Efficient Project Gradient Descent for Ensemble Adversarial Attack
Fisrt Place Solution for IJCAI19 Targeted Adversarial Attack competition; Sumbitted to IJCAI19 AIBS WORKSHOP
Fanyou Wu, Rado Gazo, Eva Haviarova and Bedrich Benes @ Purdue University
Recent advances show that deep neural networks are not robust to deliberately crafted adversarial examples which many are generated by adding human imperceptible perturbation to clear input. Consider
- Download model parameters
- Use Docker to set environment.
run.sh ./dev_data ./outputto run codes.