Seq2Seq-SC: End-to-End Semantic Communication Systems with Pre-trained Language Model
This work addresses communication efficiency for future wireless systems like 6G, but appears incremental as it builds on existing semantic communication approaches with pre-trained models.
The authors tackled the problem of communication efficiency in semantic networks by proposing seq2seq-SC, a system compatible with 5G NR that uses a pre-trained language model for generalized text datasets, and demonstrated that it outperforms previous models in extracting semantically meaningful information while maintaining superior performance.
In this work, we propose a realistic semantic network called seq2seq-SC, designed to be compatible with 5G NR and capable of working with generalized text datasets using a pre-trained language model. The goal is to achieve unprecedented communication efficiency by focusing on the meaning of messages in semantic communication. We employ a performance metric called semantic similarity, measured by BLEU for lexical similarity and SBERT for semantic similarity. Our findings demonstrate that seq2seq-SC outperforms previous models in extracting semantically meaningful information while maintaining superior performance. This study paves the way for continued advancements in semantic communication and its prospective incorporation with future wireless systems in 6G networks.