Deep Semantic Inference over the Air: An Efficient Task-Oriented Communication System

2025-08-18

AI Summary

In task-oriented wireless communications, it is challenging to simultaneously achieve high classification accuracy, low computational latency, and minimal communication overhead. Method: This paper proposes a joint optimization framework integrating model splitting and semantic feature compression. Specifically, the ResNet backbone is partitioned between edge and cloud; only lightweight semantic feature vectors are uploaded, coupled with a channel-adaptive compression strategy to enable over-the-air semantic inference on CIFAR-10/100. Contribution/Results: Under ≥85% of baseline accuracy, the framework reduces edge computational load by up to 62% and communication overhead by over 70% compared to end-to-end transmission. Crucially, this work establishes, for the first time, the quantitative trade-off among semantic feature dimensionality, channel conditions, and task accuracy—providing a scalable design paradigm for low-latency, low-overhead intelligent semantic wireless communications.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Search and Optimization: Learning to Search
• Computer Vision: Learning & Optimization for CV

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

Empowered by deep learning, semantic communication marks a paradigm shift from transmitting raw data to conveying task-relevant meaning, enabling more efficient and intelligent wireless systems. In this study, we explore a deep learning-based task-oriented communication framework that jointly considers classification performance, computational latency, and communication cost. We adopt ResNets-based models and evaluate them on the CIFAR-10 and CIFAR-100 datasets to simulate real-world classification tasks in wireless environments. We partition the model at various points to simulate split inference across a wireless channel. By varying the split location and the size of the transmitted semantic feature vector, we systematically analyze the trade-offs between task accuracy and resource efficiency. Experimental results show that, with appropriate model partitioning and semantic feature compression, the system can retain over 85% of baseline accuracy while significantly reducing both computational load and communication overhead.

Problem

• Optimizing task-oriented communication for efficiency and accuracy
• Balancing classification performance with computational and communication costs
• Exploring model partitioning to reduce resource use while maintaining accuracy

Innovation

• Deep learning-based task-oriented communication framework
• ResNets models for semantic feature extraction
• Model partitioning for accuracy-resource trade-offs