------------------------------------------------------------ CALL FOR PAPERS
Abstraction in Reinforcement Learning Workshop at ICML/UAI/COLT 2009 June 18, 2009, Montreal, Canada Submission deadline: April 15, 2009 http://www-all.cs.umass.edu/~gdk/arl ------------------------------------------------------------ Although reinforcement learning methods have been effectively applied to a number of problems of practical importance, successful large-scale applications remain the exception rather than the norm. Problems with large state spaces still pose considerable challenges to existing algorithms. Abstraction is the process of factoring out irrelevant details, in other words, of focusing only on the information that is relevant for a particular purpose. For a number of years, the research community has been exploring various forms of abstraction as potential mechanisms for scaling up reinforcement learning algorithms to large, complex problems. State abstraction approaches and temporal abstraction methods have become well established, while recent representation-discovery methods have shown a great deal of promise. The goal of this workshop is to promote interaction between researchers that work on various forms of abstraction in reinforcement learning, to explore possible areas of synergy between existing approaches, and to open up discussion on novel techniques that can harness the existing strengths of different types of abstractions. We welcome submissions on all aspects of abstraction in Reinforcement Learning, including, but not limited to, papers addressing the following topics: Representation: Novel representational frameworks for temporal and state abstraction; experiences with existing frameworks. Discovery: Methods that allow artificial agents to perform state and temporal abstraction autonomously, using their experience in their environment. Algorithms: Learning and planning algorithms that can fully take advantage of temporal abstraction by reasoning at the correct temporal granularity in the presence of actions at different time scales. Applications: Descriptions of real-world applications that make effective use of various abstraction methods; suggestions for real-world and simulated environments that can support ongoing research in the area. Synergy: Methods that use one type of abstraction to discover or improve the performance of another type of abstraction. Overview/Methodology: Reviews of existing methods; comments on methodology; new research directions. Submissions will be reviewed by the program committee on the basis of relevance, significance, technical quality, and clarity. SUBMISSION INSTRUCTIONS Please email your submissions to a...@cs.umass.edu by April 15, 2009. Submissions should be in pdf format and follow ICML 2009 formatting guidelines. They should not exceed 6 pages in length. IMPORTANT DATES Submissions due: April 15, 2009 Author notification: May 8, 2009 Revised papers due: June 1, 2009 Workshop date: June 18, 2009 ORGANIZERS Ozgur Simsek, Max Planck Institute for Human Development George Konidaris, University of Massachusetts Amherst PROGRAM COMMITTEE Andrew Barto, University of Massachusetts Amherst Carlos Diuk, Rutgers University Alan Fern, Oregon State University Nick Jong, University of Texas Austin Anders Jonsson, Universitat Pompeu Fabra Ben Kuipers, University of Michigan Lihong Li, Rutgers University Shie Mannor, McGill University Yael Niv, Princeton University Sarah Osentoski, University of Massachusetts Amherst Ron Parr, Duke University Jan Peters, Max Planck Institute for Biological Cybernetics Doina Precup, McGill University Balaraman Ravindran, Indian Institute of Technology Madras Stuart Russell, University of California Berkeley Erik Talvitie, University of Michigan Tom Walsh, Rutgers University Alicia P. Wolfe, University of Massachusetts Amherst A few more PC invitations are pending. _______________________________________________ uai mailing list uai@ENGR.ORST.EDU https://secure.engr.oregonstate.edu/mailman/listinfo/uai