AdvSecureNet: A Python Toolkit for Adversarial Machine Learning

TL;DR

AdvSecureNet is a comprehensive, flexible, and high-performance Python toolkit for adversarial machine learning, supporting multi-GPU setups, multiple interfaces, and reproducibility, aimed at enhancing research and defense strategies.

Contribution

It introduces the first multi-GPU compatible toolkit supporting both CLI and API, with external YAML configuration for improved versatility and reproducibility.

Findings

Supports multiple attack and defense methods

Enables high-performance multi-GPU experiments

Provides extensive evaluation metrics

Abstract

Machine learning models are vulnerable to adversarial attacks. Several tools have been developed to research these vulnerabilities, but they often lack comprehensive features and flexibility. We introduce AdvSecureNet, a PyTorch based toolkit for adversarial machine learning that is the first to natively support multi-GPU setups for attacks, defenses, and evaluation. It is the first toolkit that supports both CLI and API interfaces and external YAML configuration files to enhance versatility and reproducibility. The toolkit includes multiple attacks, defenses and evaluation metrics. Rigiorous software engineering practices are followed to ensure high code quality and maintainability. The project is available as an open-source project on GitHub at https://github.com/melihcatal/advsecurenet and installable via PyPI.