Defending Against Neural Network Model Inversion Attacks via Data Poisoning | IEEE Journals & Magazine | IEEE Xplore