Semantic Knowledge Base Synchronization by Limited Feedback Assistance
Qi Cao, Fei Wan, Yameng Du, Baoming Bai · 2023
To address the challenge of maintaining fast and reliable communication in the face of a staggering amount of data, a new communication paradigm called semantic communication has gained significant interest. Semantic communication is expected to play a crucial role in future communication systems. A major challenge in semantic communication is dealing with the mismatched knowledge base between the transmitter and receiver. To overcome this problem, this paper proposes a feedback-aided semantic communication system. In this system, the receiver can enhance semantic decoding performance through simple query and answer interactions with the transmitter. The interpretation of meaning at the knowledge base of the receiver could be updated based on the feedback received from the transmitter. This paper presents a theoretical analysis that establishes a relationship between the number of feedback and the correct semantic decoding probability. This paper also analyzes the performance of the greedy policy and identifies the conditions under which the greedy policy is the optimal policy. Simulation results validate this result.