Sign In

Communications of the ACM

ACM TechNews

AI Networks Vulnerable to Attacks

View as: Print Mobile App Share:
The research team has made QuadAttacK publicly available, so that the research community can use it themselves to test neural networks for vulnerabilities.

The researchers developed a piece of software called QuadAttacK, which can be used to test any deep neural network for adversarial vulnerabilities.

Credit: bigstock

Tianfu Wu at North Carolina State University and colleagues developed the QuadAttacK software to test deep neural networks for adversarial vulnerabilities.

Wu explained that a trained artificial intelligence (AI) system tested with clean data will behave as predicted.

"QuadAttacK watches these operations and learns how the AI is making decisions related to the data," he said. "This allows QuadAttacK to determine how the data could be manipulated to fool the AI. QuadAttacK then begins sending manipulated data to the AI system to see how the AI responds. If QuadAttacK has identified a vulnerability, it can quickly make the AI see whatever QuadAttacK wants it to see."

The researchers used their QuadAttacK software to test four deep neural networks that are widely used and found that all four to be vulnerable.

From NC State University News
View Full Article


Abstracts Copyright © 2023 SmithBucklin, Washington, D.C., USA


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account