Answers are grounded in this paper and cite their sources.
The increase in the rate of federated edge ecosystems growth has enlarged the targets of Advanced Persistent Threats (APTs), which penetrate distributed and heterogeneous infrastructures in stealthy, multi-stage attacks.In order to overcome the shortcomings of centralised and reactive intrusion-detection systems, the research presents the adaptive cognitive graph fusion (ACGF) framework, a privacy-preserving, neuro-symbolic and federated method of proactive APT defence.ACGF employs Graph Neural Networks (GCNs) and Bidirectional LSTMs together with a Threat Intelligence Knowledge Graph (TIKG) in a Graph Contrastive Transformer (GCT) that aligns spatiotemporal representations with isolating semantic threat relations.Reinforcement-Learning-based Cognitive Trust Calibrator dynamically optimizes the contribution of federated edge nodes to withstand poisoned or unreliable updates, whereas Adaptive Computation Time (ACT) and Dynamic Model Pruning are optimizable and reduce the latency and energy usage of edge deployment.Tested on CIC-IDS2018, TON_IoT, and DAPT2020 data, ACGF attains an accuracy of up to 0.996, a macro-F1 of up to 0.996, and an AUROC of up to 0.998, which is higher than the state-of-the-art graph and federated baselines.The framework forecasts future APT schemes with 87.4% top-1 and 96.8% top-3 precision, and achieves a median model inference latency of 18.6 ms in the reported edge-device benchmark.These results reveal that neural and symbolic intelligence can have their cognitive fusion to allow interpretable, self-evolving, resourceefficient cyber defense of federated edge networks.
No takes yet. Share an insight, caveat, or question.
Masoody et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: