Alert button

Dynamic Stochastic Ensemble with Adversarial Robust Lottery Ticket Subnetworks

Add code
Alert button
Oct 06, 2022
Figure 1 for Dynamic Stochastic Ensemble with Adversarial Robust Lottery Ticket Subnetworks
Figure 2 for Dynamic Stochastic Ensemble with Adversarial Robust Lottery Ticket Subnetworks
Figure 3 for Dynamic Stochastic Ensemble with Adversarial Robust Lottery Ticket Subnetworks
Figure 4 for Dynamic Stochastic Ensemble with Adversarial Robust Lottery Ticket Subnetworks

Share this with someone who'll enjoy it:

Adversarial attacks are considered the intrinsic vulnerability of CNNs. Defense strategies designed for attacks have been stuck in the adversarial attack-defense arms race, reflecting the imbalance between attack and defense. Dynamic Defense Framework (DDF) recently changed the passive safety status quo based on the stochastic ensemble model. The diversity of subnetworks, an essential concern in the DDF, can be effectively evaluated by the adversarial transferability between different networks. Inspired by the poor adversarial transferability between subnetworks of scratch tickets with various remaining ratios, we propose a method to realize the dynamic stochastic ensemble defense strategy. We discover the adversarial transferable diversity between robust lottery ticket subnetworks drawn from different basic structures and sparsity. The experimental results suggest that our method achieves better robust and clean recognition accuracy by adversarial transferable diversity, which would decrease the reliability of attacks.

Share this with someone who'll enjoy it: