LS8 aktuell
Allgem. Info
Suche Mail an Webadmin@ls8
  Universität Dortmund UniDo-Logo
  Fachbereich Informatik Lehrstuhl für Künstliche Intelligenz UniDo-Logp

IJCAI 99 Workshop: Machine Learning for Information Filtering

Call for Participation
Workshop on
"Machine Learning for Information Filtering"
at IJCAI 99
August 1st, 1999
Stockholm, Sweden
The enormous growth of on-line information and electronic commerce has brought about a comparable growth in research on methods for automatically organizing and personalizing information.  The "information filtering" task has simultaneously emerged as an active research topic in several disciplines, including information retrieval, human computer interaction, natural language processing, and machine learning.  The information filtering task manifests itself in many theoretically challenging and commercially important applications, such as electronic commerce and marketing, search engines, information push applications, browsing assistants, and adaptive Web sites.  

The goal of this workshop is to bring together researchers working on information filtering from many subfields of AI, while emphasizing the machine learning techniques and algorithms many of these subfields share. These techniques include  

  • text classification methods (probabilistic methods, support vector machines, first order methods, use of unlabeled data, etc.)
  • collaborative filtering methods (use of complex user and object profiles (e.g. citation structure), novel clustering models and methods, etc.)
  • other methods for learning user preferences (learning orderings, etc.) 
  • combinations of approaches and multi-strategy learning
  • representational issues (knowledge representation, NLP techniques, representing interest, representing information objects,  feature selection, term weighting, data transformation, latent semantic indexing, etc.)
  • clustering methods (similarity measures, mixture models, etc.)
  • scalability issues
  • formal models and theory
  • handling different media (text, images, sound, etc.)
  • evaluation techniques
Besides these topics, the workshop covers all theoretical and methodological issues concerning information filtering. Submissions describing innovative applications of information filtering are also encouraged. By bringing together industrial representatives with researchers, the workshop will  
  • show how problems from industry present new research issues.
  • identify ways in which research results may be put in more widespread practice in an industrial setting.


  • Prof. Oren Etzioni (Univ. of. Washington)
  • Jan Pedersen (Director, Advanced Technology, InfoSeek Corp.)


April 15, 1999 Submission deadline
May 3, 1999 Notification of acceptance 
May 24, 1999 Camera-ready copy due 
August 1, 1999 Workshop


The workshop will be one full day, including invited talks, paper presentations, poster presentations, and numerous opportunities for discussion. Depending on submissions, there will be joint sessions with the workshop "Text Mining: Foundations, Techniques and Applications" on topics of common interest. Participation in the workshop is limited according to IJCAI regulations. All workshop participants have to register for the IJCAI conference. The working notes of the workshop will be published online. 


Those interested in making a presentation should submit a full paper electronically either as a Postscript or PDF to The first page of submitted papers should include: title, author names and affiliations, a brief abstract. It should also name a designated contact person with his or her postal address, electronic mail address, telephone and fax number.  Submissions should not exceed 8 pages (including the title page) according to the IJCAI formatting instructions and should be printable on 8.5" x 11" or A4 paper.  

Those interested in participating in the workshop, but not submitting a paper, should submit a one-page abstract of their research interests in learning methods for information filtering.  


Authors with contributions accepted for oral presentation will have 20 minutes for their talk (including time for questions).  Authors with poster presentations will have 5 minutes for a short summary of their poster. The size of the poster boards is 96cm x 242cm. An overhead projector, a data projector, and a PC running Windows and Netscape will be available. If you need any other equipment, please let us know as soon as possible.


 Thorsten Joachims (Chair)  
 Universitaet Dortmund  
 FB Informatik, LS8  
 Baroper Str. 301  
 44221 Dortmund, Germany  
 Phone: +49 231 755 5102  
 FAX: +49 231 755 5105  
 Andrew McCallum  
 Just Research  
 4616 Henry Street  
 Pittsburgh, PA 15213  
 Phone: (412) 683-9132  
 FAX: (412) 683-4175  
 Mehran Sahami  
 Epiphany, Inc.  
 2300 Geng Road  
 Palo Alto, CA 94303  
 Phone: (650) 496-2399  
 FAX: (650) 496-2431  
 Lyle Ungar 
 University of Pennsylvania  
 Dept of Comp. and Info. Sci.  
 200 S. 33rd St.  
 Philadelphia, PA 19104-6389  
 Phone: (215) 898-7449  
 FAX: (215) 898-0587  

Last modified November 30th, 1998 by Thorsten Joachims <
This Document is not available in english language [Index] [Forschung] [Lehre] [Allgemeines] [MitarbeiterInnen] [Intern] [Suchen] [webadmin@ls8]