检索结果-内蒙古大学图书馆

2022 IEEE International Conference on Multimedia and Expo, ICME 2022

作者： Liu, Xiaoliang Shen, Furao Zhao, Jian Nie, Changhai Nanjing University National Key Laboratory for Novel Software Technology China Nanjing University Department of Computer Science and Technology China School of Electronic Science and Engineering Nanjing University China

ISBN: (数字)9781665485630

ISBN: (纸本)9781665485630

Deep neural networks are powerful, but they also have short-comings such as their sensitivity to adversarial examples, noise, blur, occlusion, etc. Moreover, ensuring the reliability and robustness of deep neural network models is crucial for their application in safety-critical areas. Much previous work has been proposed to improve specific robustness. However, we find that the specific robustness is often improved at the sacrifice of the additional robustness or generalization ability of the neural network model. In particular, adversarial training methods significantly hurt the generalization performance on unperturbed data when improving adversarial robustness. In this paper, we propose a new data processing and training method, called AugRmixAT, which can simultaneously improve the generalization ability and multiple robustness of neural network models. Finally, we validate the effectiveness of AugRmixAT on the CIFAR-10/100 and Tiny-ImageNet datasets. The experiments demonstrate that AugR-mixAT can improve the model's generalization performance while enhancing the white-box robustness, black-box robustness, common corruption robustness, and partial occlusion robustness. © 2022 IEEE.

关键词： Neural network models

来源：评论

学校读者我要写书评

暂无评论

A hierarchical grid architecture based on computation/application metadata

引用

2nd International Workshop on Grid and Cooperative Computing, GCC 2003

作者： Dou, Wan-Chun Sun, Juan Yang, Da-Gang Cai, Shi-Jie State Key Laboratory for Novel Software Technology Department of Computer Science and Technology Nanjing University 210093 China

ISBN: (纸本)3540219889

Grid technology is the collection of tools and services that integration widely distributed, heterogeneous, and multi-organizational resource. For better understanding grid’s structure, an object-oriented framework of C/A component is explored. Then, a hierarchical grid architecture combining data Grid, information Grid, and knowledge Grid is explored based on a general hierarchical grid architecture. The conclusions are presented at last. © Springer-Verlag Berlin Heidelberg 2004.

关键词： Grid computing

来源：评论

学校读者我要写书评

暂无评论

A text-context-aware CNN network for multi-oriented and multi-language scene text detection 15

A text-context-aware CNN network for multi-oriented and mult...

引用

15th IAPR International Conference on Document Analysis and Recognition, ICDAR 2019

作者： Xiao, Yao Xue, Minglong Lu, Tong Wu, Yirui Palaiahnakote, Shivakumara National Key Lab for Novel Software Technology Nanjing University China College of Computer and Information Hohai University China Department of Computer System and Information Technology University of Malaya Malaysia

ISBN: (纸本)9781728128610

The existing deep learning based state-of-theart scene text detection methods treat scene texts a type of general objects, or segment text regions directly. The latter category achieves remarkable detection results on arbitraryorientation and large aspect ratios of scene texts based on instance segmentation algorithms. However, due to the lack of context information with consideration of scene text unique characteristics, directly applying instance segmentation to text detection task is prone to result in low accuracy, especially producing false positive detection results. To ease this problem, we propose a novel text-context-aware scene text detection CNN structure, which appropriately encodes channel and spatial attention information to construct context-aware and discriminative feature map for multi-oriented and multi-language text detection tasks. With high representation ability of textcontext-aware feature map, the proposed instance segmentation based method can not only robustly detect multi-oriented and multi-language text from natural scene images, but also produce better text detection results by greatly reducing false positives. Experiments on ICDAR2015 and ICDAR2017-MLT datasets show that the proposed method has achieved superior performances in precision, recall and F-measure than most of the existing studies. © 2019 IEEE.

关键词： Aspect ratio

来源：评论

学校读者我要写书评

暂无评论

SNN - A neural network based combination of software reliability growth models

SNN - A neural network based combination of software reliabi...

引用

1st International Conference on Information science and Engineering, ICISE2009

作者： Li, Ang Gu, Qing Feng, Guang-Cheng Chen, Dao-Xu State Key Laboratory for Novel Software Technology Department of Computer Science and Technology Nanjing University Nanjing China School of Computer Science and Engineering University of Aizu Japan

ISBN: (纸本)9780769538877

Applying SRGMs (software Reliability Growth Models) to real projects is a major concern in software reliability. Sometimes, it is hard to decide the best model for a specific project. Researchers have made a first step on solving this problem by combination, but the effect was limited in accuracy and adaptability. Aiming to improve the usability of the SRGMs, we propose a neural network based combination method to build accurate and adaptive SNN (Selective Neural Network) model. It avoids relying on a single model, thus reduces the risk to produce inaccurate predictions and improves the average performance in accuracy. Neural network and multi-criteria model selection strategy enable the SNN model to be adapted to various projects, producing accurate predictions. Experiment results show that the SNN model makes a notable improvement in accuracy compared with its component models and other combinational models do. ©2009 IEEE.

关键词： software reliability

来源：评论

学校读者我要写书评

暂无评论

Container lifecycle-aware scheduling for serverless computing

Container lifecycle-aware scheduling for serverless computin...

引用

作者： Wu, Song Tao, Zhiheng Fan, Hao Huang, Zhuo Zhang, Xinmin Jin, Hai Yu, Chen Cao, Chun 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 Wuhan China State Key Laboratory for Novel Software Technology Department of Computer Science and Technology Nanjing University Nanjing China

Elastic scaling in response to changes on demand is a main benefit of serverless computing. When bursty workloads arrive, a serverless platform launches many new containers and initializes function environments (known as cold starts), which incurs significant startup latency. To reduce cold starts, platforms usually pause a container after it serves a request, and reuse this container for subsequent requests. However, this reuse strategy cannot efficiently reduce cold starts because the schedulers are agnostic of container lifecycle. For example, it may ignore soon available containers or evict soon needed containers. We propose a container lifecycle-aware scheduling strategy for serverless computing, CAS. The key idea is to control distribution of requests and determine creation or eviction of containers according to different lifecycle phases of containers. We implement a prototype of CAS on OpenWhisk. Our evaluation shows that CAS reduces 81% cold starts and therefore brings a 63% reduction at 95th percentile latency compared with native scheduling strategy in OpenWhisk when there is worker contention between workloads, and does not add significant performance overhead. © 2021 John Wiley & Sons Ltd.

关键词： Containers

来源：评论

学校读者我要写书评

暂无评论

A utility-driven scheduling algorithms in cloud computing

引用

Revista Tecnica de la Facultad de Ingenieria Universidad del Zulia 2016年第2期39卷 282-287页

作者： Chen, Tao School of Management Science and Engineering Anhui University of Finance and Economics Bengbu Anhui233030 China Department of Computer Science and Technology Tongji University Shanghai201804 China State Key Laboratoryfor Novel Software Technology Nanjing University China

Cloud computing is a successful business model and utility paradigm which enables people to use computing power via Internet at anytime anywhere. How to schedule a task to cloud computing is an important issue. In cloud computing some users focus execution time the others maybe expense on tasks execution, so how to satisfy the users and achieve the balance between execution time and expense is very vital. Based on classic Min-Min algorithm, an efficient task scheduling algorithm was proposed where Cobb-Douglas utility function was employed to express user.s preference to time and cost, and maximized user.s utility. Multi-objective optimal solution has been obtained via the utility function. Experimental results evaluating such a mechanism show that the algorithm validate vividly.

关键词： Distributed computer systems

来源：评论

学校读者我要写书评

暂无评论

Fusion of visible and thermal images for facial expression recognition

引用

Frontiers of computer science 2014年第2期8卷 232-242页

作者： Shangfei WANG Shan HE Yue WU Menghua HE Qiang JI School of Computer Science and Technology University of Science and Technology of China Hefei 230027 China Key Lab of Computing and Communicating Software of Anhui Province Hefei 230027 China Department of Electrical Computer and Systems Engineering Rensselaer Polytechnic Institute Troy NY12180-3590 USA

Most present research into facial expression recognition focuses on the visible spectrum, which is sen- sitive to illumination change. In this paper, we focus on in- tegrating thermal infrared data with visible spectrum images for spontaneous facial expression recognition. First, the ac- tive appearance model AAM parameters and three defined head motion features are extracted from visible spectrum im- ages, and several thermal statistical features are extracted from infrared （IR） images. Second, feature selection is per- formed using the F-test statistic. Third, Bayesian networks BNs and support vector machines SVMs are proposed for both decision-level and feature-level fusion. Experiments on the natural visible and infrared facial expression （NVIE） spontaneous database show the effectiveness of the proposed methods, and demonstrate thermal 1R images＇ supplementary role for visible facial expression recognition.

关键词： facial expression recognition feature-level fu-sion decision-level fusion support vector machine Bayesiannetwork thermal infrared images visible spectrum images

来源：评论

学校读者我要写书评

暂无评论

Efficient software product-line model checking using induction and a SAT solver

引用

Frontiers of computer science 2018年第2期12卷 264-279页

作者： Fei HE Yuan GAO Liangze YIN Tsinghua National Laboratory for Information Science and Technology (TNList) Tsinghua University Beijing 100084 China Key Laboratory for Information System Security Ministry of Education Beijing 100084 China School of Software Tsinghua University Beijing 100084 China Department of Computer Science and Technology National University of Defense Technology Changsha 410073 China

software product line （SPL） engineering is increasingly being adopted in safety-critical systems. It is highly desirable to rigorously show that these systems are designed correctly. However, formal analysis for SPLs is more difficult than for single systems because an SPL may contain a large number of individual systems. In this paper, we propose an efficient model-checking technique for SPLs using induction and a SAT （Boolean satisfiability problem） solver. We show how an induction-based verification method can be adapted to the SPLs, with the help of a SAT solver. To combat the state space explosion problem, a novel technique that exploits the distinguishing characteristics of SPLs, called feature cube enlargement, is proposed to reduce the verification efforts. The incremental SAT mechanism is applied to further improve the efficiency. The correctness of our technique is proved. Experimental results show dramatic improvement of our technique over the existing binary decision diagram （BDD）-based techniques.

关键词： software product line model checking satisfiability

来源：评论

学校读者我要写书评

暂无评论

Active BSVM learning for relevance feedback in content-based sketch retrieval

Active BSVM learning for relevance feedback in content-based...

引用

5th IASTED International Conference on Signal Processing, Pattern Recognition, and Applications, SPPRA 2008

作者： Liang, Shuang Sun, Zhengxing State Key Lab. for Novel Software Technology Nanjing University Nanjing 210093 China Department of Computer Science and Technology Nanjing University Nanjing 210093 China

ISBN: (纸本)9780889867178

The availability of relevance feedback is held back by the problem of the imbalance and limited size of labeled training data, as well as the real-time requirement of online interaction demands. In this paper, we propose a relevance feedback algorithm called active biased SVM (BSVM) learning, in which biased classification and active learning are employed to address these difficulties. The algorithm is applied to content-based sketch retrieval (CBSR), and the experiments prove both the effectiveness and efficiency of the proposed approach.

关键词： Active learning Biased SVM (BSVM) Content-based sketch retrieval (CBSR) Relevance feedback

来源：评论

学校读者我要写书评

暂无评论

A compositional multiple policies operating system security model

A compositional multiple policies operating system security ...

引用

8th International Workshop on Information Security Applications, WISA 2007

作者： Lei, Xia Wei, Huang Hao, Huang State Key Laboratory for Novel Software Technology Department of Computer Science and Technology Nanjing University 22 Hankou Road Nanjing 210093 China

ISBN: (纸本)354077534X

Multilevel security policies aim at only confidentiality assurance, with less consideration on integrity assurance and weakness in expressing channel control policies. Besides, the trusted subjects it introduces to handle the information flow "downgrade" have many security flaws. Moreover, increasing diversity of the computing environments results in various security requirements. However, current mainstream security models are aiming at only one or few requirements of them each. The Multi-Policy Views Security Model is presented, which is based on the MLS model, combining the domain and role attributes to the model, to enforce the expression power in channel control policies, make permission management more fine-grained and enhance the ability of confining the permission of the trusted subjects. Moreover, MPVSM has integrated the properties and functions of MLS, Domain-Type and Role Based models into one unified model. It is able to enforce multi-policy views in operating system in a flexible way. © Springer-Verlag Berlin Heidelberg 2007.

关键词： Security of data

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：