softsys4ai/athena
Athena: A Framework for Defending Machine Learning Systems Against Adversarial Attacks
PythonMIT
Issues
- 1
- 1
Organize the defense in a folder
#23 opened by MENG2010 - 1
- 1
- 5
- 1
End-to-end demo for our defense approach
#25 opened by oceank - 4
Normalized l2-dissimilarity
#35 opened by ScottLiao920 - 1
- 1
blackbox attack
#3 opened by MENG2010 - 1
- 1
- 1
- 1
Evaluate White-box Threat Model
#28 opened by MENG2010 - 1
Evaluate Grey-box Threat Model
#29 opened by MENG2010 - 1
Support MIM attack
#26 opened by MENG2010 - 4
Get scripts run on rci nodes.
#2 opened by MENG2010 - 0
construct tests for this project
#7 opened by MENG2010 - 1
pgd attack
#4 opened by MENG2010 - 0
DEEPFOOL (l_inf norm)
#5 opened by MENG2010 - 3
Issues running script on gcloud
#9 opened by MENG2010 - 1
support one-pixel attack
#1 opened by MENG2010 - 1
- 2
- 1
make current defense approaches (ensemble models) to not depend on a specific AE type
#13 opened by oceank - 1
- 2
poisson-noise transformation model crashes when predicting the AE, jsma (theta-10, gamma-30)
#18 opened by oceank - 7
Tune and fix bugs for new transformations
#22 opened by MENG2010 - 0
- 1
Detection as a defense
#11 opened by oceank - 0
use FLAG to manage configurations
#16 opened by MENG2010 - 0
manage the information with logging
#15 opened by MENG2010