发帖
雷达卡
ProFlip: Targeted Trojan Attack with Progressive Bit Flips
Huili Chen Cheng Fu Jishen Zhao Farinaz Koushanfar
University of California, San Diego
Email: {huc044, cfu, jzhao, farinaz}@ucsd.edu
Abstract training-time attack that tampers with model weights by in-
jecting incorrectly labeled data into the training set [7, 28].
The security of Deep Neural Networks (DNNs) ...
京公网安备 11010802022788号
