|
Activity Number:
|
204
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #307041 |
|
Title:
|
Adversarial Learning
|
|
Author(s):
|
Bowei Xi*+ and Murat Kantarcioglu and Chris Clifton
|
|
Companies:
|
Purdue University and The University of Texas at Dallas and Purdue University
|
|
Address:
|
150 N. University Street, West Lafayette, IN, 47907,
|
|
Keywords:
|
data mining ; adversarial learning ; game theory
|
|
Abstract:
|
Many data mining applications, both current and proposed, are faced with an active adversary. Problems range from the annoyance of spam to the damage of computer hackers to the destruction of terrorists. These problems pose a significant new challenge: The behavior of a class (the adversary) may adapt to avoid detection. In all of these cases, data mining has been proposed as a solution: from training spam filter to using data mining to identify terrorists. We use a game theoretic approach to identify a steady-state: What happens when both parties are doing the best they can to achieve their conflicting goals? We will demonstrate that in a spam email setting.
|
