MiniFool generates uncertainty-aware adversarial perturbations and uses the attack strength needed to flip a classification as a measure of how robust the network's decision is.
CMS.G6CE.1ITV
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
MiniFool -- Physics-Constraint-Aware Minimizer-Based Adversarial Attacks in Deep Neural Networks
MiniFool generates uncertainty-aware adversarial perturbations and uses the attack strength needed to flip a classification as a measure of how robust the network's decision is.