Deep Learning Enabled Semantic Communication Systems

Системы семантической связи на основе глубокого обучения
Zhijin Qin, Geoffrey Ye Li, Huiqiang Xie, Biing-Hwang Fred Juang
2021-01-01

DeepSCTransformersemantic communicationsentence similaritytransfer learning
Recently, deep learned enabled end-to-end communication systems have been developed to merge all physical layer blocks in the traditional communication systems, which make joint transceiver optimization possible. Powered by deep learning, natural language processing has achieved great success in analyzing and understanding a large amount of language texts. Inspired by research results in both areas, we aim to provide a new view on communication systems from the semantic level. Particularly, we propose a deep learning based semantic communication system, named DeepSC, for text transmission. Based on the Transformer, the DeepSC aims at maximizing the system capacity and minimizing the semantic errors by recovering the meaning of sentences, rather than bit- or symbol-errors in traditional communications. Moreover, transfer learning is used to ensure the DeepSC applicable to different communication environments and to accelerate the model training process. To justify the performance of semantic communications accurately, we also initialize a new metric, named sentence similarity. Compared with the traditional communication system without considering semantic information exchange, the proposed DeepSC is more robust to channel variation and is able to achieve better performance, especially in the low signal-to-noise (SNR) regime, as demonstrated by the extensive simulation results.
1
DeepSC is a Transformer-based semantic communication system that optimizes text transmission by recovering sentence meaning rather than minimizing bit or symbol errors.
2
Simulations show that DeepSC is more robust to channel variations and outperforms conventional systems, particularly in low-SNR conditions.
3
The authors introduce sentence similarity as a metric for evaluating semantic communication performance.
4
The system jointly optimizes communication components end-to-end to maximize capacity while reducing semantic errors in transmitted text.
5
Transfer learning enables DeepSC to adapt to different communication environments and accelerates model training.

DeepSC: a deep learning based semantic communication system for text transmission (Transformer-based end-to-end communication system)

Meaning-preserving text transmission, including semantic-error minimization, system capacity, robustness to channel variation, and performance in low-SNR conditions

Publication Details
Publication Date
2021-01-01
Journal
Publisher
ISSN
Cited by
1754
Access Type
Author Information
Authors
Zhijin Qin
Geoffrey Ye Li
Huiqiang Xie
Biing-Hwang Fred Juang
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%