The Internet of Things (IoT) plays a vital role in the digital age by interconnecting numerous heterogeneous devices. This complexity and ubiquity expose IoT networks to diverse and sophisticated cyber threats. IoT environments require rigorous collaborative intrusion detection system (IDS) that operates in distributed environments and heterogeneous data traffic. While federated learning offers a promising, privacy-conscious training model, most traditional approaches to IDS fail to detect the structural relationships between connected entities. On the other hand, Graph Neural Networks (GNNs) have attracted significant attention in Network Intrusion Detection Systems (NIDS) for their effectiveness in modeling complex network traffic flows in real-world environments. However, existing GAD methods are generally designed for centralized training, thereby posing privacy leakage risks. Despite progress, current mainstream Federated Graph anomaly detection (FGAD) methods still face challenges. A key limitation is that most existing approaches focus on node-level analysis while disregarding inter-node relationships making them ineffective against sophisticated attacks. To tackle this problem, this paper presents an Edge Federated Graph Anomaly Detection (E-FGAD) framework for IoT environments that combines centralized self-supervised pre-training with distributed supervised learning over edge embeddings. During the supervised phase, parameters are optimized in a federated manner using FedAvg with FedProx and server momentum. Our framework operates on graphs, where nodes represent IP endpoints and edges represent NetFlow records with traffic statistics as features, aiming to capture the flow of interactions between entities. We evaluate experiments on two real-world datasets, NF-BoT-IoT-v3 and NF-ToN-IoT-v3, in both binary and multi-class settings. E-FGAD achieves a maximum detection accuracy of 99.32%, a Macro-F1 of 89.73%, and a Weighted-F1 of 99.2%. Our framework outperforms centralized and federated baselines, demonstrating its effectiveness in detecting attacks while preserving privacy.
Nuha A. Hamad, Khairul Azmi Abu Bakar, Faizan Qamar et al.· IEEE Access· 0 citations
Introduction Breast cancer remains one of the leading causes of cancer-related mortality rate worldwide, and the identification of effective drug combinations is an essential requirement in pharmaceutical research. The integration of Artificial Intelligence (AI) in processing large volumes of chemical and biological data combines molecular representation, predictive modeling and structured support within a single accessible tool, which accelerates early-stage candidate identification for breast cancer research while promoting reproducibility, transparency and user centered design. Aim The current research focuses on developing and designing “PredictRx” which is an artificial intelligence based driven decision support tool which tends to benefit healthcare practioners to analyze the combination of drug which can be utilized for breast cancer patients. Methodology PredictRx was developed using molecular descriptors, physicochemical properties, and drug interaction datasets collected from publicly available biomedical databases. The tool integrates in total six supervised and unsupervised learning techniques to examine the structural similarities between compounds and predict the potential drug interactions for breast cancer. Various machine learning techniques, including Random Forest, Support Vector Machine, Logistic Regression, K-Means Clustering, DBSCAN, and Agglomerative Clustering, to analyse structural similarities and predict potential drug interactions and synergy patterns. Model performance was evaluated using Classification matrix, Silhouette Score, Calinski-Harabasz Index, and Davies-Bouldin Index. The tool was deployed as a browser-accessible web application for real-time interaction and visualization. Result The results suggests that Random Forest has the highest predictive performance accuracy of 1, and Agglomerative clustering delivered strongest scores (Silhouette Score: 0.6946; Davies-Bouldin Index: 0.2457). The current tool was deployed as a browser accessible web tool with possibility of real time interaction and result visualization. PredictRx is a distinctive easy to use, and interpretable screening tool focused on drug compatibility and synergy analysis. EDA further identified molecular weight, lipophilicity, and structural similarity as important contributors to drug compatibility prediction. Conclusion PredictRx shows how AI-driven predictive modeling which can speed up molecular screening and early-stage breast cancer medication discovery. The technology facilitates the effective identification of appropriate drug combinations and offers a scalable foundation for upcoming AI-assisted pharmaceutical research by combining clustering, classification, molecular representation, and visualization into a single interpretable platform.
Ritu Chauhan, Neha Pandey, M. Zuhairi· Frontiers in Artificial Inte...· 0 citations