This code repository prepared for our ICSE 2024 paper titled MALCERTAIN:Enhancing Deep Neural Network Based Android Malware Detection by Tackling Prediction Uncertainty and our TOSEM 2026 paper titled Towards Improved DNN-Based Android Malware Detection via Uncertainty Estimation.
In this paper, we take the first step to explore how we can leverage the prediction uncertainty to improve DNN-based Android malware detection models. Our key insight is if we can identify uncertainty metrics that differ greatly between correct and incorrect predictions, we can use these metrics to pinpoint the potentially incorrectly-classified samples and correct their classification results accordingly. We developed our code based on this project.
We develop the codes on Windows operation system, and run the codes on Ubuntu 20.04. The codes depend on Python 3.8.10. Other packages (e.g., TensorFlow) can be found in the ./requirements.txt.