Workshop "Machine Learning for User Modeling"
Seventh International Conference on User Modeling Banff, Canada, 20-24 June 1999

• Call for Participation
• Workshop Topic
• Workshop Schedule
• Workshop Papers
• Organizers

Workshop Topic

The information about the user that is available to an interactive computer system is usually limited, so that user model acquisition is a difficult problem. Classical acquisition methods like user interviews, application-specific heuristics, and stereotypical inferences often are not flexible enough.

Machine Learning is concerned with the formation of models from observations. Hence, learning algorithms seem to be promising candidates for user model acquisition systems. Recently, a growing interest in the application of machine learning techniques for user modeling purposes has been apparent. At the last user modeling conference (UM97), a first workshop on "Machine Learning for User Modeling" took place, and a special interest group on that topic was initiated. The ML4UM SIG now has both a web site and a mailing list with about 150 subscribers.

This workshop has three goals: First, it attempts to be a forum for user modeling researchers who want to discuss specific problems of using machine learning for user modeling. Second, it shall provide an opportunity for user modeling researchers who are not so familiar but want to get acquainted with the possibilities of using machine learning for user modeling. These "novices" shall benefit from the insights of more experienced colleagues, while the latter shall be pointed to new applications and problems. Third, the workshop shall function as a SIG meeting, where joint activities of interested attendants can be planned.

Workshop Schedule

Workshop times: Wednesday, June 23, 14:00 - 18:30

On Wednesday, there is no official conference event in the evening. Moreover, since some participants can be expected to leave Banff early, and some seem to have problems attending the workshop on Thursday at all, it seems preferrable to run all sessions on Wednesday.

• 14:00-15:00 Session 1 (Chair: Mathias Bauer)
  Information Filtering: What has been learned, and what are the prospects?
  Panelists: Martin Müller, Asa Rudstroem, Daniel Billsus, Sofus Macskassy



• 15:00-15:45 Session 2 (Chair: Asa Rudstroem)
  Learning about the user, about all users, or from all users: Dealing with information about individuals and/or user groups
  Panelists: Tony Jameson, Frank Linton, Haym Hirsh (representing Basu/Hirsh)



• 15:45-16:00 Break



• 16:00-16:30 Session 3 (Chair: Wolfgang Pohl)
  Learning about user actions: Action Sequences and Plan Recognition
  Panelists: Frank Linton, Mathias Bauer



• 16:30-17:15 Session 4 (Chair: Haym Hirsh)
  Machine Learning for User Modeling: Specific Problems
  Panelists: Wolfgang Pohl, Asa Rudstroem, Daniel Billsus



• 17:15-17:30 Break



• 17:30-18:00 Session 5 (Concluding Plenary Discussion)
  ML4UM - future issues and directions



• 18:00-18:30 ML4UM SIG issues

Workshop Papers

• Chumki Basu, Haym Hirsh: Learning User Models for Recommendation HTML PS PDF
  


• Mathias Bauer: Machine Learning for Plan Recognition PS
  


• Daniel Billsus: Improving User Model Acquisition from Labeled Text Documents PDF
  


• Frank Linton: ML4UM Workshop Submission HTML
  


• Sofus A. Macskassy, Haym Hirsh, Aynur Dayanik: EmailValet: Learning Email Preferences for Wireless Platforms PS
  


• Martin Müller: Machine Learning based User Modeling for WWW Search PS HTML
  


• Asa Rudstroem: ML4UM in Edited Information Services HTML
  


• Ingo Schwab, Wolfgang Pohl: Learning Information Interest from Positive Examples PS
  


• Frank Wittig, Barbara Großmann-Hutter, Anthony Jameson: Learning Bayesian Networks with Hidden Variables for User Modeling PDF PS
  

Organizers

• Mathias Bauer, DFKI, Germany (bauer@dfki.de)
• Piotr Gmytrasiewicz, Univ. of Texas at Arlington, U.S.A. (piotr@huckle.uta.edu)
• Wolfgang Pohl, GMD FIT, Germany (Wolfgang.Pohl@gmd.de)