Secure Semantic Consensus as the Control Plane of Federated Agentic Networks

Bashir A. K., YAHYA H., Wu H., Pei J.

IEEE Communications Standards Magazine, 2026 (ESCI, Scopus)

Özet

Federated agentic networks are emerging in semantic communication systems, where distributed agents exchange task-level meanings, intentions, and reasoning outcomes instead of raw data or model parameters. However, heterogeneous agents may generate conflicting interpretations, leak sensitive context, or submit misleading but syntactically valid semantic states. This article introduces Secure Semantic Consensus (SSC) as a standardizable control-plane abstraction for federated agentic networks. SSC coordinates protected semantic states through reporting, privacy-preserving filtering, semantic alignment, trust-aware consensus, integrity verification, and feedback-based conflict resolution. By separating local perception and reasoning from network-level semantic coordination, SSC enables interoperable, privacy-preserving, and trustworthy collaboration without transmitting raw observations, private memory traces, or full model parameters. The article further discusses representative edge use cases and identifies standardization challenges for semantic-state description, trust evidence, feedback procedures, lightweight deployment, and cross-domain governance.