Immune to hacks: Inoculating deep neural networks to thwart attacks

If a sticker on a banana can make it show up as a toaster, how might strategic vandalism warp how an autonomous vehicle perceives a stop sign? Now, an immune-inspired defense system for neural networks can ward off such attacks, designed by engineers, biologists and mathematicians at the University

Related Keywords

Michigan ,United States ,Ren Wang ,Silvia Cardarelli ,Indika Rajapakse ,Stephen Lindsly ,Defense Advanced Research Projects Agency ,Department Of Defense ,Army Research Office ,Johnh Holland Distinguished University Professor ,University Of Michigan ,Robust Adversarial Immune Inspired Learning ,Alfred Hero ,Distinguished University Professor ,Image Credit ,Robust Deep ,Nearest Neighbor ,Nearest Neighbors ,Projected Gradient Descent ,Betty Williams Professor ,Army Research ,