Analysis of Causative Attacks against SVMs Learning from Data Streams

By Cody Burkard, Brent Lagesse. 2017

59 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Networks and Communications

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

Citations in the corpus

Stronger data poisoning attacks break data sanitization defenses (2018, background): “This makes indiscriminate attacks statistically interesting, as they get at fundamental properties of the model: how might an attacker that only controls 1% of the training data bring about a 10% increase in test error? A different type of attack is a targeted attack, in which the attacker seeks to ”
Certified Defenses for Data Poisoning Attacks (2017, methodology): “This is in contrast to targeted attacks on individual examples or sub-populations (e.g., Burkard and Lagesse, 2017).”
Reflection Backdoor: A Natural Backdoor Attack on Deep Neural Networks (2020, background): “The objective for data poisoning attack is to disrupt the proper training of DNN models by reducing their test-time performance on all or a specific subset of test examples [5,7,31,46,47,69].”
Clean-Label Backdoor Attacks (2018, background): “The most natural form of data poisoning aims to use the malicious samples aim to reduce the test accuracy of the resulting model (Xiao et al., 2012; 2015; Newell et al., 2014; Mei & Zhu, 2015; Burkard & Lagesse, 2017).”
MetaPoison: Practical General-purpose Clean-label Data Poisoning (2020, background): “Data poisoning attacks have been explored for classical scenarios (Biggio et al., 2012; Steinhardt et al., 2017; Burkard & Lagesse, 2017) which allow both training inputs and labels to be modiﬁed.”
Hidden Trigger Backdoor Attacks (2019, methodology): “d Laskov 2012; Biggio et al. 2013) where data poisoning was used to ﬂip the results of a SVM classiﬁer. More advanced methods were proposed in (Xiao et al. 2015; Koh and Liang 2017; Mei and Zhu 2015; Burkard and Lagesse 2017; Newell et al. 2014) which change the result of the classiﬁer on the clean ”
A Survey on Security Threats and Defensive Techniques of Machine Learning: A Data Driven View (2018, methodology): “, SVM [43], [44], neural networks [45], Latent Dirichlet Allocation [46], principle component analysis (PCA) and LASSO [47].”
Dataset Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses (2020, background): “Another work poisons SVM on data streams [44] and is the first to study targeted poisoning rather than general performance degradation.”