Differential Analysis of Triggers and Benign Features for Black-Box DNN Backdoor Detection. (arXiv:2307.05422v1 [cs.CR])

This paper proposes a data-efficient detection method for deep neural
networks against backdoor attacks under a black-box scenario. The proposed
approach is motivated by the intuition that features corresponding to triggers
have a higher influence in determining the backdoored network output than any
other benign features. To quantitatively measure the effects of triggers and
benign features on determining the backdoored network output, we introduce five
metrics. To calculate the five-metric values for a given input, we first
generate several synthetic …

analysis attacks backdoor backdoor attacks box data detection features higher influence intuition network networks neural networks scenario under