With the advent of 6G era, technologies integrating artificial intelligence and network communication have emerged as pivotal forces in shaping the future of connectivity. Among these innovative technologies, semantic communication stands out as it focuses on transmitting semantic representations rather than raw data sequences, thereby enhancing scalability, network efficiency, and performance. In this study, we propose a novel semantic communication framework utilizing Bayesian Rein-forcement Learning. Our proposed framework takes into account the relationship between the receiver (i.e., the edge servers) with robust computing capabilities and extensive knowledge bases, and the sender (i.e., the user device) with limited computing power and knowledge repositories, offloading the computational burden to the receiver. The receiver then proceeds with learning, considering the uncertainty of the model. This proposed framework is intended for application in semantic communication, which is not yet in use in actual communication systems, and has potential applications in areas such as edge computing and IoT.
This work was partially supported by the National Research Foundation of Korea (NRF) grant funded by the Ministry of Science and ICT (MSIT) (NRF-2020R1A2C1102284 and NRF-2021R1A2C1012776), and also supported by the BK21 FOUR program of the NRF of Korea funded by the Ministry of Education (NRF-5199991514504).