DW-FedADMM improves federated learning for Android malware detection
Original title: Consensus-Constrained Dual-Federated Optimization with Client-Side Dynamic Weighting for Distributed Android Device Threat Defense
DW-FedADMM uses dynamic client-side weighting in federated optimization, achieving up to 97.64% accuracy on Drebin dataset with partial participation.
DW-FedADMM is a federated optimization framework combining primal-dual coordination with performance-dependent client-side weighting for distributed Android threat defense. It modulates local states before augmented-Lagrangian update. Evaluated on Drebin, Malgenome, and TUANDROMD datasets, it obtains highest accuracy on Drebin and best mean accuracy in 9 of 18 configurations, with 95.42% pooled accuracy.
- highest seed-averaged accuracy on Drebin (97.64% +/- 0.20%)
- best mean accuracy in 9 of 18 dataset-client-round configurations
- reaches 95.42% accuracy, 94.62% F1-score, 96.06% precision, and 99.45% AUC
- reduces variation across seeds on Drebin with longer budget
From the original article
1 min read at arxiv.orgSmartphones, smart-home hubs, and wearables are increasingly exposed to sophisticated Android malware, while centralized threat-detection pipelines raise privacy and scalability concerns. Federated learning (FL) enables collaborative model training without collecting raw client data, but partial participation and variation in local model quality can destabilize global optimization.
By Mohammad Mashreghi · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.