Multiple classifier systems for robust classifier design in adversarial environments

By Battista Biggio, Giorgio Fumera, Fabio Roli. 2010

199 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

Wild Patterns: Ten Years After the Rise of Adversarial Machine Learning (2017, methodology): “These recent developments have enabled a deeper understanding on how to defend against evasion attacks in spam ﬁltering and malware detection, also clarifying (in a formal manner) the intuitive idea of uniform feature weights only heuristically provided in [42, 76].”
Evasion Attacks against Machine Learning at Test Time (2013, background): “, to avoid detection, spam emails are often modified by obfuscating common spam words or inserting words associated with legitimate emails [3, 9, 16, 19].”
Poisoning Attacks against Support Vector Machines (2012, background): “…or proposed as solutions to a number of such security-sensitive tasks including spam, worm, intrusion and fraud detection (Meyer & Whateley, 2004; Biggio et al., 2010; Stolfo et al., 2003; Forrest et al., 1996; Bolton & Hand, 2002; Cova et al., 2010; Rieck et al., 2010; Curtsinger et al., 2011;…”
Yes, Machine Learning Can Be More Secure! A Case Study on Android Malware Detection (2017, methodology): “tic approaches in [8], [26], but it is also able to outperform them in terms of security.”
On Security and Sparsity of Linear Classifiers for Adversarial Settings (2016, methodology): “For instance, the widely-used SpamAssassin anti-spam filter exploits a linear classifier [7,5].”
Adversarial Feature Selection Against Evasion Attacks (2016, background): “It has been, therefore, widely adopted in previous work [10], [11], [15], [16], [23], [26], and in several real anti-spam filters, like SpamAssassin and SpamBayes.”
Support vector machines under adversarial label contamination (2015, background): “Keywords: Support Vector Machines, Adversarial Learning, Label Noise, Label Flip Attacks”
On Evaluating Adversarial Robustness (2019, background): “This long line of research (Dalvi et al., 2004; Lowd & Meek, 2005; Barreno et al., 2006; 2010; Globerson & Roweis, 2006; Kołcz & Teo, 2009; Barreno et al., 2010; Biggio et al., 2010; Šrndic & Laskov, 2013) has recently begun seeing significant study as machine learning becomes more widely used.”
One-and-a-Half-Class Multiple Classifier Systems for Secure Learning Against Evasion Attacks at Test Time (2015, methodology): “The reason is that, in these applications, false positives are more harmful than false negatives, and thus, they should be kept very low [18,6].”
Secure Kernel Machines against Evasion Attacks (2016, background): “Then, we have shown that infinity-norm regularization is the optimal solution against worst-case sparse attacks (as the ones only qualitatively envisioned in [5, 18]).”
The relationship between trust in AI and trustworthy machine learning technologies (2019, unknown): “, in the training stage [26, 13, 11, 61, 85].”
A Survey on Security Threats and Defensive Techniques of Machine Learning: A Data Driven View (2018, methodology): “On the other hand, improving the robustness of learning algorithms is another feasible defending technique, e.g., Bootstrap Aggregating and Random Subspace Method (RSM) [78], [79].”
Pattern Recognition Systems under Attack: Design Issues and Research Challenges (2014, background): “Non-stationarity has also been considered, modeled as p(X,Y ; t), t being a parameter denoting time, but only of non-adversarial nature.”
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.”
A unifying view on dataset shift in classification (2012, methodology): “This type of problem is receiving an increasing amount of attention in the machine learning field [16,6,10,35], and usually copes with non-stationary environments due to the existence of an adversary that tries to work around the existing classifier’s learned concepts.”
Design of robust classifiers for adversarial environments (2011, methodology): “We consider GWI and BWO attacks in this experiment, and simulate them by modifying up to nMAX words in each spam email, as in [2], [13].”
Towards Adversarial Malware Detection (2018, background): “Developing specific learning approaches that distribute feature importance more evenly can significantly improve robustness to such adversarial evasion attempts [12, 29, 51, 68].”