Learning from Imbalanced Data Sets II

ICML'2003 Workshop:
Learning from Imbalanced Data Sets II

Thursday, August 21, 2003

Washington, DC


Nitesh Chawla Business Analytic Solutions, CIBC (chawla@morden.csee.usf.edu)
Nathalie Japkowicz      University of Ottawa (nat@site.uottawa.ca)
Aleksander Kolcz America Online, Inc. (ark@pikespeak.uccs.edu)

Workshop Description:

Overview: Recent years brought increased interest in applying machine learning techniques to difficult "real-world" problems, many of which are characterized by imbalanced learning data, where at least one class is under-represented relative to others. Examples include (but are not limited to): fraud/intrusion detection, risk management, medical diagnosis/monitoring, bioinformatics, text categorization and personalization of information. The problem of imbalanced data is often associated with asymmetric costs of misclassifying elements of different classes. Additionally the distribution of the test data may differ from that of the learning sample and the true misclassification costs may be unknown at learning time.

The AAAI-2000 Workshop on "Learning from Imbalanced Data Sets" provided the first venue where this important problem was explicitly addressed and has been received with much interest. The related ICML-2000 Workshop "Cost-Sensitive Learning" provided another venue for addressing the problem of asymmetric costs of different classes and features. Although much awareness of the issues related to data imbalance has been raised, many of the key problems still remain open and are in fact encountered more often, especially when applied to massive datasets. We believe that it would be of value to the machine learning community to not only examine the progress achieved in this area over the last three years but also discuss the current school of thought on research in learning from imbalanced datasets. Based on our understanding of class imbalance problem, the following topics of discussion are proposed (but not limited to):

Proposed Format: The workshop will open with an invited talk by Foster Provost that will introduce and overview the topic. Presentations will then be organized into several sessions corresponding roughly to the to the categories identified above. The workshop will conclude with a discussion during which a distinguished guest will comment on the presentations of the day, and open the floor for general discussion.

Proposed Length: One Day during which each panel will be allocated 1 to 2 hours, depending on the number of contributions and the expected length of the discussion session.

Workshop Notes: The accepted papers will be available electronically from the workhop website, and also as printed workshop notes to the attendees.

Submissions: Authors are invited to submit papers on the topics outlined above or on other related issues. Submissions should not exceed 8 pages, and should be in line with the ICML style sheet . Electronic submissions, in PDF format, are prefered and should be sent to Nitesh Chawla at chawla@morden.csee.usf.edu. If electronic submissions are inconvenient, please send four hard copies of your submission to:

Dr. Nitesh Chawla
Business Analytic Solutions, TBRM,
CIBC, BCE Place,
161 Bay Street, 11th Floor,
Toronto, Ontario M5J 2S8,


Invited Speakers:

Foster Provost                   New York University, USA
Charles Elkan                   University of California San Diego, USA
Naoki Abe TJ Watson Research Research Center, IBM, USA

Program Committee:

Kevin Bowyer University of Notre Dame, USA
Chris Drummond National Research Council, Canada
Charles Elkan University of California San Diego, USA
Marko Grobelnik Jozef Stefan Institute, Slovenia
Larry Hall University of South Florida, USA
Robert Holte University of Alberta, Canada
W. Philip Kegelmeyer       Sandia National Labs, USA
Miroslav Kubat         University of Miami, USA
Aleksandar Lazarevic University of Minnesotta, USA
Charles Ling University of Western Ontario, Canada
Dragos Margineantu Boeing Corporation, USA
Foster Provost New York University, USA
Gary Weiss AT&T Labs, USA