Port 22

New Framework Released to Protect Machine Learning Systems From Adversarial Attacks

Microsoft, in collaboration with MITRE, IBM, NVIDIA, and Bosch, has released anew open frameworkthat aims to help security analysts detect, respond to, and remediate adversarial attacks against machine learning (ML) systems. Called theAdversarial ML Threat Matrix, the initiative is an attempt to organize the different techniques employed by malicious adversaries in subverting ML systems. Just