use SOM and K-Means in Intrusion Dete...

引用

International Conference on E-Business and E-Government (ICEE)

作者： Wang, Huai-Bin Yang, Hong-Liang Xu, Zhi-Jian Yuan, Zheng Tianjin Key Lab. of Intelligent Computing and Novel Software Technology Tianjin University of Technology 300191 Tianjin China Key Laboratory of Computer Vision and System Ministry of Education 300191 Tianjin China Key Laboratory of Computer Vision and System Ministry of Education China

ISBN: (纸本)9780769539973

Improving detection definition is a pivotal problem for intrusion detection. Many intelligent algorithms were used to improve the detection rate and reduce the false rate. Traditional SOM cannot provide the precise clustering results to us, while traditional K-Means depends on the initial value serious and it is difficult to find the center of cluster easily. Therefore, in this paper we introduce a new algorithm, first, we use SOM gained roughly clusters and center of clusters, then, using K-Means refine the clustering in the SOM stage. At last of this paper we take KDD CUP-99 dataset to test the performance of the new algorithm. The new algorithm overcomes the defects of traditional algorithms effectively. Experimental results show that the new algorithm has a good stability of efficiency and clustering accuracy. © 2010 IEEE.

关键词： Self organizing maps

来源：评论

学校读者我要写书评

暂无评论

Coarray Tensor Train Decomposition for Bistatic MIMO Radar with Uniform Planar Array

引用

IEEE Transactions on Antennas and Propagation 2025年

作者： Xie, Qianpeng Wang, Zhanling Wen, Fangqing He, Jin Truong, Trieu-Kien National University of Defense Technology State Key Laboratory of Complex Electromagnetic Environment Effects on Electronics and Information System Changsha410073 China Hubei University of Automotive Technology Institute of Vehicle Information Control and Network Technology Shiyan442002 China China Three Gorges University Hubei Key Laboratory of Intelligent Vision Based Monitoring for Hydroelectric Engineering Yichang443002 China Shanghai Jiao Tong University Shanghai Key Laboratory of Navigation and Location Based Services School of Electronic Information and Electrical Engineering Shanghai200240 China

Recently, tensor network decomposition has attracted increasing attention due to its high efficacy in modeling multi-dimensional correlations in high-order tensors. Considering this advantage, this study employs tensor train decomposition (TTD) to achieve joint two-dimensional direction-of-departure (2D-DOD) and direction-of-arrival (2D-DOA) estimation in bistatic multiple-input multiple-output (MIMO) radar systems with uniform planar array (UPA) configurations. First, a five-dimensional (5D) tensor is defined to incorporate the two-dimensional (2D) spatial information from transmit and receive UPAs, which is then combined with temporal information. Next, a low-rank four-dimensional (4D) tensor is generated by performing an autocorrelation operation along the temporal dimension of the 5D tensor, forming a structure that encapsulates the difference coarrays of the transmit and receive UPAs. Then, a TTD framework is designed to decompose the 4D tensor into interconnected lower-order TT-cores. Furthermore, a precise bidirectional mapping is performed between the TT-cores and the original Vandermonde factor matrices. Finally, leveraging distinct combinations of TT-cores, two innovative methods are developed to recover the Vandermonde factor matrices. The proposed methods conduct automatic pairing of 2D-DOD and obtain 2D-DOA estimates by enforcing consistent permutation orders across all factor matrices. The proposed methods are verified by simulation experiments. The results confirm that the proposed methods can surpass the existing tensor-based methods in terms of estimation accuracy. © 1963-2012 IEEE.

关键词： Matrix algebra

来源：评论

学校读者我要写书评

暂无评论

License plate locating algorithm based on multi-feature in color image

License plate locating algorithm based on multi-feature in c...

引用

2010 International Conference on Educational and Information technology, ICEIT 2010

作者： Sun, Jing Xue, Yan-Bing Key Laboratory of Computer Vision and System Ministry of Education Tianjin University of Technology Tianjin 300191 China Tianjin Key Laboratory of Intelligence Computing and Novel Software Technology Tianjin University of Technology 300191 Tianjin China

ISBN: (纸本)9781424480340

A new method for license plate locating based on multi-feature is introduced in this paper. The method proposed can overcome the shortcoming of early methods based on single feature. At first, the algorithm employs HSV color space to get rid of non-license plate area. Then concomitant color feature, statistical letter frequency and geometry feature are used to detect the possible regions. The algorithm is neither confined to the size and position of the plate, nor limited to the complicated background or illumination condition of the image. Experiments show that the method is effective. © 2010 IEEE.

关键词： Color

来源：评论

学校读者我要写书评

暂无评论

A method of genetic algorithm optimized extended kalman particle filter for nonlinear system state estimation

A method of genetic algorithm optimized extended kalman part...

引用

5th International Conference on Natural Computation, ICNC 2009

作者： Yang, Shuying Huang, Wenjuan Ma, Qin Tianjin Key Laboratory of Intelligence Computing and Novel Software Technology Tianjin University of Technology 300191 Tianjin China Key Laboratory of Computer Vision and System Ministry of Education Tianjin University of Technology Tianjin 300191 China

ISBN: (纸本)9780769537368

A new method of genetic algorithm (GA) optimized the extended kalman particle filter (EKPF) is proposed in this paper. The algorithm of extended kalman particle filter is a suboptimal filtering algorithm with good performance for target tracking and non-linear tracking problem. In the implementation of the extended kalman particle filter, a resampling scheme is used to decrease the degeneracy phenomenon and improve estimation performance. However, the target tracking mutation system status has poorer filtering precision. In order to overcome the problem of the extended kalman particle filter, a novel filtering method called the genetic particle filter (GA-EKPF) is proposed in this paper. The genetic mechanism provides an important guiding ideology to solve the deprivation of particles. The proposed algorithm overcomes the deprivation of particles and enhances the filtering precision. Experimental results show that the performance of modified extended kalman particle filter superiors to the standard particle filter (PF) and some other modified PFs. © 2009 IEEE.

关键词： Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

Particle swarm optimized unscented particle filter for target tracking

Particle swarm optimized unscented particle filter for targe...

引用

2009 2nd International Congress on Image and Signal Processing, CISP'09

作者： Shuying, Yang Ma, Qin Wenjuan, Huang Tianjin Key Laboratory of Intelligence Computing and Novel Software Technology Tianjin University of Technology 300191 Tianjin China Key Laboratory of Computer Vision and System Ministry of Education Tianjin University of Technology Tianjin 300191 China

ISBN: (纸本)9781424441310

In this paper, a novel particle swarm optimized (PSO) unscented particle filter (PSO-UPF) algorithm is proposed for target tracking. Unscented particle filter (UPF) can obtain the better sequential importance sampling than the traditional PF algorithm. Then we use PSO to optimize the state equation of UPF. So that the particle set can tend to the high likelihood region before the weights updated, thereby the impoverishment of particles can be overcome. While, the optimization process makes the particles which far away from the true state tend to the region where the true state has a greater probability of emergence, it can enhance the effect of each particle. Experiment results show that our modified particle filter algorithm uses fewer particles than the general particle filters and its performance outperforms them. And the accuracy of video tracking is improved. ©2009 IEEE.

关键词： Target tracking

来源：评论

学校读者我要写书评

暂无评论

An improved RED congestion algorithm based on partition of network segment

An improved RED congestion algorithm based on partition of n...

引用

5th International Conference on Natural Computation, ICNC 2009

作者： CtandongWang Li, Ting Wang, Huaibin Tianjin Key Laboratory of Intelligent Computing and Novel Software Technology Tianjin University of Technology Tianjin 300191 China Key Laboratory of Computer Vision and System Ministry of Education Tianjin University of Technology Tianjin 300191 China

ISBN: (纸本)9780769537368

Aiming at the growing problem of network congestion, an improved random early detection algorithm is proposed for the internal network and applied on a TCP congestion control algorithm based on partition of network segment. By using known network information and network topology information, the method is more effective, and better results are achieved. © 2009 IEEE.

关键词： Network topology

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：