Smart Cities AVNs: Transparent and Privacy-Preserving Cyber Attack Prediction with Explainable Federated Feature Selection
Abstract
Abstract In recent years, the rapid integration of Autonomous Vehicular Networks (AVNs) into smart cities has increased the exposure of connected mobility systems to cyber-attacks. Their distributed, dynamic, and heterogeneous nature makes reliable intrusion detection challenging, particularly when Vehicular Intrusion Detection Systems (VIDS) must preserve data privacy, maintain high predictive performance, and provide transparent decisions. Traditional centralized IDS/ML approaches require pooled traffic data and commonly rely on centralized feature selection, creating privacy, communication, and interpretability limitations. Federated Learning (FL) enables collaborative learning without sharing raw vehicular data; however, conventional FL gives limited attention to distributed feature selection, non-IID client heterogeneity, and decision transparency. Explainable Artificial Intelligence (XAI) can improve trust by revealing the factors underlying model predictions, yet it is commonly applied only after model development rather than being integrated with distributed learning and feature-selection processes. To address these limitations, this research proposes an explainable federated VIDS for multiclass cyber-attack prediction in smart-city AVNs. Four class-aware non-IID clients independently estimate feature relevance using Mutual Information and regularized Random Forest importance, while sample-size weighting and cross-client agreement are used to derive a global feature…
Read the paper
Related studies
- Active Learning during Federated Learning for Object Detection
- Fairness-aware federated learning
- Decentralized Federated Learning
- Building a Personally Identifiable Information Recognizer in a Privacy Preserved Manner Using Automated Annotation and Federated Learning
- Preservation of Patient Level Privacy: Federated Classification and Calibration Models
- Efficient explainable federated learning
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.