rl-policies-attacks-defenses
rl-policies-attacks-defenses copied to clipboard
davide97l
Adversarial attacks on Deep Reinforcement Learning (RL)
can youo tell us which version of gym and atari-py you have used for this repo?
- [x] Pong - [x] Breakout - [ ] Pacman - [ ] Enduro
I think one file is missed in this repo： log_def/PongNoFrameskip-v4/dqn/policy.pth 
I want to know how do you set the acts_mask in critical point attack and critical strategy attack? how to set this argument for other atari games? Thank a lot!...
I found there are so many adversarial attacks in your code but in your paper, you only used FGSM and PGD Did you use other attcks to do experiment? if...
'Attack **Pong-PPO** policy with **Uniform Attack** with attack frequenc7 0.5. Moreover, let's suppose we don't know the agent policy is PPO and let's perform attacks on a **A2C** policy trained...
I cannot git clone the tianshou github, then l download the latest type, but this type report error. Can you tell me the type about tianshou
