Birmingham City University Develops New Defense Mechanism Against Cyberattacks On AI Systems
Image Transformation Method Outperforms State-Of-The-Art Random Noise Defense
AI systems used in healthcare are vulnerable to adversarial cyberattacks, which are a growing concern, said Atif Azad, a professor of AI at Birmingham City University. Azad’s research group has developed a method that trains AI to become more resilient to cyber threats through the use of random image adjustments.
