Fig. 5From: Spears and shields: attacking and defending deep model co-inference in vehicular crowdsensing networksInference accuracy with respect to privacy budgetBack to article page