Generative AI Driven Task-Oriented Adaptive Semantic Communications

Yuzhou Fu

doi:10.1109/TWC.2025.3644355

Return

Generative AI Driven Task-Oriented Adaptive Semantic Communications

2026-01-01

PRE

DOI：10.1109/TWC.2025.3644355

Original

Original request for help

Save

Abstract

En 中文

Task-Oriented Semantic Communication (TOSC) has been regarded as a promising communication framework, serving for various Artificial Intelligence (AI) task driven applications. The existing TOSC frameworks focus on extracting the full semantic features of source data and learning low-dimensional channel inputs to transmit them within limited bandwidth resources. Although transmitting full semantic features can preserve the integrity of data meaning, this approach does not attain the performance threshold of the TOSC. In this paper, we propose a Task-oriented Adaptive Semantic Communication (TasCom) framework to effectively facilitate the inference of different AI tasks. Based on the Generative AI (GAI) techniques, we first propose a Joint Source-Channel Coding (JSCC) that which only extracts and fuses task-related semantic features, and then transmits them to achieve efficient task-oriented semantic transmission. Then, we propose a generative training algorithm to train the proposed JSCC for optimal performance. Furthermore, an Adaptive Coding Controller (ACC) is proposed to find the optimal coding scheme for the proposed JSCC, which allows the semantic features with significant contributions to the task inference to preferentially occupy limited bandwidth resources for wireless transmission. The simulation results show that the proposed TasCom outperforms the existing TOSC and traditional codec schemes on the object detection and instance segmentation tasks under all considered channel conditions.

Keywords:

Task-oriented semantic communication

semantic extraction

joint source-channel coding

generative AI