Document is current

Article History

Accepted: 8 November 2024

First Online: 26 May 2025

Declarations

:

: The author has no relevant financial or nonfinancial interests to disclose.

: This work proposes an attack sample generation algorithm to evaluate and improve the adversarial robustness of machine learning models. All experiments were performed on publicly available datasets (e.g., MNIST, CIFAR-10), with no access to private or sensitive information. The study adheres to the ethical guidelines of Southwest Petroleum University and is intended solely for defensive research. Any misuse of the generated samples is strictly prohibited.