|
|
|
ASPRET™ (Anti-Spoofing Pattern REcognition Technology™) is the result of a research started to give answers to some issues raised in the DARPA GARD (Guaranteeing AI Robustness against Deception) program. |
In the figure above: a CNN
neural network has been trained to recognize fighter jet models. A GAN neural
network was trained to generate physically feasible deception patterns for targeted
black-box attacks. |
|
|
©2024_Luca_Marchese_All_Rights_Reserved Aerospace_&_Defence_Machine_Learning_Company VAT:_IT0267070992 NATO_CAGE_CODE:_AK845 Email:_luca.marchese@synaptics.org |
Contacts_and_Social_Media |