Classifier evaluation and attribute selection against active adversaries

By Murat Kantarcioglu, Bowei Xi, Chris Clifton. 2010

88 citations recorded by the source corpus.

Many data mining applications, such as spam filtering and intrusion detection, are faced with active adversaries. In all these applications, the future data sets and the training data set are no longer from the same population, due to the transformations employed by the adversaries. Hence a main assumption for the existing classification techniques no longer holds and initially successful classifiers degrade easily. This becomes a game between the adversary and the data miner: The adversary modifies its strategy to avoid being detected by the current classifier; the data miner then updates its classifier based on the new threats. In this paper, we investigate the possibility of an equilibrium in this seemingly never ending game, where neither party has an incentive to change. Modifying the classifier causes too many false positives with too little increase in true positives; changes by the adversary decrease the utility of the false negative items that are not detected. We develop a game theoretic framework where equilibrium behavior of adversarial classification applications can be analyzed, and provide solutions for finding an equilibrium point. A classifier’s equilibrium performance indicates its eventual success or failure. The data miner could then select attributes based on their equilibrium performance, and construct an effective classifier. A case study on online lending data demonstrates how to apply the proposed game theoretic framework to a real application.

Subfields: Artificial Intelligence, Information Systems, Management Science and Operations Research

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

In Management Science and Operations Research

Citations in the corpus

A Game Theoretical Model for Adversarial Learning (2009, background): “As a result, a new subfield of “adversarial learning” has emerged to understand and design classifier which are robust to adversarial transformation[1], [2], [3], [4], [5], [6], [7].”
Mining adversarial patterns via regularized loss minimization (2010, background): “But a common problem for Kantarcioglu et al. (2009) and Liu and Chawla (2009 )i s that the strategies of the adversary are stochastically sampled (e.g., Monte Carlo integration in Kantarcioglu et al. 2009 ) and then among the samples the best fit is selected (e.g., genetic algorithm in Liu and Chaw”
Stackelberg games for adversarial prediction problems (2011, background): “Some authors [9, 12] study the case in which the data generator acts as leader and the learner as follower.”
Adversarial machine learning (2019, background): “Subsequent game theoretic approaches to learning have extended this setting and solved for equilibria of the game [8, 32].”