wanghangpsu / acgan-ada Goto Github PK

Hi again,
This time I had a question regarding how the process of decision making regarding the All-AD method is carried out as the description in the paper is not very clear to me.

You form a vector using the three detection criteria and you model this using a multivariate gaussian and make a decision based on a desired p-value.

Since this part is seemingly not present in your codebase I would like to ask if you could elaborate on how exactly the All-AD detection works and how it is implemented.

Hi.
First I would like to thank you for you work on "Anomaly Detection of Adversarial Examples using Class-conditional Generative Adversarial Networks". You mentioned here that a grid search was conducted to obtain the thresholds for different detection methods (D-AD, All-AD and G-AD). I would like to know if you could share the exact values for these thresholds as I couldn't find them in the code.