检索结果-内蒙古大学图书馆

IEEE 15th International Conference on data mining Workshops (ICDMW)

作者： Qi, Zhiquan Tian, Yingjie Niu, Lingfeng Meng, Fan Cui, Limeng Shi, Yong Chinese Acad Sci Res Ctr Fictitious Econ & Data Sci Key Lab Big Data Min & Knowledge Management Beijing 100190 Peoples R China

ISBN: (纸本)9781467384933

How to balance the speed and the quality is always a challenging issue in pedestrian detection. In this paper, we introduce the Learning model Using Privileged Information (LUPI), which can accelerate the convergence rate of learning and effectively improve the quality without sacrificing the speed. In more detail, we give the clear definition of the privileged information, which is only available at the training stage but is never available for the testing set, for the pedestrian detection problem and show how much the privileged information helps the detector to improve the quality. All experimental results show the robustness and effectiveness of the proposed method, at the same time show that the privileged information offers a significant improvement.

关键词： pedestrian detection privileged information support vector machine

来源：评论

学校读者我要写书评

暂无评论

ESL-SELO:A Robust Image Denoising Algorithm with Penalty

引用

Acta Mathematicae Applicatae Sinica 2017年第3期33卷 753-770页

作者： Guo-hua WANG San-guo ZHANG Peng-jie DAI School of Mathematical Sciences University of Chinese Academy of SciencesBeijing 100049China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of SciencesBeijing100049China School of Business Remin University of ChinaBeijing 100872China

Robust image recovery methods have been attracted more and more attention in recent decades for its good property of tolerating system errors or measuring noise. In this paper, we propose a new robust method （ESL-SELO） to recover nosing image, which combine exponential loss function and seamless-L0 （SELO） penalty function to guarantee both accuracy and robustness of the estimator. Theoretical result showed that our method has a local optimal solution and good asymptotic properties. Finally, we compare our method with other methods in simulation which shows better robustness and takes much less time.

关键词： denoising algorithm penalty function robust SELO

来源：评论

学校读者我要写书评

暂无评论

Single Image Super-Resolution Reconstruction Based on Edge-Preserving with External and Internal Gradient Prior knowledge 13th

Single Image Super-Resolution Reconstruction Based on Edge-P...

引用

13th Asian Conference on Computer Vision (ACCV)

作者： Wang, Ruxin Han, Congying Li, Mingqiang Guo, Tiande Univ Chinese Acad Sci Sch Math Sci Key Lab Big Data Min & Knowledge Management Beijing Peoples R China

ISBN: (纸本)9783319544076;9783319544069

Single image super-resolution (SISR) reconstruction is currently a very fundamental and significant task in image processing. Instead of upscaling the image in spatial domain, we propose a novel SISR method based on edge preserving integrating the external gradient priors by deep learning method (auto-encoder network) and internal gradient priors using non-local total variation (NLTV). The gradient domain effectively reflects the high frequency details and edge information of nature image to some extent. The joint perspective exploits the complementary advantages of external and internal gradient prior knowledge for reconstructing the HR image. The experimental results demonstrate the effectiveness of our approach over several state-of-art SISR methods.

关键词： Optical resolving power

来源：评论

学校读者我要写书评

暂无评论

Reliability and Joint Reliability Importance in a Consecutive-k-Within-m-out-of-n:F System With Markov-Dependent Components

引用

IEEE TRANSACTIONS ON RELIABILITY 2016年第2期65卷 802-815页

作者： Zhu, Xiaoyan Boushaba, Mahmoud Reghioua, Mohamed Univ Chinese Acad Sci Sch Econ & Management Beijing Peoples R China Chinese Acad Sci Key Lab Big Data Min & Knowledge Management Beijing Peoples R China Ecole Normale Super Constantine Constantine Algeria

This paper studies a consecutive-k-within-m-out-of-n:F system with Markov-dependent components;that is, the reliability of a component depends on its neighbors. Using probability generating functions, the closed-form formula for reliability of the consecutive-k-within-m-out-of-n:F system with Markov-dependent components and a closed-form formula for joint reliability importance (JRI) of two components in such a system are derived. The JRI of two components evaluates the interaction effect between the components on contributing to system reliability. A formula of the JRI of more than two components are also derived and presented. Many real systems and procedures, such as radar detection systems, pipeline systems, quality inspection procedures, and so on, can be modeled as a consecutive-k-within-m-out-of-n:F system, in which components are Markov-dependent. The present results can evaluate the reliability of these systems or the accuracy of the procedures as well as the contributions of components to the system reliability or the accuracy of the procedures. The applications of the present formulas are demonstrated through the numerical examples. The examples show the changes of system radiabilities and the changes among the JRI values of different pairs of components in consecutive-k-within-m-out-of-n:F systems. The JRI values of Markov-dependent components are also compared to the JRI values of s-independent components.

关键词： Consecutive-k-within-m-out-of-n:F system Joint reliability importance Markov-dependent components probability generating function

来源：评论

学校读者我要写书评

暂无评论

Large-scale structural learning and predicting via hashing approximation

引用

NEURAL COMPUTING & APPLICATIONS 2019年第7期31卷 2889-2903页

作者： Chen, Dandan Tian, Yingjie Univ Chinese Acad Sci Sch Math Sci Beijing 100049 Peoples R China Chinese Acad Sci Res Ctr Fictitious Econ & Data Sci Key Lab Big Data Min & Knowledge Management Beijing 100190 Peoples R China

By combining the structural information with nonparallel support vector machine, structural nonparallel support vector machine (SNPSVM) can fully exploit prior knowledge to directly improve the algorithm's generalization capacity. However, the scalability issue how to train SNPSVM efficiently on data with huge dimensions has not been studied. In this paper, we integrate linear SNPSVM with b-bit minwise hashing scheme to speedup the training phase for large-scale and high-dimensional statistical learning, and then we address the problem of speeding-up its prediction phase via locality-sensitive hashing. For one-against-one multi-class classification problems, a two-stage strategy is put forward: a series of hash-based classifiers are built in order to approximate the exact results and filter the hypothesis space in the first stage and then the classification can be refined by solving a multi-class SNPSVM on the remaining classes in the second stage. The proposed method can deal with large-scale classification problems with a huge number of features. Experimental results on two large-scale datasets (i.e., news20 and webspam) demonstrate the efficiency of structural learning via b-bit minwise hashing. Experimental results on the ImageNet-BOF dataset, and several large-scale UCI datasets show that the proposed hash-based prediction can be more than two orders of magnitude faster than the exact classifier with minor losses in quality.

关键词： Nonparallel support vector machine Structural information Locality-sensitive hashing Minwise hashing

来源：评论

学校读者我要写书评

暂无评论

Enhanced Semantic Head for Cascade Instance Segmentation

Enhanced Semantic Head for Cascade Instance Segmentation

引用

2022 IEEE International Conference on Multimedia and Expo, ICME 2022

作者： Huang, Xuerong Su, Li Li, Guorong Zhang, Xinfeng Qing, Laiyun Huang, Qingming University of Chinese Academy of Sciences Beijing China Cas Key Lab of Big Data Mining and Knowledge Management Beijing China Institute of Computing Technology Cas Key Lab of Intelligent Information Processing Beijing China

ISBN: (数字)9781665485630

ISBN: (纸本)9781665485630

Recently, cascade instance segmentation inspired by cascade object detection has achieved notable performance. Due to the lack of global information, many methods suffer from incomplete segmentation such as missing edge regions and discontinuities within instances. To solve this problem, we proposed an effective and flexible semantic head to extract enhanced spatial context information. A vision transformer is utilized to generate global context features, and a convolution network is adopted to generate spatial context features. After combining the two modules, we obtain enhanced semantic segmentation features for segmentation. Extensive experiments show that the enhanced semantic head achieves 40.6% and 42.3% mask AP for cascade predictor HTC and DSC, which surpass about 0.9 and 1.4 percentage points respectively. The enhanced semantic head is universal and effective to improve the performance of different cascade predictors. © 2022 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

A linearly convergent stochastic recursive gradient method for convex optimization

引用

OPTIMIZATION LETTERS 2020年第8期14卷 2265-2283页

作者： Liu, Yan Wang, Xiao Guo, Tiande Univ Chinese Acad Sci Sch Math Sci Beijing 100049 Peoples R China Univ Chinese Acad Sci Key Lab Big Data Min & Knowledge Management Beijing 100049 Peoples R China

The stochastic recursive gradient algorithm (SARAH) (Nguyen et al. in: Neural information processing systems, pp 2613-2621, 2017) attracts much interest recently. It admits a simple recursive framework for updating stochastic gradient estimates. Motivated by this, in this paper we propose a new stochastic recursive gradient method, called SARAH-I. Different from SARAH, SARAH-I incorporates importance sampling strategy and computes the full gradient at the last iterate in each inner iteration. We show that the sequence of distances between iterates and the optima set is linearly convergent under both strong convexity and non-strong convexity conditions. Furthermore, we propose to use the Barzilai-Borwein (BB) approach to adaptively compute step sizes for SARAH-I, and name the resulting method as SARAH-I-BB. We establish its convergence and complexity properties in different cases. Finally numerical tests are reported to indicate promising performances of proposed algorithms.

关键词： Stochastic optimization Stochastic gradient BB method Linear convergence rate Complexity

来源：评论

学校读者我要写书评

暂无评论

Trade credit or credit insurance? A green supply chain finance design scheme with multi-objective programming

引用

Journal of Intelligent and Fuzzy Systems 2023年第2期45卷 2707-2721页

作者： Zhang, Linzi Shi, Yong School of Economics and Management University of Chinese Academy of Sciences Beijing China University of Birmingham Edgbaston Birmingham United Kingdom Research Center on Fictitious Economy Data Science Chinese Academy of Science Beijing China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of Sciences Beijing China

Classical supply chain finance (SCF) primarily focuses on the financial service among all upstream and downstream supply chain participants. Due to the continuously deteriorating of the ecological environment, an environmental-friendly SCF system is urgently needed. In this paper, we propose a novel SCF design scheme with environmental concerns, i.e., green supply chain finance (GSCF), consider the financing channels both from banks and from consumers, and design a bi-objective optimization model that depicts the trade-off between the benefit and the emission. Further, an improved normalized normal constraint (INNC) Pareto method is developed to address the optimal financing strategy of the bi-objective model. We then conduct a numerical case of a Taiwanese steel firm to verify the effectiveness and efficiency of our method. Results show that our model provides a portfolio of optimal solutions on Pareto frontier which can be applied as an effective decision support system when designing a GSCF. Furthermore, the sensitivity analysis also presents the impact of environmental investment cost, technological ratio of companies and the interest rate of trade credit on the optimal configuration of the GSCF. © 2023 - IOS Press. All rights reserved.

关键词： Multiobjective optimization

来源：评论

学校读者我要写书评

暂无评论

A branch-and-bound algorithm for solving max-k-cut problem

引用

JOURNAL OF GLOBAL OPTIMIZATION 2021年第2期81卷 367-389页

作者： Lu, Cheng Deng, Zhibin North China Elect Power Univ Sch Econ & Management Beijing 102206 Peoples R China Chinese Acad Sci Univ Chinese Acad Sci Sch Econ & Management Key Lab Big Data Min & Knowledge Management Beijing 100190 Peoples R China

The max-k-cut problem is one of the most well-known combinatorial optimization problems. In this paper, we design an efficient branch-and-bound algorithm to solve the max-k-cut problem. We propose a semidefinite relaxation that is as tight as the conventional semidefinite relaxations in the literature, but is more suitable as a relaxation method in a branch-and-bound framework. We then develop a branch-and-bound algorithm that exploits the unique structure of the proposed semidefinite relaxation, and applies a branching method different from the one commonly used in the existing algorithms. The symmetric structure of the solution set of the max-k-cut problem is discussed, and a strategy is devised to reduce the redundancy of subproblems in the enumeration procedure. The numerical results show that the proposed algorithm is promising. It performs better than Gurobi on instances that have more than 350 edges, and is more efficient than the state-of-the-art algorithm bundleBC on certain types of test instances.

关键词： Max-k-cut Semidefinite relaxation Branch-and-bound algorithm

来源：评论

学校读者我要写书评

暂无评论

The long-run dynamic relationship between exchange rate and its attention index: Based on DCCA and TOP method

引用

PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS 2016年第0期453卷 108-115页

作者： Wang, Xuan Guo, Kun Lu, Xiaolin Univ Chinese Acad Sci Sch Management Beijing Peoples R China Chinese Acad Sci Res Ctr Fictitious Econ & Data Sci Beijing Peoples R China Chinese Acad Sci Key Lab Big Data Min & Knowledge Management Beijing 100864 Peoples R China

The behavior information of financial market plays a more and more important role in modern economic system. The behavior information reflected in INTERNET search data has already been used in short-term prediction for exchange rate, stock market return, house price and so on. However, the long-run relationship between behavior information and financial market fluctuation has not been studied systematically. Further, most traditional statistic methods and econometric models could not catch the dynamic and non-linear relationship. An attention index of CNY/USD exchange rate is constructed based on search data from 360 search engine of China in this paper. Then the DCCA and Thermal Optimal Path methods are used to explore the long-run dynamic relationship between CNY/USD exchange rate and the corresponding attention index. The results show that the significant interdependency exists and the change of exchange rate is 1-2 days lag behind the attention index. (C) 2016 Elsevier B.V. All rights reserved.

关键词： Exchange rate Attention index Thermal optimal path DCCA Web-searching

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：