检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Li, Bo Yun, Jing College of Data Science and Application Inner Mongolia University of Technology Inner Mongolia Huhhot010080 China Inner Mongolia Autonomous Region Engineering and Technology Research Center of Big Data-Based Software Service Inner Mongolia Huhhot010080 China Inner Mongolia Key Laboratory of Beijiang Cyberspace Security Inner Mongolia Huhhot010080 China

Multi-view clustering can explore common semantics from multiple views and has received increasing attention in recent years. However, current methods focus on learning consistency in representation, neglecting the contribution of each view’s complementarity aspect in representation learning. This limit poses a significant challenge in multi-view representation learning. This paper proposes a novel multi-view clustering framework that introduces a disentangled variational autoencoder that separates multi-view into shared and private information, i.e., consistency and complementarity information. We first learn informative and consistent representations by maximizing mutual information across different views through contrastive learning. This process will ignore complementary information. Then, we employ consistency inference constraints to explicitly utilize complementary information when attempting to seek the consistency of shared information across all views. Specifically, we perform a within-reconstruction using the private and shared information of each view and a cross-reconstruction using the shared information of all views. The dual consistency constraints are not only effective in improving the representation quality of data but also easy to extend to other scenarios, especially in complex multi-view scenes. This could be the first attempt to employ dual consistent constraint in a unified MVC theoretical framework. During the training procedure, the consistency and complementarity features are jointly optimized. Extensive experiments show that our method outperforms baseline methods. © 2025, CC BY.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

An Auto-Parallel Method for Deep Learning Models Based on Genetic Algorithm

An Auto-Parallel Method for Deep Learning Models Based on Ge...

引用

International Conference on Parallel and Distributed Systems (ICPADS)

作者： Yan Zeng ChengChuang Huang YiJie Ni ChunBao Zhou JiLin Zhang Jue Wang MingYao Zhou MeiTing Xue YunQuan Zhang School of Computer Science and Technology Hangzhou Dianzi University Hangzhou China Key Laboratory for Modeling and Simulation of Complex Systems Ministry of Education Hangzhou China Data Security Governance Zhejiang Engineering Research Center Hangzhou China School of ITMO Joint Institute Hangzhou Dianzi University Hangzhou China Institute of Computer Network Information Center of the Chinese Academy of Sciences Beijing China HuaWei State Key Laboratory of Computer Architecture Institute of Computing Technology of the Chinese Academy of Sciences Beijing China

As the size of datasets and neural network models increases, automatic parallelization methods for models have become a research hotspot in recent years. The existing auto-parallel methods based on machine learning or graph algorithms still have issues with search efficiency and applicability. This paper proposes an automatic parallel method based on a dual-population genetic algorithm, TGA, which transforms model partitioning and placement into an integer linear programming problem and constructs a cost model to evaluate the solution. The solution space is built using the neural network’s dataflow graph and device cluster’s topology, and the dual-population genetic algorithm is used to search for the optimal model parallel strategy. Experiments with various models show that the proposed method can improve single-step execution time by up to 42% compared to the Baechi method and up to 37.7% compared to the Hierarchical method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

AdAUC: End-to-end Adversarial AUC Optimization Against Long-tail Problems

arXiv

引用

arXiv 2022年

作者： Hou, Wenzheng Xu, Qianqian Yang, Zhiyong Bao, Shilong He, Yuan Huang, Qingming Key Laboratory of Intelligent Information Processing Institute of Computing Technology CAS Beijing China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China State Key Laboratory of Information Security Institute of Information Engineering CAS Beijing China School of Cyber Security University of Chinese Academy of Sciences Beijing China Alibaba Group Beijing China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of Sciences Beijing China Artificial Intelligence Research Center Peng Cheng Laboratory Shenzhen China

It is well-known that deep learning models are vulnerable to adversarial examples. Existing studies of adversarial training have made great progress against this challenge. As a typical trait, they often assume that the class distribution is overall balanced. However, long-tail datasets are ubiquitous in a wide spectrum of applications, where the amount of head class instances is larger than the tail classes. Under such a scenario, AUC is a much more reasonable metric than accuracy since it is insensitive toward class distribution. Motivated by this, we present an early trial to explore adversarial training methods to optimize AUC. The main challenge lies in that the positive and negative examples are tightly coupled in the objective function. As a direct result, one cannot generate adversarial examples without a full scan of the dataset. To address this issue, based on a concavity regularization scheme, we reformulate the AUC optimization problem as a saddle point problem, where the objective becomes an instance-wise function. This leads to an end-to-end training protocol. Furthermore, we provide a convergence guarantee of the proposed algorithm. Our analysis differs from the existing studies since the algorithm is asked to generate adversarial examples by calculating the gradient of a min-max problem. Finally, the extensive experimental results show the performance and robustness of our algorithm in three long-tail datasets. © 2022, CC BY.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Discrete unified gas kinetic scheme for the conservative Allen-Cahn equation

arXiv

引用

arXiv 2022年

作者： Zhang, Chunhua Liang, Hong Guo, Zhaoli Wang, Lian-Ping Guangdong Provincial Key Laboratory of Turbulence Research and Applications Department of Mechanics and Aerospace Engineering Southern University of Science and Technology Guangdong Shenzhen518055 China Department of Physics Hangzhou Dianzi University Hangzhou310018 China State Key Laboratory of Coal Combustion Huazhong University of Science and Technology Wuhan430074 China Center for Complex Flows and Soft Matter Research Southern University of Science and Technology Guangdong Shenzhen518055 China Guangdong-Hong Kong-Macao Joint Laboratory for Data-Driven Fluid Mechanics and Engineering Applications Southern University of Science and Technology Guangdong Shenzhen518055 China

In this paper, the discrete unified gas kinetic scheme (DUGKS) with an improved microflux across the cell interface for the conservative Allen-Cahn equation (CACE) is proposed. In the context of DUGKS, the recovered kinetic equation from the flux evaluation with linear reconstruction in the previous DUGKS is analyzed. It is found that the calculated microflux across the cell interface is only the solution to the target kinetic equation with first order accuracy, which can result in an inaccurate CACE since the force term is involved or the first moment of the collision model has no conservation property. To correctly recover the kinetic equation up to the second order accuracy, the value of the distribution function that will propagate along the characteristic line with ending point at the cell interface is appropriated by the parabolic reconstruction instead of the linear reconstruction. To validate the accuracy of the present DUGKS for the CACE, several benchmark problems, including the diagonal translation of a circular interface, the rotation of a Zaleska disk and the deformation of a circular interface, have been simulated. Numerical results show that the present DUGKS scheme is able to capture the interface with improved accuracy when compared with the previous DUGKS. Copyright © 2022, The Authors. All rights reserved.

关键词： Distribution functions

来源：评论

学校读者我要写书评

暂无评论

Heterogeneous graph completion collaborative network for attribute-missing heterogeneous graph representation learning

引用

Expert Systems with Applications 2025年 290卷

作者： Yang, Yuanjun Ou, Weihua Fu, Sichao Chen, Jiamin Wu, Yunshun School of Mathematical Sciences Guizhou Normal University Guiyang China School of Big Data and Computer Science Guizhou Normal University Guiyang China School of Electronic Information and Communication Huazhong University of Science and Technology Hubei China Engineering Research Center of Micro-Nano and Intelligent Manufacturing Ministry of Education Kaili Guizhou China

Heterogeneous graph representation learning has attracted significant attention and found widespread application in various real-world scenarios. Due to various uncontrollable factors, heterogeneous graph data often suffers from missing attributes, which affects the performance of existing heterogeneous graph representation learning. Recently proposed attribute-missing heterogeneous graph representation learning methods provide an effective solution to the above issues. However, previous completion methods overlook nodes’ high-order semantic information, leading to insufficient utilization of high-order semantic information and ultimately resulting in poor performance of downstream tasks. Additionally, the process of lacking effective heterogeneous graphs’ information collaborative interaction leads to insufficient representation learning. To solve the above problems, we propose a Heterogeneous Graph Completion Collaborative Network called HGCCN for attribute-missing heterogeneous graph representation learning, which first utilizes high-order diffusion completion method to complete missing attributes, and then introduces a graph transformer with a multi-head attention mechanism to adaptively perform collaborative interaction among multiple completed semantic attributes. Firstly, we construct multiple meta-path subgraphs to explore the nodes’ high-order semantic information in the completion process. Secondly, we propose a novel diffusion completion method that leverages high-order semantic information to more effectively complete missing attributes. Thirdly, we introduce a graph transformer to achieve the heterogeneous graphs’ high-order semantic information collaborative interaction, further obtaining the enhanced completed node attributes. In summary, compared to existing methods, the advantage of HGCCN lies in its ability to utilize high-order semantic information from different meta-path subgraphs for completion, as well as enhancing the completion embeddings th

关键词： Graph embeddings

来源：评论

学校读者我要写书评

暂无评论

SCB: Flexible and Efficient Asymmetric Computations Utilizing Symmetric Cryptosystems Implemented with Intel SGX

SCB: Flexible and Efficient Asymmetric Computations Utilizin...

引用

IEEE International Conference on Performance, Computing and Communications (IPCCC)

作者： Wenyi Ouyang Qiongxiao Wang Wei Wang Jingqiang Lin Yaxi He Institute of Information Engineering Chinese Academy of Sciences Beijing China School of Cyber Security University of Chinese Academy of Sciences Beijing China Data Assurance and Communications Security Research Center Chinese Academy of Sciences Beijing China Beijing Research Institute University of Science and Technology of China Beijing China School of Cyber Security University of Science and Technology of China Hefei China

ISBN: (纸本)9781665443326

The wide spread of cloud computing and the rapid growth of online activities raise the demand for privacy enhancing cryptography, such as identity-based encryption (IBE), attribute-based encryption (ABE), and group/ring signature, which are mostly asymmetric. However, implementing a practical asymmetric cryptosystem with privacy enhancing functionalities faces the big challenge of large computing overhead that stems from the hard mathematical problems they are based on, which in reality hinders the further deployment of such *** this paper, we propose "SGX Cipher Box" (SCB), a method of constructing flexible and efficient asymmetric computations by encapsulating symmetric algorithms into Intel SGX enclaves. Two types of enclaves are designed for serving computations with the public key and the private key, respectively. The security assumptions of asymmetric algorithms can be achieved in SCB based on guarantees offered by SGX. SCB is of great flexibility that it facilitates implementing various asymmetric computations with limited enclaves. We present four SCB-based cryptosystems, provisioning asymmetric primitives, IBE, ABE, and ring signature, respectively. The four prototype cryptosystems are implemented and benchmarked in terms of computing efficiency. Experimental results show that SCB surpasses conventional asymmetric cryptosystem implementations in performance.

关键词： Privacy Cloud computing Identity-based encryption Conferences Public key Prototypes Computational efficiency

来源：评论

学校读者我要写书评

暂无评论

Modeling and analysis of the dynamic mechanism of rumor propagation considering the process of rumor refuting

Nonlinear Science

引用

Nonlinear science 2025年 3卷

作者： Xutao Li Junshan Pan Yong Hu Hubei Provincial Engineering Research Center of Intelligent Connected Vehicle Network Security Key Laboratory of Intelligent Sensing System and Security (Hubei University) Ministry of Education School of Cyber Science and Technology Hubei University Wuhan 430062 Hubei China School of Physics Huazhong University of Science and Technology Wuhan 430074 Hubei China

The boost of rumor spreading by the Internet has caused a series of social problems. Refuting the rumor timely can alleviate the harm caused by rumor dissemination. To investigate the propagation mechanism of rumor affected by rumor refuting on online social platforms, we propose the Susceptible-Infective-Remover-Debunker-Remover (SIRDR) model to describe the propagation dynamics of rumor and rumor refutation. To conform closely to the actual scenarios of rumor dissemination, we consider a refined division of the population involved in propagation and introduce an information exposure attractiveness index and forgetting mechanisms to characterize individual spreading behavior. We then investigate the extent to which the suppression of rumor dissemination by refuting rumors at different spreading stages. The proposed model is validated by the analysis of real datasets collected from Sina Weibo and Twitter. The R-squared values of data 1 are 0.947 and 0.985, the R-squared values of data 2 are 0.993 and 0.976, while the R-squared values of data 3 are 0.948 and 0.956, indicating that our model is very consistent with the real rumor propagation process. Especially, we find out that the suppression effect of rumor refutation on rumor propagation varies greatly with its release time. Our work may offer theoretical support for rumors governance in cyberspace.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Evaluating the Cache Side Channel Attacks Against ECDSA 15th

Evaluating the Cache Side Channel Attacks Against ECDSA

引用

15th International Conference on Information security and Cryptology, Inscrypt 2019

作者： Ma, Ziqiang Cai, Quanwei Lin, Jingqiang Jing, Jiwu Ye, Dingfeng Meng, Lingjia State Key Laboratory of Information Security Institute of Information Engineering Chinese Academy of Sciences Beijing China Data Assurance and Communication Security Research Center CAS Beijing China School of Cyber Security University of Chinese Academy of Sciences Beijing China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China

ISBN: (纸本)9783030429201

Various attacks are proposed against different ECDSA implementations: the key-related data are acquired through cache side channels, and then processed to recover the private key. For each cache side channel attack, the requirements of the data qualified for sequent processing vary greatly, and the success probability of private key recovery relies on both the acquired data and the parameters of data processing. So it is difficult to tell, for a certain ECDSA implementation, (a) how many signatures does a cache side channel attack need to recover the private key? or which attack performs the best? and (b) what kind of threat level exists due to potential side channel attacks, if the ECDSA implementation runs for a number of signatures on an unprotected system with cache side channels? Currently, there is no quantitative metric to fairly answer the questions. Such a metric to evaluate cache side channel attacks, will provide a reference for the adversaries to choose the suitable attack, and also for the defenders to set up protections for the certain ECDSA implementation (e.g., updating the private key after it has been used for a certain number of signatures). In this paper, we design an evaluation approach to quantitatively compare the cache side channel attacks against ECDSA. The expected minimum number of signatures needed for at least one successful private key recovery, is proposed as the metric, and this metric considers both the data requirements and the success probability. We apply the approach to evaluate various cache side channel attacks against ECDSA. By calculating the metric, we obtain (a) for each attack, the optimal parameters with the minimum number of signatures needed, and (b) for each ECDSA implementation, the minimum number of signatures that will be enough for at least one successful private key recovery of some cache side channel attacks. © 2020, Springer Nature Switzerland AG.

关键词： Side channel attack

来源：评论

学校读者我要写书评

暂无评论

VulDeeLocator: A deep learning-based fine-grained vulnerability detector

arXiv

引用

arXiv 2020年

作者： Li, Zhen Zou, Deqing Xu, Shouhuai Chen, Zhaoxuan Zhu, Yawei Jin, Hai The National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab Big Data Security Engineering Research Center School of Cyber Science and Engineering Huazhong University of Science and Technology Wuhan430074 China School of Cyber Security and Computer Hebei University Baoding071002 China The Department of Computer Science University of Colorado Colorado Springs CO80918 United States University of Texas San Antonio United States

Automatically detecting software vulnerabilities is an important problem that has attracted much attention from the academic research community. However, existing vulnerability detectors still cannot achieve the vulnerability detection capability and the locating precision that would warrant their adoption for real-world use. In this paper, we present a vulnerability detector that can simultaneously achieve a high detection capability and a high locating precision, dubbed Vulnerability Deep learning-based Locator (VulDeeLocator). In the course of designing VulDeeLocator, we encounter difficulties including how to accommodate semantic relations between the definitions of types as well as macros and their uses across files, how to accommodate accurate control flows and variable define-use relations, and how to achieve high locating precision. We solve these difficulties by using two innovative ideas: (i) leveraging intermediate code to accommodate extra semantic information, and (ii) using the notion of granularity refinement to pin down locations of vulnerabilities. When applied to 200 files randomly selected from three real-world software products, VulDeeLocator detects 18 confirmed vulnerabilities (i.e., true-positives). Among them, 16 vulnerabilities correspond to known vulnerabilities;the other two are not reported in the National Vulnerability database (NVD) but have been "silently" patched by the vendor of Libav when releasing newer versions. Copyright © 2020, The Authors. All rights reserved.

关键词： Location

来源：评论

学校读者我要写书评

暂无评论

Delay-Optimal Edge Cache Replacement with Non-Markovian Content Fetching

Delay-Optimal Edge Cache Replacement with Non-Markovian Cont...

引用

GLOBECOM 2020 - 2020 IEEE Global Communications Conference

作者： Nannan Zhang Wei Wang Pan Zhou Aiping Huang College of Information Science and Electronic Engineering Zhejiang Provincial Key Laboratory of Information Processing Communication and Networking Zhejiang University Hangzhou China Hubei Engineering Research Center on Big Data Security School of Cyber Science and Engineering Huazhong University of Science and Technology Wuhan China

ISBN: (数字)9781728182988

ISBN: (纸本)9781728182995

Leveraging the content caching technology, popular contents can be in close proximity to the users, which reduces the delay effectively. In this paper, we dedicate to minimizing the long-term average delay for the cache-enabled wireless networks, in which the fetching delay is considered when the uncached contents are retrieved from remote servers. We construct a decision-theoretic framework for this delay-optimal content-caching problem. The most challenging obstacle is that the consideration of imperfect content fetching breaks the Markovian property in the decision-theoretic framework, which makes the problem difficult to solve. To overcome this obstacle, we propose an equivalent transformation by analyzing the queue dynamics during the content fetching delay, so that the transformed problem can be modeled as an infinite horizon semi-Markov decision process (SMDP). To handle the curse of dimensionality for solving the SMDP, we decompose the global optimality equation into several per-content equations, and propose a low-complexity delay-optimal content caching algorithm. Simulation results show that the proposed algorithm achieves significantly lower delay than conventional caching algorithms.

关键词： Delays Servers Mathematical model Wireless networks Queueing analysis Optimization Time measurement

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：