检索结果-内蒙古大学图书馆

18th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2012

作者： Gong, Pinghua Ye, Jieping Zhang, Changshui State Key Laboratory on Intelligent Technology and Systems Department of Automation Tsinghua University Beijing 100084 China Center for Evolutionary Medicine and Informatics Biodesign Institute Arizona State University Tempe AZ 85287 United States

ISBN: (纸本)9781450314626

Multi-task learning (MTL) aims to improve the performance of multiple related tasks by exploiting the intrinsic relationships among them. Recently, multi-task feature learning algorithms have received increasing attention and they have been successfully applied to many applications involving high dimensional data. However, they assume that all tasks share a common set of features, which is too restrictive and may not hold in real-world applications, since outlier tasks often exist. In this paper, we propose a Robust Multi-Task Feature Learning algorithm (rMTFL) which simultaneously captures a common set of features among relevant tasks and identifies outlier tasks. Specifically, we decompose the weight (model) matrix for all tasks into two components. We impose the well-known group Lasso penalty on row groups of the first component for capturing the shared features among relevant tasks. To simultaneously identify the outlier tasks, we impose the same group Lasso penalty but on column groups of the second component. We propose to employ the accelerated gradient descent to efficiently solve the optimization problem in rMTFL, and show that the proposed algorithm is scalable to large-size problems. In addition, we provide a detailed theoretical analysis on the proposed rMTFL formulation. Specifically, we present a theoretical bound to measure how well our proposed rMTFL approximates the true evaluation, and provide bounds to measure the error between the estimated weights of rMTFL and the underlying true weights. Moreover, by assuming that the underlying true weights are above the noise level, we present a sound theoretical result to show how to obtain the underlying true shared features and outlier tasks (sparsity patterns). Empirical studies on both synthetic and real-world data demonstrate that our proposed rMTFL is capable of simultaneously capturing shared features among tasks and identifying outlier tasks. © 2012 ACM.

关键词： Learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Multi-way Constrained Spectral Clustering by Nonnegative Restriction

Multi-way Constrained Spectral Clustering by Nonnegative Res...

引用

International Conference on Pattern Recognition

作者： Han Hu Jiahuan Zhou Jianjiang Feng Jie Zhou State Key Laboratory on Intelligent Technology and Systems TNList Department of Automation Tsinghua University

ISBN: (纸本)9781467322164

Clustering often benefits from side information. In this paper, we consider the problem of multi-way constrained spectral clustering with pairwise constraints which encode whether two nodes belong to the same cluster or not. Due to the nontransitive property of cannot-link constraints, it is hard to incorporate cannot-link constraints into the framework. We settle this difficulty by restricting the spectral vectors with nonnegative elements. An iterative method is proposed to optimize the objective. Experiments on several publicly available datasets demonstrate the effectiveness of our algorithm.

关键词： iterative methods optimisation pattern clustering vectors

来源：评论

学校读者我要写书评

暂无评论

Mining sub-categories for object detection

Mining sub-categories for object detection

引用

International Conference on Pattern Recognition

作者： Jifeng Dai Jianjiang Feng Jie Zhou State Key Laboratory on Intelligent Technology and Systems TNList Department of Automation Tsinghua University Beijing China

ISBN: (纸本)9781467322164

The visual concept of an object category is usually composed of a set of sub-categories corresponding to different sub-classes, perspectives, spatial configurations and etc. Existing detector training algorithms usually require extensive supervisory information to achieve a satisfactory performance for sub-categorization. In this paper, we propose a detector training algorithm which can automatically mine meaningful sub-categories utilizing only the image contents within the training bounding boxes. The number of sub-categories can also be determined automatically. The mined sub-categories are of medium size and could be further labeled for a variety of applications like sub-category detection, meta-data transferring and etc. Promising detection results are obtained on the challenging PASCAL VOC dataset.

关键词： Detectors Training Object detection Visualization Linear programming Integrated circuits Birds

来源：评论

学校读者我要写书评

暂无评论

Cooperative spectrum sensing based on spatially-correlated Rayleigh shadowing

Cooperative spectrum sensing based on spatially-correlated R...

引用

IEEE International Conference on Communications in China (ICCC)

作者： Yong Zhang Zhong Chen Feifei Gao Xianda Zhang State Key Laboratory of Intelligent Technology and Systems Tsinghua National Laboratory for Information Science and Technology Department of Automation Tsinghua University Beijing China

Cooperative spectrum sensing has been proposed to help raise the sensitivity requirement under the fading propagation in Cognitive Radio Network. In this paper, we introduce a new explanation that the best probability of detection or false alarm can be achieved from only parts of the cooperative secondary users under different scenarios. Specifically, we analyze from the new perspective of spatially-correlated shadowing and formulate the probability of detection as a function of the correlation coefficient. The simulation shows that, given a targeted cooperative probability of false alarm, the cooperative probability of detection decreases as correlation increases. The range of an appropriate distance between secondary users is derived based on the accuracy and sensitivity of the Cognitive Radio system. According to the correlation coefficient between two SUs, we can determine the cooperative SUs' set to get better performance of Cognitive Radio Network.

关键词： Sensors Correlation Shadow mapping Signal to noise ratio Cognitive radio Exponential distribution Sensitivity

来源：评论

学校读者我要写书评

暂无评论

Learning similarity metric with SVM

Learning similarity metric with SVM

引用

International Joint Conference on Neural Networks

作者： Xiaoqiang Zhu Pinghua Gong Zengshun Zhao Changshui Zhang State Key Laboratory on Intelligent Technology and Systems Tsinghua National Laboratory for Information Science and Technology (TNList) Department of Automation Tsinghua University College of Information and Electrical Engineering Shandong University of Science and Technology

ISBN: (纸本)9781467314886

In this paper, we show how to learn a good similarity metric for SVM classification. We present a novel approach to simultaneously learn a Mahalanobis similarity metric and an SVM classifier. Different from previous approaches, we optimize the Mahalanobis metric directly for minimizing the SVM classification error. Our formulation generalizes the traditional large margin principle used in standard SVM, that is, we maximize the margin-radius-ratio. The learned similarity metric significantly improves the classification performance of standard SVM. Empirical studies on real datasets show the proposed approach achieves higher or comparable classification accuracies compared with state-of-the-art similarity learning methods.

关键词： Support Vector Network Metric Metronidazole

来源：评论

学校读者我要写书评

暂无评论

Efficient power control algorithms for V-BLAST system with per-antenna power constraints

Efficient power control algorithms for V-BLAST system with p...

引用

IEEE Conference on Wireless Communications and Networking

作者： Kun Wang Xianda Zhang Linglong Dai State Key Laboratory of Intelligent Technology and Systems Tsinghua National Laboratory for Information Science and Technology (TNList) Department of Automation Tsinghua University Beijing China Department of Electronic Engineering Tsinghua University Beijing China

In this paper, three novel efficient power control algorithms are proposed for V-BLAST system with both total power constraint (TPC) and per-antenna power constraints (PAPC). Through analysis of the Karush-Kuhn-Tucker (KKT) conditions of the optimization problem, we find out that the antennas with the optimal power control can be divided into two sets. The optimal allocated power (OAP) of each antenna in the first set is its per-antenna power constraint, and the OAPs of the antennas in the second set can be obtained easily through a new optimization problem with TPC only. Based on the above analysis, three novel efficient power allocation algorithms are developed, which have much lower computational complexities compared with the existing dual update method. Simulation results are presented to demonstrate the efficiency of the proposed power control schemes.

关键词： Power control Antennas Bit error rate Complexity theory Algorithm design and analysis Receivers Resource management

来源：评论

学校读者我要写书评

暂无评论

Shared control teleoperation for targets acquisition

Shared control teleoperation for targets acquisition

引用

Chinese Control and Decision Conference, CCDC

作者： Dongli Wang Huaping Liu Fuchun Sun Yonggui Du Automation college of information engineering Taiyuan university of technology 030024 china Tsinghua national laboratory for information science and technology state key laboratory of intelligent technology and systems department of computer science and technology tsinghua university Beijing 100084 china

It is very difficult to capture the target in the microgravity environment, simply by relying on ground operators. Based on shared control theory, this paper effectively combines the decision-making ability of ground operators and the independent ability of space robot, through making full use of independent control ability of the space robot, and reducing operators' workload and working time, to achieve more valid capture for targets. Shared control is more accurate capture of the target compared with teleoperation in experimental.

关键词： Aerospace electronics Robot kinematics Grippers Control systems Solid modeling Heuristic algorithms

来源：评论

学校读者我要写书评

暂无评论

Optimal sensing based resource allocation in multiuser cognitive radio networks

Optimal sensing based resource allocation in multiuser cogni...

引用

IEEE Conference on Wireless Communications and Networking

作者： Zhong Chen Feifei Gao Xianda Zhang James C. F. Li Ming Lei State Key Laboratory of Intelligent Technology and Systems Tsinghua National Laboratory for Information Science and Technology Department of Automation Tsinghua University Beijing China National Mobile Communications Research Laboratory South East University Nanjing China NEC Laboratories Beijing China

In this paper, spectrum-sensing based resource allocation for a cognitive radio network that contains multiple secondary users (SUs) and a primary user (PU) is investigated. Each SU firstly performs the individual spectrum sensing and forwards the result to a fusion center to make the global decision. If PU is determined absent, then SUs access the primary band with their regular transmit power. Otherwise SUs still access the licensed band but with a limited transmit power to avoid harmful interference to PU. We target at designing the optimal sensing time, bandwidth allocation, power allocation, and the detection threshold to maximize the total achievable rate of SUs subject to the constraints of their peak transmit powers as well as the interference constraint to PU. The original optimization is divided into several subproblems that can be solved separately with low computational complexity. Numerical results demonstrate the superiority of the proposed one over the existing candidates.

关键词： Sensors Resource management Interference Bandwidth Cognitive radio Optimization Joints

来源：评论

学校读者我要写书评

暂无评论

Histogram of the Oriented Gradient for Face Recognition

引用

Tsinghua Science and technology 2011年第2期16卷 216-224页

作者：舒畅丁晓青方驰 State Key Laboratory of Intelligent Technology and System Department of Electronic EngineeringTsinghua University

The histogram of oriented gradient has been successfully applied in many research fields with excellent performance especially in pedestrian detection. However, the method has rarely been applied to face recognition. Aimed to develop a fast and efficient new feature for face recognition, the original HOG and its variations were applied to evaluate the effects of different factors. An information theory-based criterion was also developed to evaluate the potential classification power of different features. Comparative experiments show that even with a relatively simple feature descriptor, the proposed HOG feature achieves almost the same recognition rate with much lower computational time than the widely used Gabor feature on the FRGC and CAS-PEAL databases.

关键词： face recognition feature histogram of oriented gradient

来源：评论

学校读者我要写书评

暂无评论

Multi-Stage Multi-Task Feature Learning* 12

Multi-Stage Multi-Task Feature Learning*

引用

Annual Conference on Neural Information Processing systems

作者： Pinghua Gong Jieping Ye Changshui Zhang State Key Laboratory on Intelligent Technology and Systems Tsinghua National Laboratory for Information Science and Technology (TNList) Department of Automation Tsinghua University Beijing 100084 China Computer Science and Engineering Center for Evolutionary Medicine and Informatics The Biodesign Institute Arizona State University Tempe AZ 85287 USA

ISBN: (纸本)9781627480031

Multi-task sparse feature learning aims to improve the generalization performance by exploiting the shared features among tasks. It has been successfully applied to many applications including computer vision and biomedical informatics. Most of the existing multi-task sparse feature learning algorithms are formulated as a convex sparse regularization problem, which is usually suboptimal, due to its looseness for approximating an ?_0o-type regularizer. In this paper, we propose a non-convex formulation for multi-task sparse feature learning based on a novel regularizer. To solve the non-convex optimization problem, we propose a Multistage Multi-Task Feature Learning (MSMTFL) algorithm. Moreover, we present a detailed theoretical analysis showing that MSMTFL achieves a better parameter estimation error bound than the convex formulation. Empirical studies on both synthetic and real-world data sets demonstrate the effectiveness of MSMTFL in comparison with the state of the art multi-task sparse feature learning algorithms.

关键词： Learning learning algorithms Non-Convex Biomedical informatics Theoretical analysis Parameter estimation images computer vision state of the art Dataset Convexity Religious Missions SYNTHETICS

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：