TESTING DATA CONTROL: The attacker may utilize this to add perturbations to testing samples at model deployment time, as performed in evasion attacks to generate adversarial examples or in backdoor poisoning attacks.
Apostol Vassilev • Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations
Sometimes a computer will send out an explorer packet to a destination that will return the path it took, and the computer will insert instructions into the packet header telling the bridge or router the exact path the frame or packet should travel. This is called source routing, and is very much frowned upon – we should let the network devices fig
... See morePhil Martin • Simple CISSP
In the simplest form, we can use the model’s detection confidence to determine a score. But even here there are quite a few options to choose from:
- Lowest confidence - the score is the lowest confidence of all detected objects
- Average confidence - average of all confidences of detected objects
- Minimizing confidence delta - difference between confidenc
Active Learning with Domain Experts, a Case Study in Machine Learning
Little Snitch on OS X to detect and decide to allow or block when a specific application is connecting out, and learn where it’s connecting to. You can use Wireshark for further analysis, mentioned below.