Skip to content

A lightweight plugin that improves malware classifiers' robustness against adversarial attacks on Windows executables (EXEmples). Based on the research paper "Updating Windows Malware Detectors: Balancing Robustness and Regression against Adversarial EXEmples" (2025).

License

Notifications You must be signed in to change notification settings

matouskozak/EXE-scanner

Repository files navigation

EXE-scanner

EXE-scanner, lightweight plugin ready to be deployed with main malware classifiers to increase robustness against adversarial EXEmples. See the provided tutorial on how to train and use your own EXE-scanner.

image

Download

Download dataset and pretrained models from here: https://kaggle.com/datasets/fca93b34e3d3ed8936fb76cc06b4a7a94f9f296eebd675de2fab682857e24232

Citing

If you use this work, pleace cite the following paper:

@article{kozak2025updating,
  title={Updating Windows Malware Detectors: Balancing Robustness and Regression against Adversarial EXEmples},
  author={Kozak, Matous and Demetrio, Luca and Trizna, Dmitrijs and Roli, Fabio},
  journal={Computers \& Security},
  volume={155},
  pages={104466},
  year={2025},
  publisher={Elsevier},
  doi={https://doi.org/10.1016/j.cose.2025.104466}
}

About

A lightweight plugin that improves malware classifiers' robustness against adversarial attacks on Windows executables (EXEmples). Based on the research paper "Updating Windows Malware Detectors: Balancing Robustness and Regression against Adversarial EXEmples" (2025).

Topics

Resources

License

Stars

Watchers

Forks