检索结果-内蒙古大学图书馆

A parallel scheduling algorithm for reinforcement learning in large state space

Frontiers of computer science 2012年第6期6卷 631-646页

作者： Quan LIU Xudong YANG Ling JING Jin LI Jiao LI Institute of Computer Science and Technology Soochow UniversitySuzhou 215006China Department of Computer Science and Technology Nanjing UniversityNanjing 210093China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin UniversityChangchun 130012China

The main challenge in the area of reinforcement learning is scaling up to larger and more complex problems. Aiming at the scaling problem of reinforcement learning, a scalable reinforcement learning method, DCS-SRL, is proposed on the basis of divide-and-conquer strategy, and its convergence is proved. In this method, the learning problem in large state space or continuous state space is decomposed into multiple smaller subproblems. Given a specific learning algorithm, each subproblem can be solved independently with limited available resources. In the end, component solutions can be recombined to obtain the desired result. To ad- dress the question of prioritizing subproblems in the scheduler, a weighted priority scheduling algorithm is proposed. This scheduling algorithm ensures that computation is focused on regions of the problem space which are expected to be maximally productive. To expedite the learning process, a new parallel method, called DCS-SPRL, is derived from combining DCS-SRL with a parallel scheduling architecture. In the DCS-SPRL method, the subproblems will be distributed among processors that have the capacity to work in parallel. The experimental results show that learning based on DCS-SPRL has fast convergence speed and good scalability.

关键词： divide-and-conquer strategy parallel schedule scalability large state space continuous state space

来源：评论

学校读者我要写书评

暂无评论

Proceedings of the 8th Annual Conference on the science of Dissemination and Implementation : Washington, DC, USA. 14-15 December 2015

引用

Implementation science : IS 2016年第2期11 Suppl 2卷 100页

作者： Chambers, David Simpson, Lisa Hill-Briggs, Felicia Neta, Gila Vinson, Cynthia Beidas, Rinad Marcus, Steven Aarons, Gregory Hoagwood, Kimberly Schoenwald, Sonja Evans, Arthur Hurford, Matthew Rubin, Ronnie Hadley, Trevor Barg, Frances Walsh, Lucia Adams, Danielle Mandell, David Martin, Lindsey Mignogna, Joseph Mott, Juliette Hundt, Natalie Kauth, Michael Kunik, Mark Naik, Aanand Cully, Jeffrey McGuire, Alan White, Dominique Bartholomew, Tom McGrew, John Luther, Lauren Rollins, Angie Salyers, Michelle Cooper, Brittany Funaiole, Angie Richards, Julie Lee, Amy Lapham, Gwen Caldeiro, Ryan Lozano, Paula Gildred, Tory Achtmeyer, Carol Ludman, Evette Addis, Megan Marx, Larry Bradley, Katharine VanDeinse, Tonya Wilson, Amy Blank Stacey, Burgin Powell, Byron Bunger, Alicia Cuddeback, Gary Barnett, Miya Stadnick, Nicole Brookman-Frazee, Lauren Lau, Anna Dorsey, Shannon Pullmann, Michael Mitchell, Shannon Schwartz, Robert Kirk, Arethusa Dusek, Kristi Oros, Marla Hosler, Colleen Gryczynski, Jan Barbosa, Carolina Dunlap, Laura Lounsbury, David O’Grady, Kevin Brown, Barry Damschroder, Laura Waltz, Thomas Ritchie, Mona Atkins, David Imel, Zac E. Xiao, Bo Can, Doğan Georgiou, Panayiotis Narayanan, Shrikanth Berkel, Cady Gallo, Carlos Sandler, Irwin Brown, C. Hendricks Wolchik, Sharlene Mauricio, Anne Marie Mehrotra, Sanjay Chandurkar, Dharmendra Bora, Siddhartha Das, Arup Tripathi, Anand Saggurti, Niranjan Raj, Anita Hughes, Eric Jacobs, Brian Kirkendall, Eric Loeb, Danielle Trinkley, Katy Yang, Michael Sprowell, Andrew Nease, Donald Lyon, Aaron Lewis, Cara Boyd, Meredith Melvin, Abigail Nicodimos, Semret Liu, Freda Jungbluth, Nathanial Flynn, Allen Landis-Lewis, Zach Sales, Anne Baloh, Jure Ward, Marcia Zhu, Xi Bennett, Ian Unutzer, Jurgen Mao, Johnny Proctor, Enola Vredevoogd, Mindy Chan, Ya-Fen Williams, Nathaniel Green, Phillip Bernstein, Steven Rosner, June-Marie DeWitt, Michelle Tetrault, Jeanette Dziura, James Hsiao, Allen Sussman, Scott O’Connor, Patrick Toll, Benjamin Jones, Michael Gassaway, Julie Tobin, Jonathan Zatzic Division of Cancer Control and Population Sciences National Cancer Institute National Institutes of Health Rockville USA AcademyHealth Washington USA Department of Medicine and Welch Center for Prevention Epidemiology and Clinical Research Johns Hopkins Medical Institutions Baltimore USA Psychiatry University of Pennsylvania Philadelphia USA Family Medicine and Community Health University of Pennsylvania Philadelphia USA School of Social Policy and Practice University of Pennsylvania Philadelphia USA Psychiatry UC San Diego La Jolla USA Child & Adolescent Psychiatry The Child Study Center at NYU Langone Medical Center New York NY New York USA Psychiatry and Behavioral Sciences MUSC Charleston USA DBHIDS City of Philadelphia Philadelphia USA CBH City of Philadelphia Philadelphia USA Psychiatry University of Pennsylvania Perelman School of Medicine Philadelphia USA Health Services Research & Development Department of Veterans Affairs & Baylor College of Medicine Houston USA Treatment Core Department of Veterans Affairs Waco USA National Center for PTSD Executive Division Department of Veterans Affairs White River Junction USA HSR&D HSR&D Center for Health Information and Communication Roudebush VA Medical Center Indianapolis USA Psychology Indiana University Purdue University Indianapolis Indianapolis USA Department of Psychiatric Rehabilitation and Counseling Rutgers University Newark USA Health Services Research & Development Richard L. Roudebush VAMC Indianapolis USA IUPUI Department of Psychology ACT Center of Indiana Indianapolis USA Human Development Washington State University Pullman USA Prevention Science Washington State University Pullman USA Group Health Research Institute Group Health Seattle USA Behavioral Health Services Group Health Seattle USA Preventive Care Group Health Seattle USA Health Services Research & Development Primary and Specialty Medical Care Service VA Puget Sound Seattle USA School of Social Work Un

A1 Introduction to the 8th Annual Conference on the science of Dissemination and Implementation: Optimizing Personal and Population Health David Chambers, Lisa Simpson D1 Discussion forum: Population health D&I research Felicia Hill-Briggs D2 Discussion forum: Global health D&I research Gila Neta, Cynthia Vinson D3 Discussion forum: Precision medicine and D&I research David Chambers S1 Predictors of community therapists’ use of therapy techniques in a large public mental health system Rinad Beidas, Steven Marcus, Gregory Aarons, Kimberly Hoagwood, Sonja Schoenwald, Arthur Evans, Matthew Hurford, Ronnie Rubin, Trevor Hadley, Frances Barg, Lucia Walsh, Danielle Adams, David Mandell S2 Implementing brief cognitive behavioral therapy (CBT) in primary care: Clinicians' experiences from the field Lindsey Martin, Joseph Mignogna, Juliette Mott, Natalie Hundt, Michael Kauth, Mark Kunik, Aanand Naik, Jeffrey Cully S3 Clinician competence: Natural variation, factors affecting, and effect on patient outcomes Alan McGuire, Dominique White, Tom Bartholomew, John McGrew, Lauren Luther, Angie Rollins, Michelle Salyers S4 Exploring the multifaceted nature of sustainability in community-based prevention: A mixed-method approach Brittany Cooper, Angie Funaiole S5 Theory informed behavioral health integration in primary care: Mixed methods evaluation of the implementation of routine depression and alcohol screening and assessment Julie Richards, Amy Lee, Gwen Lapham, Ryan Caldeiro, Paula Lozano, Tory Gildred, Carol Achtmeyer, Evette Ludman, Megan Addis, Larry Marx, Katharine Bradley S6 Enhancing the evidence for specialty mental health probation through a hybrid efficacy and implementation study Tonya VanDeinse, Amy Blank Wilson, Burgin Stacey, Byron Powell, Alicia Bunger, Gary Cuddeback S7 Personalizing evidence-based child mental health care within a fiscally mandated policy reform Miya Barnett, Nicole Stadnick, Lauren Brookman-Frazee, Anna Lau S8 Leveraging an existing

关键词： Veteran Affair Motivational Interview Qualitative Comparative Analysis Community That Care Supplemental Nutrition Assistance Program

来源：评论

学校读者我要写书评

暂无评论

Building linked open university data: tsinghua university open data as a showcase

Building linked open university data: tsinghua university op...

引用

Joint International Semantic Technology Conference, JIST 2011

作者： Ma, Yuanchao Xu, Bin Bai, Yin Li, Zonghui Knowledge Engineering Group Department of Computer Science and Technology Tsinghua University Beijing China

ISBN: (纸本)9783642299223

Linked Open University Data applies semantic web and linked data technology to university data scenario, aiming at building interlinked semantic data around university information, providing possibility for unified inner- and inter- school information query and comparison. This paper proposes a general process of building linked open university data, with procedures covering choosing datasets and vocabularies, collecting and processing data, building RDF and interlink, etc. Tsinghua University Open Data is used to demonstrate the process. Tsinghua University consist of 5 well-formed, interconnected datasets, with a number of interesting applications has been built on top of them. Finally, remarkable points about data collecting and processing is discussed. © 2012 Springer-Verlag.

关键词： Linked data

来源：评论

学校读者我要写书评

暂无评论

ExerLink: Enabling pervasive social exergames with heterogeneous exercise devices 12

ExerLink: Enabling pervasive social exergames with heterogen...

引用

10th International Conference on Mobile Systems, Applications, and Services, MobiSys'12

作者： Park, Taiwoo Hwang, Inseok Lee, Uichin Lee, Sunghoon Ivan Yoo, Chungkuk Lee, Youngki Jang, Hyukjae Choe, Sungwon Peter Park, Souneil Song, Junehwa Department of Computer Science KAIST Korea Republic of Department of Knowledge Service Engineering KAIST Korea Republic of Computer Science Department UCLA United States

ISBN: (纸本)9781450313018

We envision that diverse social exercising games, or exergames, will emerge, featuring much richer interactivity with immersive game play experiences. Further, the recent advances of mobile devices and wireless networking will make such social engagement more pervasive - people carry portable exergame devices (e.g., jump ropes) and interact with remote users anytime, anywhere. Towards this goal, we explore the potential of using heterogeneous exercise devices as game controllers for a multi-player social exergame;e.g., playing a boat paddling game with two remote exercisers (one with a jump rope, and the other with a treadmill). In this paper, we propose a novel platform called ExerLink that converts exercise intensity to game inputs and intelligently balances intensity/delay variations for fair game play experiences. We report the design considerations and guidelines obtained from the design and development processes of game controllers. We validate the efficacy of game controllers and demonstrate the feasibility of social exergames with heterogeneous exercise devices via extensive human subject studies. © 2012 ACM.

关键词： Rope

来源：评论

学校读者我要写书评

暂无评论

Reinforcement learning transfer via common subspaces

Reinforcement learning transfer via common subspaces

引用

2011 Adaptive and Learning Agents Workshop, ALA 2011, Held at Autonomous Agents and Multi-Agent Systems Conference, AAMAS 2011

作者： Ammar, Haitham Bou Taylor, Matthew E. Department of Knowledge Engineering Maastricht University Netherlands Department of Computer Science Lafayette College United States

ISBN: (纸本)9783642284984

Agents in reinforcement learning tasks may learn slowly in large or complex tasks - transfer learning is one technique to speed up learning by providing an informative prior. How to best enable transfer between tasks with different state representations and/or actions is currently an open question. This paper introduces the concept of a common task subspace, which is used to autonomously learn how two tasks are related. Experiments in two different nonlinear domains empirically show that a learned inter-state mapping can successfully be used by fitted value iteration, to (1) improving the performance of a policy learned with a fixed number of samples, and (2) reducing the time required to converge to a (near-) optimal policy with unlimited samples. © 2012 Springer-Verlag Berlin Heidelberg.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Reinforcement learning transfer using a sparse coded inter-task mapping

Reinforcement learning transfer using a sparse coded inter-t...

引用

9th European Workshop on Multi-Agent Systems, EUMAS 2011

作者： Ammar, Haitham Bou Taylor, Matthew E. Tuyls, Karl Weiss, Gerhard Department of Knowledge Engineering Maastricht University Netherlands Department of Computer Science Lafayette College United States

ISBN: (纸本)9783642347986

Reinforcement learning agents can successfully learn in a variety of difficult tasks. A fundamental problem is that they may learn slowly in complex environments, inspiring the development of speedup methods such as transfer learning. Transfer improves learning by reusing learned behaviors in similar tasks, usually via an inter-task mapping, which defines how a pair of tasks are related. This paper proposes a novel transfer learning technique to autonomously construct an inter-task mapping by using a novel combinations of sparse coding, sparse projection learning, and sparse pseudo-input gaussian processes. Experiments show successful transfer of information between two very different domains: the mountain car and the pole swing-up task. This paper empirically shows that the learned inter-task mapping can be used to successfully (1) improve the performance of a learned policy on a fixed number of samples, (2) reduce the learning times needed by the algorithms to converge to a policy on a fixed number of samples, and (3) converge faster to a near-optimal policy given a large amount of samples. © 2012 Springer-Verlag Berlin Heidelberg.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A Genetic Algorithm (GA) based Load Balancing Strategy for Cloud Computing

引用

Procedia Technology 2013年 10卷 340-347页

作者： Kousik Dasgupta Brototi Mandal Paramartha Dutta Jyotsna Kumar Mandal Santanu Dam Department of Computer Science and Engineering Kalyani Government Engineering College Kalyani-741 235 India Deptt. of CSE Supreme Knowledge Foundation Group of Inst. Mankundu-712 139 India Deptt. of CSE Kalyani University Kalyani-741 235 India Deptt. of CSS Visva-Bharati University Santiniketan-731 235 India Deptt. of CSE Future Institute of Engineering and Management Kolkata-700 150 India

The next-generation of cloud computing will thrive on how effectively the infrastructure are instantiated and available resources utilized dynamically. Load balancing which is one of the main challenges in Cloud computing, distributes the dynamic workload across multiple nodes to ensure that no single resource is either overwhelmed or underutilized. This can be considered as an optimization problem and a good load balancer should adapt its strategy to the changing environment and the types of tasks. This paper proposes a novel load balancing strategy using Genetic Algorithm (GA). The algorithm thrives to balance the load of the cloud infrastructure while trying minimizing the make span of a given tasks set. The proposed load balancing strategy has been simulated using the CloudAnalyst simulator. Simulation results for a typical sample application shows that the proposed algorithm outperformed the existing approaches like First Come First Serve (FCFS), Round Robing (RR) and a local search algorithm Stochastic Hill Climbing (SHC).

关键词： Cloud Computing Load balancing Genetic Algorithm

来源：评论

学校读者我要写书评

暂无评论

Reinforcement learning transfer via sparse coding

Reinforcement learning transfer via sparse coding

引用

24th Benelux Conference on Artificial Intelligence, BNAIC 2012

作者： Ammar, Haitham Bou Tuyls, Karl Taylor, Matthew E. Driessens, Kurt Weiss, Gerhard Department of Knowledge Engineering Maastricht University Netherlands Departement of Computer Science Lafayette College United States

来源：评论

学校读者我要写书评

暂无评论

A semi-supervised learning framework to cluster mixed data types

A semi-supervised learning framework to cluster mixed data t...

引用

4th International Conference on knowledge Discovery and Information Retrieval, KDIR 2012

作者： Abdullin, Artur Nasraoui, Olfa Department of Computer Engineering and Computer Science Knowledge Discovery and Web Mining Lab University of Louisville Louisville KY United States

ISBN: (纸本)9789898565297

We propose a semi-supervised framework to handle diverse data formats or data with mixed-type attributes. Our preliminary results in clustering data with mixed numerical and categorical attributes show that the proposed semi-supervised framework gives better clustering results in the categorical domain. Thus the seeds obtained from clustering the numerical domain give an additional knowledge to the categorical clustering algorithm. Additional results show that our approach has the potential to outperform clustering either domain on its own or clustering both domains after converting them to the same target domain. Copyright © 2012 SciTePress - science and Technology Publications.

关键词： Clustering algorithms

来源：评论

学校读者我要写书评

暂无评论

Stream-dashboard: A framework for mining, tracking and validating clusters in a data stream

Stream-dashboard: A framework for mining, tracking and valid...

引用

1st International Workshop on Big Data, Streams and Heterogeneous Source Mining: Algorithms, Systems, Programming Models and Applications, BigMine-12 - Held in Conjunction with SIGKDD Conference

作者： Hawwash, Basheer Nasraoui, Olfa Department of Computer Engineering and Computer Science Knowledge Discovery and Web Mining Lab University of Louisville Louisville KY United States

ISBN: (纸本)9781450315470

Clustering data streams is a challenging problem that has received significant attention in the recent decade. In this paper, we address the hitherto inadequately addressed challenge of managing the output of stream clustering. This task comprises the continuous cluster model validation, monitoring, trend and change detection, and summarization of the cluster mining output. For this purpose, we propose a complete framework to mine, track and validate the clusters in a data stream. The proposed framework keeps track of each discovered cluster model of the stream through time, while quantifying, modeling and summarizing the cluster evolution trends, and storing a summary of the cluster models and the evolution trends only at milestones corresponding to the occurrence of significant changes. Our experiments demonstrate the accuracy of the proposed framework in tracking cluster evolution over time, while generating a concise summary of the evolution trends over the lifetime of the stream. Copyright © 2012 ACM.

关键词： Data mining

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：