Lightweight Task-Oriented Semantic Communication Empowered by Large-Scale AI Models

2025-06-16

IEEE Transactions on Vehicular Technology

AI Summary

To address the high computational overhead of large models and the slow, channel-agnostic inference of standard knowledge distillation (KD) in task-oriented semantic communication, this paper proposes a channel-aware fast knowledge distillation framework. Our method introduces three key innovations: (1) a pre-stored compression mechanism that eliminates redundant inference; (2) a channel-adaptive module enabling dynamic semantic adjustment based on real-time channel conditions; and (3) an information-bottleneck-driven loss function that jointly optimizes semantic fidelity and channel robustness. Experiments demonstrate that the proposed approach achieves comparable task accuracy while reducing model size by 3.2×, decreasing inference latency by 67%, and cutting training data requirements by 45%. It significantly outperforms existing KD and semantic communication baselines in efficiency, adaptability, and resource efficiency.

Technology Category

• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Multiagent Systems: Agent Communication
• Machine Learning: Learning on the Edge & Model Compression

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Recent studies have focused on leveraging large-scale artificial intelligence (LAI) models to improve semantic representation and compression capabilities. However, the substantial computational demands of LAI models pose significant challenges for real-time communication scenarios. To address this, this paper proposes utilizing knowledge distillation (KD) techniques to extract and condense knowledge from LAI models, effectively reducing model complexity and computation latency. Nevertheless, the inherent complexity of LAI models leads to prolonged inference times during distillation, while their lack of channel awareness compromises the distillation performance. These limitations make standard KD methods unsuitable for task-oriented semantic communication scenarios. To address these issues, we propose a fast distillation method featuring a pre-stored compression mechanism that eliminates the need for repetitive inference, significantly improving efficiency. Furthermore, a channel adaptive module is incorporated to dynamically adjust the transmitted semantic information based on varying channel conditions, enhancing communication reliability and adaptability. In addition, an information bottleneck-based loss function is derived to guide the fast distillation process. Simulation results verify that the proposed scheme outperform baselines in term of task accuracy, model size, computation latency, and training data requirements.

Problem

• Reduce LAI model complexity for real-time communication
• Improve distillation efficiency with pre-stored compression mechanism
• Enhance communication reliability via channel adaptive module

Innovation

• Knowledge distillation reduces LAI model complexity
• Pre-stored compression mechanism enhances distillation efficiency
• Channel adaptive module improves communication reliability