Semantic Communication With Adaptive Universal Transformer

• DOI: 10.1109/lwc.2021.3132067
• Publication Date: 2021-12-02T20:26:10Z
• Authors: Qingyang Zhou; Rongpeng Li; Zhifeng Zhao; Chenghui Peng; Honggang Zhang
• Keywords: Signal Processing (eess.SP); FOS: Computer and information sciences; Computer Science - Computation and Language; FOS: Electrical engineering, electronic engineering, information engineering; 0202 electrical engineering, electronic engineering, information engineering; 02 engineering and technology; Electrical Engineering and Systems Science - Signal Processing; Computation and Language (cs.CL)

With the development of deep learning (DL), natural language processing (NLP) makes it possible for us to analyze and understand a large amount texts. Accordingly, we can achieve semantic communication in terms joint source channel coding over noisy with help NLP. However, existing method realize this goal is use fixed transformer NLP while ignoring difference information contained each sentence. To solve problem, propose new system based on Universal Transformer. Compared traditional transformer, an adaptive circulation mechanism introduced Through introduction mechanism, be more flexible transmit sentences different information, better end-to-end performance under various conditions.