AI Sec Watch: A Security Intelligence Platform for AI Systems

Luu, T.J.

Rethinking Fake Adversarial Examples for Single-Step Adversarial Training

inforesearchPeer-Reviewed

researchsafety

Source: IEEE Xplore (Security & AI Journals)July 20, 2026

Summary

Adversarial training (a method where AI models learn to defend against adversarial attacks, which are inputs designed to fool the model) typically requires expensive multi-step calculations, so researchers use single-step versions to save computing power. However, single-step methods create some misleading adversarial examples, called "fakers," that actually weaken the model instead of strengthening it. This paper proposes FAST (Faker-Alleviating Single-step adversarial Training), which reduces the impact of fakers during training by dynamically adjusting how the model learns from difficult examples and introducing auxiliary samples to stabilize the training process.

Solution / Mitigation

The paper proposes FAST, which "consists of two main components. First, it dynamically adjusts the label-smoothing level for adversarial examples according to their learning difficulty, making fakers easier for the model to learn. Second, it introduces an auxiliary sample with a weak adversarial effect, derived from the single-step adversarial example, which is used to dynamically ease the alignment with clean data and stabilize the optimization process." The code is available at https://github.com/mesunhlf/FAST.