Abstract:Deep learning (DL) based semantic communication methods have been explored for the efficient transmission of images, text, and speech in recent years. In contrast to traditional wireless communication methods that focus on the transmission of abstract symbols, semantic communication approaches attempt to achieve better transmission efficiency by only sending the semantic-related information of the source data. In this paper, we consider semantic-oriented speech transmission which transmits only the semantic-relevant information over the channel for the speech recognition task, and a compact additional set of semantic-irrelevant information for the speech reconstruction task. We propose a novel end-to-end DL-based transceiver which extracts and encodes the semantic information from the input speech spectrums at the transmitter and outputs the corresponding transcriptions from the decoded semantic information at the receiver. In particular, we employ a soft alignment module and a redundancy removal module to extract only the text-related semantic features while dropping semantically redundant content, greatly reducing the amount of semantic redundancy compared to existing methods. We also propose a semantic correction module to further correct the predicted transcription with semantic knowledge by leveraging a pretrained language model. For the speech to speech transmission, we further include a CTC alignment module that extracts a small number of additional semantic-irrelevant but speech-related information, such as duration, pitch, power and speaker identification of the speech for the better reconstruction of the original speech signals at the receiver. We also introduce a two-stage training scheme which speeds up the training of the proposed DL model. The simulation results confirm that our proposed method outperforms current methods in terms of the accuracy of the predicted text for the speech to text transmission and the quality of the recovered speech signals for the speech to speech transmission, and significantly improves transmission efficiency. More specifically, the proposed method only sends 16% of the amount of the transmitted symbols required by the existing methods while achieving about a 10% reduction in WER for the speech to text transmission. For the speech to speech transmission, it results in an even more remarkable improvement in terms of transmission efficiency with only 0.2% of the amount of the transmitted symbols required by the existing method while preserving the comparable quality of the reconstructed speech signals.

Task-Oriented Communication with Reliability-Driven Retransmission Request

Semantic-aware Speech to Text Transmission with Redundancy Removal

Semantic-preserved Communication System for Highly Efficient Speech Transmission

Latency-minimizing Semantic Communication with Dynamic Model Partitioning

Deep Learning Enabled Semantic Communication Systems for Video Transmission

Deep Learning-Enabled Semantic Communication Systems With Task-Unaware Transmitter and Dynamic Data

Deep Learning-Based Image Semantic Communication System

Task-oriented Explainable Semantic Communications

Joint URLLC Traffic Scheduling and Resource Allocation for Semantic Communication Systems

A Unified Multi-Task Semantic Communication System with Domain Adaptation

A Highly Reliable Encoding and Decoding Communication Framework Based on Semantic Information

Deep Learning Enabled Task-Oriented Semantic Communication for Memory-Limited Devices

Adaptive Modulation and Retransmission Scheme for Semantic Communication Systems

Knowledge Distillation Based Semantic Communications For Multiple Users

Rate Adaptive Mechanism for Semantic Communication Systems: A Robustness Verification Approach

Reliable Semantic Communication System Enabled by Knowledge Graph

A Robust Semantic Text Communication System

Deep Learning based Semantic Communications: An Initial Investigation

Distortion Resilience for Goal-Oriented Semantic Communication

Semantic Entropy Can Simultaneously Benefit Transmission Efficiency and Channel Security of Wireless Semantic Communications

Reinforcement Learning-powered Semantic Communication via Semantic Similarity