diffai
A certifiable defense against adversarial examples by training neural networks to be provably robust
217 stars
16 watching
26 forks
Language: Python
last commit: 2 months ago abstract-interpretationattackdefenseneural-networkpytorchrobust