method: Baseline - Multilayer Perceptron2025-03-18
method: Baseline - DREBIN2024-09-04
Authors: .
Description: Reproduction of the detector from Arp, Daniel, et al. “Drebin: Effective and explainable detection of android malware in your pocket.” NDSS. Vol. 14. 2014.
method: Continual-Positive Congruent Training2025-02-20
Authors: Daniele Ghiani
Affiliation: University of Cagliari
Description: Continual-Positive Congruent Training (C-PCT) is an adaptation of the well-known PCT strategy to a Continual Learning (CL) scenario.
In CL, data arrives incrementally, following a temporal order. Each time a new batch of data becomes available, the model is updated using only the most recent data.
Following this scenario the training data has been divided into 12 quarters.
The PCT regularization loss term encourages the new model under training, which is updated solely with the latest data, to mimic the previous model's behavior when it made correct predictions.
| False Positive Rate | Detection Rate | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Date | Method | False Positive Rate | Clean data | 25 manipulated features | 50 manipulated features | 100 manipulated features | |||
| 2025-03-18 | Baseline - Multilayer Perceptron | 0.32% | 68.48% | 6.80% | 4.00% | 2.64% | |||
| 2024-09-04 | Baseline - DREBIN | 0.36% | 77.28% | 0.48% | 0.08% | 0.00% | |||
| 2025-02-20 | Continual-Positive Congruent Training | 0.40% | 59.60% | 4.56% | 1.20% | 0.24% | |||
| 2024-09-04 | Baseline - SecSVM | 0.44% | 75.36% | 14.96% | 10.64% | 5.44% | |||
| 2025-03-10 | SVM-CB (b=0.2, n=100) | 0.82% | 75.36% | 4.16% | 3.28% | 2.32% | |||
| 2025-03-10 | SVM-CB (b=0.8, n=100) | 0.82% | 75.36% | 4.16% | 3.28% | 2.32% | |||
| 2025-04-04 | DeepTrust | 0.98% | 78.32% | 43.28% | 33.84% | 19.92% | |||