1 paper
Mohamed Suliman, Anisa Halimi, Swanand Kadhe +2
Data forging attacks provide counterfactual proof that a model was trained on a given dataset, when in fact, it was trained on another. These attacks work by forging (replacing) mi…