Program
Schedule
 
Keynotes
Invited Talks
 
Papers
Posters
IT Papers
 
  Panels
Tutorials
Workshops
 
Exhibits
KDD Cup
 
Attending
Registration
Hotel
Airline
 
Organization
Chairs
PC
 

KDD-2001 Panels


When and How to Subsample (Monday)

CHAIR: Pedro Domingos

PARTICIPANTS:
  • Surajit Chaudhuri, Microsoft Research
  • David Jensen, University of Massachusetts at Amherst
  • Ronny Kohavi, Blue Martini
  • Foster Provost, New York University
ABSTRACT: Databases in the terabyte range are now common. In many domains, mining all the data available in reasonable time is already beyond the reach of current systems. Yet the size of databases continues to grow rapidly. Is subsampling unavoidable? Or should it be avoided at all costs? If we subsample, what is the best way to do it? What issues must be taken into account? In this panel we will address these and related questions, with the twin goals of developing practical guidelines and identifying key research issues.


Data Mining Startups: The Perfect Storm (Tuesday)

CHAIR: George H. John

PARTICIPANTS: ABSTRACT: In the 1990's, a perfect storm of several factors coincided to create huge opportunities for data-mining entrepreneurs -- the maturation of data mining and supporting technologies coincided with a supportive investment climate and a strong appetite among corporations for software solutions to reduce costs or enhance revenues. Today, after a deceleration in the growth of the U.S. economy and a brisk devaluation of internet companies and technology companies in general, the environment for entrepreneurship is much more challenging, although the ultimate criteria for success remain unchanged. The Data Mining Startups panel brings together four senior members of the data-mining community to talk about their experiences, what kinds of data-mining startups succeeded or failed in the 1990's, and what they think the future holds.


New Research Directions in KDD (Wednesday)

CHAIR: Johannes Gehrke

PARTICIPANTS:
  • Rakesh Agrawal, IBM Almaden Research Center
  • Daryl Pregibon, AT&T Research Labs
  • Tom Mitchell, Whizbang! Labs
  • Ted Senator, DARPA
ABSTRACT: Data mining as a discipline has matured considerably, and there exists a multitude of scalable algorithms that transform oceans of bits in very large databases into interpretable patterns and predictive models. This panel, consisting of visionaries from academia, industry, and government, will focus on the next generation of data mining research. Panelists will talk about what they think the community should (and should not) be working on, what are important problems that are being neglected, what are emerging directions, and what a decade of KDD research teaches us for the future.