Title:Semantic Information Retrieval in Wireless Networks

Abstract:Motivated by recent success of Machine Learning (ML) tools in wireless communications, the idea of semantic communication by Weaver from 1949 has received considerable attention. It breaks with the classic design paradigm of Shannon by aiming to transmit the meaning of a message, i.e., semantics, rather than its exact copy and thus allows for savings in channel uses or information rate. In this work, we extend the fundamental approach from Basu et al. for modeling semantics from logical to probabilistic entailment relations between meaning and messages. Thus, we model semantics by means of a hidden random variable and define the task of semantic communication as transmission of messages over a communication channel such that semantics is best preserved. We formulate the semantic communication design either as an Information Maximization or as an Information Bottleneck optimization problem. Finally, we propose the ML-based semantic communication system SINFONI for a distributed multipoint scenario: SINFONI communicates the meaning behind multiple messages that are observed at different senders to a single receiver for semantic retrieval. We analyze SINFONI by processing images as an example of messages. Numerical results reveal a tremendous rate normalized SNR shift up to 20 dB compared to classically designed communication systems.