A research team from Zhejiang University, Wuhan University and Adobe Research proposes Feature Importance-Aware Attacks (FIA) that drastically improve the transferability of adversarial examples, achieving superior performance compared to state-of-the-art transferable attacks.

Here is a quick read: Novel Feature Importance-Aware Transferable Adversarial Attacks Dramatically Improve Transferability.

The FIA code will be released on the project GitHub. The paper Feature Importance-Aware Transferable Adversarial Attacks is on arXiv.



Source link