Call for Papers

Theories and Applications of
Unsupervised and Semi-Supervised Learning

=========================================================================

A Special Track at the 16th IEEE International Conference on Tools with Artificial Intelligence (ICTAI-2004) 

November 15-17, 2004, Marriot Hotel, Boca Raton, Florida

 

ICTAI 2004 Conference Web Site: http://www.cse.fau.edu/~ictai04/

==========================================================================

Modern data mining and machine learning applications often involve learning from large amounts of data without output (i.e., category labels) or with only a very limited number of labels. Examples include automatic categorization of document collections, gene function analysis from gene expression (DNA microarray) data, hyper-spectral image segmentation/classification, and content-based image retrieval, etc. In these applications, category labels are usually difficult or expensive to get, or even dynamic. Traditional classification techniques have become insufficient in addressing these challenges; Various unsupervised clustering and semi-supervised learning algorithms have recently been proposed and successfully employed.

The success of unsupervised and semi-supervised learning motivates further enhancements to existing algorithms and proposing new algorithms to cope with the requirements of real world problems. While typical applications have focused on clustering and classification tasks, there is a spectrum of possible learning situations such as:

  • learning from completely unlabeled data,

  • learning from unlabeled data and both positively and negatively labeled data,

  • learning from unlabeled and only positively (or only negatively) labeled data,

  • learning from partially incorrectly labeled data and unlabeled data.

This special track solicits and welcomes papers in the general area of applications of unsupervised and semi-supervised learning as well as algorithmic enhancements to handle issues raised in real world problems.

Topics of interest (include but are not limited to)

  • Clustering with constraints

  • Applications of clustering in AI

  • Modeling of learning with partial labels

  • Semi-supervised learning methods and applications

  • Theoretical or empirical evaluation of the value of labeled and unlabeled data

  • Comparative study of existing unsupervised and semi-supervised learning methods

  • Any related applications (text clustering/classification, CBIR, scientific data analysis, data mining for intrusion detection, ......)

Key dates

 

July 2, 2004:         Extended deadline for paper submission (this new deadline is firm and no late submissions will be accepted)

August 2, 2004:       Notification of acceptance

September 3, 2004: Camera-ready papers due

 

Submission Guideline

 

Papers submitted to the special track must not have been previously published and must not be currently under consideration for publication elsewhere.  Authors must adhere to all submission procedures of ICTAI-2004, including deadlines and paper format. In addition to submitting papers via the ICTAI-2004 submission system, authors are required to submit a copy of their paper(s) by email to special track co-chair Ian Davidson (davidson@cs.albany.edu) or Shi Zhong (zhong@cse.fau.edu) on or before June 18, 2003.

 

General guideline for ICTAI-2004 repeated here: The manuscript should not exceed 20 double-spaced, single-column pages, including figures and tables. The font size should be at least 10 point.  All paper submissions will be handled electronically through our online web submission system. The submission of your paper should be in either PDF or PS file format. If for some reason you cannot submit your paper through our web-based submission system, you can email your submission to the Program Committee Chair at the address you may find here. Make sure that you include, in your email, the title, author name(s) and full address (email and postal), contact author, an abstract and the full manuscript in PDF, or PS format.

 

Note

 

It is required that each accepted paper must be presented by at least one author at the conference in order for the paper to be included in the conference proceedings. A selection of the best papers in the conference will be published in a special issue of the International Journal on Tools with Artificial Intelligence.

 

Program Committee

 

Ian Davidson, State Univ. of New York – Albany (co-chair)

Shi Zhong, Florida Atlantic Univ. (co-chair)

Chris Ding, Lawrence Berkeley National Laboratory

Mohammed Zaki, Rensselaer Polytechnic Institute

Massih Amini, University of Marie Currie

Rohan Baxter, CSIRO – Australia

Larry Hall, Univ. of South Florida

Mei-Ling Shyu, Univ. of Miami

Kiri Wagstaff, Jet Propulsion Laboratory 

Rayid Ghani, Accenture Technology Labs

 

Any questions regarding this special track or the submission procedure, please don't hesitate to contact the track co-chairs at davidson@cs.albany.edu or zhong@cse.fau.edu.

 

Last revised, June 2004