An N-ary-Based Reliable Semantic Communication System With a Correction and Reasoning Approach

Semantic communication systems enhance communication efficiency and task performance by understanding the semantics of transmitted information, offering broad application prospects. Nevertheless, existing semantic communication systems struggle to handle information with complex relational semantics. Motivated by this limitation, this article proposes the n-ary relation semantic communication system with correction and reasoning (NRSC-CR). In NRSC-CR, semantic information is extracted from transmitted texts into n-ary relations. To enhance robustness against wireless channel influence, a correction algorithm based on the rationality of primary triplets and auxiliary information pairs (CPA) is introduced at the receiver. Furthermore, by deeply learning the interactions within and between n-ary relations, a reasoning algorithm based on global and local dependency (RGL) is proposed to address the performance degradation caused by excessive influence under low signal-to-noise ratio (SNR) conditions. Extensive simulation results show that the proposed NRSC-CR has advantages in understanding complex relational semantics, channel robustness, and data transmission efficiency.