检索结果-内蒙古大学图书馆

The Journal of Machine Learning Research 2013年第1期14卷

作者： Kevin Murphy Bernhard Schölkopf Pinghua Gong Jieping Ye Changshui Zhang Google MPI for Intelligent Systems State Key Laboratory on Intelligent Technology and Systems Tsinghua National Laboratory for Information Science and Technology Department of Automation Tsinghua University Beijing China Computer Science and Engineering Center for Evolutionary Medicine and Informatics The Biodesign Institute Arizona State University Tempe AZ

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 l0-type regularizer. In this paper, we propose a non-convex formulation for multi-task sparse feature learning based on a novel nonconvex regularizer. To solve the non-convex optimization problem, we propose a Multi-Stage Multi-Task Feature Learning (MSMTFL) algorithm; we also provide intuitive interpretations, detailed convergence and reproducibility analysis for the proposed 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： multi-stage multi-task learning non-convex sparse learning

来源：评论

学校读者我要写书评

暂无评论

Hydrogen generation for sensing and control in submerged friction stir welding

Hydrogen generation for sensing and control in submerged fri...

引用

作者： Fleming, P. Bloodworth, T. Lammlein, D. Strauss, A. Cook, G. Department of Electrical Engineering and Computer Science Vanderbilt University Nashville TN United States Department of Mechanical Engineering Vanderbilt University Nashville TN United States Welding Automation Laboratory Department of Mechanical Engineering Vanderbilt University 2201 West End Avenue Nashville TN 37235 United States

Friction stir welding (FSW) is a recently developed welding technique, where the metals that are welded are not melted. FSW takes place at temperatures below the solidus points of the metals that are joined. FSW is thus a solid-state process. This can yield weldments of a significantly higher strength than those achieved by traditional fusion welding processes. This is due to the enhanced grain structure of the FSW materials. Submerged FSW is performed underwater for structural purposes such as oil platforms and pipelines, or in industrial settings and laboratories in order to further enhance the strength of the weldments, sometimes exceeding the strength of the parent materials. Submerged FSW, due to the significantly lower temperatures of the process compared to FSW in air, yields even more enhanced grain structures and greater strengths. In the experiments performed here, it is shown that the hydrogen generated during submerged FSW is produced at rates and quantities sufficient to be easily detected by means of a fuel cell. The generated hydrogen, which is a function of the FSW parameters, can be used as a means of sensing and control of submerged FSW which can be applied to robotic and automatic submerged FSW. © Authors 2011.

关键词： Friction stir welding

来源：评论

学校读者我要写书评

暂无评论

Task-based grasp adaptation on a humanoid robot

Task-based grasp adaptation on a humanoid robot

引用

10th IFAC Symposium on Robot Control, SYROCO 2012

作者： Bohg, Jeannette Welke, Kai León, Beatriz Do, Martin Song, Dan Wohlkinger, Walter Madry, Marianna Aldóma, Aitor Przybylski, Markus Asfour, Tamim Martí, Higinio Kragic, Danica Morales, Antonio Vincze, Markus Humanoids and Intelligence Systems Lab. Institute for Anthropomatics KIT Germany Autonomous Motion Lab. MPI for Intelligent Systems Germany Robotic Intelligence Laboratory Department of Computer Science and Engineering UJI Spain Computer Vision and Active Perception Lab. KTH Sweden Automation and Control Institute TUW Austria

ISBN: (纸本)9783902823113

In this paper, we present an approach towards autonomous grasping of objects according to their category and a given task. Recent advances in the field of object segmentation and categorization as well as task-based grasp inference have been leveraged by integrating them into one pipeline. This allows us to transfer task-specific grasp experience between objects of the same category. The effectiveness of the approach is demonstrated on the humanoid robot ARMAR-IIIa. © 2012 IFAC.

关键词： Anthropomorphic robots

来源：评论

学校读者我要写书评

暂无评论

Resource-distortion modeling for video streaming over mesh networks with priority-based packet scheduling

Resource-distortion modeling for video streaming over mesh n...

引用

2012 13th IEEE International Conference on Multimedia and Expo, ICME 2012

作者： Zhang, Yongfei Zhang, Yunsheng Qin, Shiyin He, Zhihai Beijing Key Lab. of Digital Media School of Computer Science and Engineering Beihang University Beijing 100191 China State Key Laboratory of Virtual Reality Technology and Systems Beihang University Beijing 100191 China School of Automation Science and Electrical Engineering Beihang University Beijing 100191 China Graduate School at Shenzhen Tsinghua University Shenzhen 518055 China Department of Electrical and Computer Engineering University of Missouri Columbia MO 65211 United States

ISBN: (纸本)9781467316590

Video streaming over mesh networks operates under stringent network resource constraints, with a large number of video sessions competing for limited network resources. In this work, we aim to establish so-called resource-distortion models to characterize the inherent relationship between the allocated network resources to each video session and its end-to-end video distortion or presentation quality at the receiver end. We observed that priority-based packet scheduling has significant impact on such resource-distortion relationship. Using ANN-based learning methods, we develop an end-to-end packet delay bound violation (PDBV) probability model for video streaming over multi-hop networks with priority-based packet scheduling. We then derive a quadratic video transmission distortion model to capture the unique behavior of priority-based packet scheduling and its impact on the end-to-end video distortion. Based on these resource-distortion models, we are able to predict the end-to-end distortion for video streaming over multi-hop networks with priority-based packet scheduling. Our extensive experimental results demonstrate that the proposed method is very accurate. © 2012 IEEE.

关键词： Video streaming

来源：评论

学校读者我要写书评

暂无评论

Towards automated task planning for service robots using semantic knowledge representation

Towards automated task planning for service robots using sem...

引用

IEEE International Conference on Industrial Informatics (INDIN)

作者： Ze Ji Renxi Qiu Alex Noyvirt Anthony Soroka Michael Packianather Rossi Setchi Dayou Li Shuo Xu School of Engineering Cardiff University UK School of Computer Science Bedfordshire University UK Shanghai Key Laboratory of Manufacturing Automation and Robotics School of Mechatronic Engineering and Automation Shanghai University China

ISBN: (纸本)9781467303125

Automated task planning for service robots faces great challenges in handling dynamic domestic environments. Classical methods in the Artificial Intelligence (AI) area mostly focus on relatively structured environments with fewer uncertainties. This work proposes a method to combine semantic knowledge representation with classical approaches in AI to build a flexible framework that can assist service robots in task planning at the high symbolic level. A semantic knowledge ontology is constructed for representing two main types of information: environmental description and robot primitive actions. Environmental knowledge is used to handle spatial uncertainties of particular objects. Primitive actions, which the robot can execute, are constructed based on a STRIPS-style structure, allowing a feasible solution (an action sequence) for a particular task to be created. With the Care-O-Bot (CoB) robot as the platform, we explain this work with a simple, but still challenging, scenario named “get a milk box”. A recursive back-trace search algorithm is introduced for task planning, where three main components are involved, namely primitive actions, world states, and mental actions. The feasibility of the work is demonstrated with the CoB in a simulated environment.

关键词： Planning Robot kinematics Ontologies Robot sensing systems Semantics Dairy products

来源：评论

学校读者我要写书评

暂无评论

Validation of an odor source identification algorithm via an underwater vehicle

Validation of an odor source identification algorithm via an...

引用

2nd International Conference on Intelligent Systems Design and engineering Applications, ISDEA 2012

作者： Kang, Xiaodong Li, Wei Xu, Hongli Feng, Xisheng Li, Yiping State Key Laboratory of Robotics Shenyang Institute of Automation Chinese Academy of Sciences Liaoning Province 110016 China Graduate School of the Chinese Academy of Sciences Beijing 100049 China Department of Computer Engineering and Science California State University Bakersfield CA 93311 United States

ISBN: (纸本)9780769546087

This paper systemically discusses a procedure for evaluating and validating an odor source identification algorithm derived from moth-inspired plume tracing strategies. We evaluate and validate the proposed source identification algorithm as follows. First, we identify the source of a virtual chemical plume with significant filament intermittency and meander via a simulated underwater vehicle. Second, we validate the source identification algorithm using the virtual chemical plume and a real underwater vehicle in a swimming pool. Finally, we run in-water tests to trace a Rhoda mine dye plume developed in near-shore ocean environments characterized by turbulence, tides and waves, and to identify its source via the underwater vehicle at Da Lian Bay in China. © 2012 IEEE.

关键词： Vehicles

来源：评论

学校读者我要写书评

暂无评论

RADD: A Replicas Adaptive Data Delivery Scheme in DTMSN

RADD: A Replicas Adaptive Data Delivery Scheme in DTMSN

引用

International Conference on Mobile Ad-hoc and Sensor Networks, MSN

作者： Yong Feng Feng Wang Qian Qian Bo Liang Yunnan Key Laboratory of Computer Technology Applications School of Information Engineering and Automation Kunming University of Science and Technology Kunming Yunnan China

ISBN: (纸本)9781467358088

Due to the intermitted connectivity feature, it is rational to use multi-replica approaches so as to enhance the data transmission performance in DTMSN. But the majority of off-the-shelf multi-replica approaches usually generate too many message copies in the network, which may quickly exhaust the limited network resource of DTMSN. Therefore, an efficient and effective data delivery scheme needs to judge and weigh the benefits and costs due to the number of replica of messages. In this paper, we propose a new Replicas Adaptive Data Delivery protocol (RADD), which makes use of a self-adapting algorithm to reduce the number of redundant messages, and attains a good network performance with low network resource expenditure. Simulation results indicate that our proposed RADD reaches comparable or higher data delivery performance at the cost of the lower transmission overhead compared to several existing multi-replica schemes in DTMSN.

关键词： DTMSN Data gathering scheme Replicas adaptive Delivery probability REPLICA Web resources Transmission Characteristics Obstetric Delivery Messages Rational number Self tuning Scheme

来源：评论

学校读者我要写书评

暂无评论

Stability and transparency improvement in haptic device employing both MR-brake and active actuator

Stability and transparency improvement in haptic device empl...

引用

2012 21st IEEE International Symposium on Robot and Human Interactive Communication, RO-MAN 2012

作者： Baser, Ozgur Konukseven, E. Ilhan Gurocak, Hakan Mechanical Engineering Department Middle East Technical University Universiteler Mah. Dumlupinar Blv. No:1 06800 Çankaya Ankara Turkey Robotics and Automation Laboratory Washington State University Vancouver WA United States School of Engineering and Computer Science Washington State University Vancouver WA 98686 United States

ISBN: (纸本)9781467346054

An ideal haptic device should transmit a wide range of stable virtual model impedances (Z-width) with high transparency. Magneto-rheological fluid (MR) brakes are advantageous in haptic devices since they are passive actuators. However, they cannot provide high transparency and smooth interaction due to high viscous friction, residual torque, slow response, sticking and hysteresis effects. On the other hand, active actuators cannot simulate high virtual impedances stably, but provide high transparency with a closed loop control algorithm. In the proposed hybrid actuation a task divider control (TDC) algorithm was developed for torque sharing between two actuators to provide a large Z-width and improve both transparency and smoothness. The algorithm employs two parameters which were estimated experimentally and extended to entire achievable impedance range by artificial neural network (ANN) and curve fitting techniques. A 1-DOF device having an excitation motor at the user side and brushless DC motor and MR-brake in the haptic side was used in the experiments. The excitation motor is used to generate a white noise torque input to simulate a user for frequency domain transparency tests. Results of the proposed and conventional closed loop impedance control (CLIC) algorithms were compared. The proposed algorithm improves the transparency of MR-brake by eliminating its drawbacks and presents a larger Z-width than the active actuator alone. © 2012 IEEE.

关键词： Hysteresis

来源：评论

学校读者我要写书评

暂无评论

A Dynamic Pseudonyms Based Anonymous Routing Protocol for Wireless Ad Hoc Networks

A Dynamic Pseudonyms Based Anonymous Routing Protocol for Wi...

引用

International Conference on Mobile Ad-hoc and Sensor Networks, MSN

作者： Ting Li Yong Feng Feng Wang Xiaodong Fu Yunnan Key Laboratory of Computer Technology Applications School of Information Engineering and Automation Kunming University of Science and Technology Kunming Yunnan China

ISBN: (纸本)9781467358088

In wireless ad hoc networks, it is an efficient approach to utilize the dynamic pseudonyms scheme to preserve the nodes' privacy. However, in the existing works, there are some problems such as weak anonymity and synchronization difficulty when updating pseudonyms. In this paper, we propose a lightweight Dynamic Pseudonyms based Anonymous Routing protocol, called DPAR. In the proposed DPAR, the security is implemented by lightweight symmetric key cryptography and hashing operations, the pseudonyms are generated by hashing operations, and dynamically updated and synchronized by elaborate message interactions. Extensive analyses show that DPAR can achieve the routing security and the nodes' anonymity with acceptable overhead.

关键词： Wireless ad hoc networks Secure routing protocol Anonymous routing Dynamic pseudonyms Mobile ad hoc networks Pseudonyms Secure Routing Protocol wireless Routing protocols

来源：评论

学校读者我要写书评

暂无评论

Online actor-critic learning for motion control of non-holonomic mobile robot

Online actor-critic learning for motion control of non-holon...

引用

2011 International Conference on Electrical, Information engineering and Mechatronics, EIEM 2011

作者： Wang, Kai Jia, Yingmin Du, Junping Yu, Fashan Beijing 100191 China Beijing 100191 China Beijing Key Laboratory of Intelligent Telecommunications Software and Multimedia School of Computer Science and Technology Beijing University of Posts and Telecommunications Beijing 100876 China School of Electrical Engineering and Automation Henan Polytechnic University Jiaozuo 454000 Henan China

ISBN: (纸本)9781447124665

This paper presents a control structure designed for non-holonomic mobile robots by an online algorithm based on policy iteration for learning the continuous-time (CT) optimal control solution with infinite horizon cost. The algorithm learns online in real-time to the solution of Hamilton-Jacobi-Bellman (HJB) equation which has been used for optimal control design. This method finds in real-time suitable approximations of both the optimal cost and control policy, while also guaranteeing closed-loop stability, which implemented as an actor/critic structure involves simultaneous continuous-time adaptation of both actor and critic neural networks (NNs). Simulation examples show the effectiveness of the new algorithm. © 2012 Springer-Verlag London Limited.

关键词： Iterative methods

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：