1 paper
Xiaojun Xu, Qi Wang, Huichen Li +3
In machine learning Trojan attacks, an adversary trains a corrupted model that obtains good performance on normal data but behaves maliciously on data samples with certain trigger…