CFPPrevious | item21 | NextDate: Fri, 24 Mar 2000 13:21:20 -0800 (PST) From: Byung-Hoon Park bhpark@eecs.wsu.edu Subject: KDD-2000 Workshop on Distributed and Parallel Knowledge Discovery, Boston, MA, Aug 20 2000 Sixth ACM SIGKDD International Conference on Knowledge Discovery & Data Mining August 20, 2000, Boston, MA, USA Workshop Web Site: http://www.eecs.wsu.edu/~hillol/DKD/dpkd2000.html Chairs: Hillol Kargupta http://www.eecs.wsu.edu/~hillol hillol@eecs.wsu.edu Joydeep Ghosh http://hercules.ece.utexas.edu/~ghosh ghosh@ece.utexas.edu Philip Chan http://www.cs.fit.edu/~pkc pkc@cs.fit.edu Vipin Kumar http://www-users.cs.umn.edu/~kumar kumar@cs.umn.edu Zoran Obradovic http://www.eecs.wsu.edu/~zoran zoran@eecs.wsu.edu Background Information Information is most useful when it can be transformed into actionable knowledge. Since the volume of data in most non-trivial processes is typically large, the role of automated data analysis in modern information processing applications is becoming increasingly important. The emerging development of large, data rich distributed systems like the internet and intranets has added another dimension to the problem---a plethora of distributed information resources. Although the communication bandwidth is increasing, by no means is it increasing at a rate that is even close to the increase of available information. As a result, downloading large data sets to a single site over limited bandwidth channels, followed by the application of centralized data analysis algorithms may not be scalable for the large, distributed data analysis applications of the future. Moreover, this may not even be feasible because of security/privacy concerns, or incompatibilities at the system or information representation levels. Examples include information retrieval in a mobile computing environment, situation monitoring using large sensor networks, distributed real-time advanced practice of telemedicine, and e-commerce applications. We need to rethink our fundamental approach toward data analysis in these emerging new environments. Distributed knowledge discovery (DKD) accepts the fact that data may be inherently distributed among different loosely coupled sites connected by a network and the sites may have heterogeneous data. It offers techniques to discover new knowledge through distributed data analysis and modeling using minimal communication of information. Design and development of distributed data analysis algorithms, architectures and their scalability, efficiency, ability to work with different kinds of data and computing platforms such as handheld devices to web server farms, security, human-computer interaction are some of the open research problems in DKD. The first workshop on Distributed Data Mining, held at the Fourth International Conference on Knowledge Discovery and Data Mining (1998) brought the interested researchers and practitioners together and created an environment for crystallizing the fast growing field of DKD. The workshop was focused on the state-of-the-art DKD algorithms, systems, and application related issues. Approximately 40 participants attended the workshop. The workshop had 13 presentations, including 3 invited talks. It was sponsored by the Intel Corporation and Magnify Inc. It resulted in a book (edited by Hillol Kargupta and Philip Chan) to be published by AAAI/MIT Press in early Summer of 2000. With the rapid growth in DKD technology both in terms of research issues and real-world applications, we expect that the second workshop will be even more successful. Further details about the (1998) workshop can be found at http://www.eecs.wsu.edu/~hillol/ddm.html. Today after about two years, this field has gained a large boost in momentum. The number of researchers and practitioners working in this area has nearly tripled. More research issues are introduced and the field is starting to crystallize its shape. We believe it is now the time for hosting another workshop to discuss the state of the art research and practice. This workshop will also extend its scope by including complementary research in parallel data mining which is also an exciting area and highly related to its distributed counterpart. Workshop Description The primary thrust of this workshop will be on knowledge discovery from distributed data. However, high quality research on high performance parallel data mining is also relevant to the workshop. The proposed workshop will provide a platform to discuss both theoretical and applied research issues in distributed and parallel knowledge discovery (DPKD). The topics of interest include, but are not limited to: 1) Theoretical foundation of DPKD. 2) Methods and algorithms: Advanced distributed and parallel data analysis and knowledge discovery algorithms. 3) Architectural issues: Architecture, control, security, and communication issues in DPKD. 4) Distributed data analysis in mobile computing environments. 5) Experimental systems: Large experimental systems, performance design issues. 6) Software agents and DPKD: Agent based approaches in DPKD. Agent interaction: cooperation, collaboration, negotiation, organizational behavior. 7) Distributed and Parallel knowledge discovery from spatial data. 8) Applications of DPKD: Application in business, science, engineering, medicine, and other disciplines. 9) Human-computer interaction in DPKD: Human-computer interaction in DPKD, multi-user interaction in DPKD. 10) Distributed Data analysis on the Internet. Important Dates March 1, 2000: Workshop Call for Papers. May 15, 2000: Papers Due. June 15, 2000: Acceptance Notification. July 15, 2000: Revision Due. August 20, 2000: Workshop. for additional information see http://www.eecs.wsu.edu/~hillol/DKD/dpkd2000.html Previous | item21 | Next |
Copyright © 2000 KDnuggets. Subscribe to KDnuggets News!