1 paper
Mih Dinh, SouYoung Jin
Large-scale image datasets frequently contain identifiable or sensitive content, raising privacy risks when training models that may memorize and leak such information. We present…