A BIC-based Mixture Model Defense against Data Poisoning Attacks on Classifiers. (arXiv:2105.13530v2 [cs.LG] UPDATED)

Data Poisoning (DP) is an effective attack that causes trained classifiers to
misclassify their inputs. DP attacks significantly degrade a classifier's
accuracy by covertly injecting attack samples into the training set. Broadly
applicable to different classifier structures, without strong assumptions about
the attacker, an {\it unsupervised} Bayesian Information Criterion (BIC)-based
mixture model defense against "error generic" DP attacks is herein proposed
that: 1) addresses the most challenging {\it embedded} DP scenario wherein, if
DP is present, the poisoned samples are …

arxiv attacks bic classifiers data defense