Knowledge-Enhanced Network Semantic Communication Based on Multi-Objective Optimization

Semantic communication represents a paradigm shift from traditional bit-level to meaning-centric wireless communication. This paper introduces Knowledge-Enhanced Network (KE-Net) semantic communication, a novel framework that leverages external knowledge graphs (KGs) to improve semantic transmission accuracy over wireless channels. The proposed approach employs a multi-objective optimization strategy combining semantic loss, knowledge integration loss, reconstruction loss, and channel-aware loss components. Numerical results show the superior performance of KE-Net semantic communication over competing techniques.