A Survey on Semantic Communication Networks: Architecture, Security, and Privacy

TL;DR

This survey presents SemComNet, a three-layer architecture (control, semantic transmission, cognitive sensing) for multi-agent semantic communications. It identifies three working modes (paired, clustered, networked), outlines enabling technologies, and discusses use cases and semantic metrics. A comprehensive taxonomy of security and privacy threats across layers is provided, along with state-of-the-art defenses spanning semantic model security, transmission security, reliability, trust management, and data/knowledge protection. The paper culminates with actionable future directions—green designs, explainability, endogenous security, and adaptive architectures—to enable robust, private, and scalable SemComNet in 6G and beyond.

Abstract

With the rapid advancement and deployment of intelligent agents and artificial general intelligence (AGI), a fundamental challenge for future networks is enabling efficient communications among agents. Unlike traditional human-centric, data-driven communication networks, the primary goal of agent-based communication is to facilitate coordination among agents. Therefore, task comprehension and collaboration become the key objectives of communications, rather than data synchronization. Semantic communication (SemCom) aims to align information and knowledge among agents to expedite task comprehension. While significant research has been conducted on SemCom for two-agent systems, the development of semantic communication networks (SemComNet) for multi-agent systems remains largely unexplored. In this paper, we provide a comprehensive and up-to-date survey of SemComNet, focusing on their fundamentals, security, and privacy aspects. We introduce a novel three-layer architecture for multi-agent interaction, comprising the control layer, semantic transmission layer, and cognitive sensing layer. We explore working modes and enabling technologies, and present a taxonomy of security and privacy threats, along with state-of-the-art defense mechanisms. Finally, we outline future research directions, paving the way toward intelligent, robust, and energy-efficient SemComNet. This survey represents the first comprehensive analysis of SemComNet, offering detailed insights into its core principles as well as associated security and privacy challenges.

Figures (11)

• Figure 1: Illustration of paired semantic communication and networked semantic communication.
• Figure 2: Organization structure of this paper.
• Figure 3: Architecture of the SemComNet.
• Figure 4: SemComNet working modes. (a) Paired SemCom: interaction between two agents via SemCom. (b) Clustered SemCom: each agent collaborates with others via paired SemCom within a cluster for a common task. (c) Networked SemCom: multiple clustered agents interact for different tasks, where intra-cluster agents interact via clustered SemCom and inter-cluster communication may assisted by semantic relay.
• Figure 5: Illustration of the construction of cross-modal knowledge graph (CKG) in li2022crossmodal, which involves four key steps: 1) data collection and preprocessing, 2) multi-modal knowledge extraction, 3) cross-modal knowledge fusion that combines extracted knowledge and existing CKG, and 4) information storage and retrieval, supported by graph databases (e.g., neo4j).