English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Recently, a backdoor data poisoning attack was proposed, which addsmislabeled examples to the training set, with an embedded backdoor pattern,aiming to have the classifier learn to classify to a target class whenever thebackdoor pattern is present in a test sample. Here, we address post-trainingdetection of innocuous perceptible backdoors in DNN image classifiers, whereinthe defender does not have access to the poisoned training set, but only to thetrained classifier, as well as unpoisoned examples. This problem is challengingbecause without the poisoned training set, we have no hint about the actualbackdoor pattern used during training. This post-training scenario is also ofgreat import because in many practical contexts the DNN user did not train theDNN and does not have access to the training data. We identify two importantproperties of perceptible backdoor patterns - spatial invariance and robustness- based upon which we propose a novel detector using the maximum achievablemisclassification fraction (MAMF) statistic. We detect whether the trained DNNhas been backdoor-attacked and infer the source and target classes. Ourdetector outperforms other existing detectors and, coupled with animperceptible backdoor detector, helps achieve post-training detection of allevasive backdoors.

Revealing Perceptible Backdoors, without the Training Set, via the Maximum Achievable Misclassification Fraction Statistic