Multiple Classifier Systems for Adversarial Classification Tasks

By Battista Biggio, Giorgio Fumera, Fabio Roli. 2009

44 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Networks and Communications, Safety, Robustness & Interpretability, Signal Processing

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 Computer Networks and Communications

In Safety, Robustness & Interpretability

In Signal Processing

Citations in the corpus

Evading Classifiers by Morphing in the Dark (2017, background): “Another proposed defensive mechanism is to use a multi-classifier system whose classification score is randomly picked from different models trained with disjoint feature sets [10].”
Optimal randomized classification in adversarial settings (2014, background): “Randomization is often noted to endow a defender with substantial power over a deterministic choice of defense [3, 4].”
Cracking Classifiers for Evasion: A Case Study on the Google's Phishing Pages Filter (2016, background): “MCSs [13][14][15], as the name suggests, uses multiple classifiers rather than only one to improve classifier’s robustness.”
Multiple classifier systems for robust classifier design in adversarial environments (2010, methodology): “In this paper we extend our recent works [ 6 , 7] with a more thorough discussion on bagging and the RSM, and a more extensive experimental investigation.”
Multiple Classifier Systems under Attack (2010, background): “In [16] we provided an empirical evidence(1) that a MCS architecture can be more robust than a single classifier architecture.”
One-and-a-Half-Class Multiple Classifier Systems for Secure Learning Against Evasion Attacks at Test Time (2015, methodology): “Linear classifiers exhibiting the aforementioned behavior have been implemented efficiently using MCSs, by averaging an ensemble of linear classifiers [23,26,27,18].”
On the hardness of evading combinations of linear classifiers (2013, background): “[2] have also analyzed the difficulty of evading combinations of linear classifiers, although they use a different measure of evasion difficulty.”
Pattern Recognition Systems under Attack: Design Issues and Research Challenges (2014, background): “Multiple Classifier Systems (MCSs) have been exploited for different purposes in adversarial settings, in particular, to improve security against evasion attempts [47, 17, 15, 16, 18] and against poisoning of the training data [9].”