Task-Oriented Communication with Hybrid-Precision Models

2026-07-18

AI Summary

This work addresses the challenge of balancing communication efficiency, on-device computational overhead, and inference accuracy in edge intelligence systems. To this end, the authors propose a mixed-precision, task-oriented communication framework wherein edge devices employ a binarized front-end tailored for split inference to extract and transmit binary features, while an edge server executes full-precision back-end processing for final inference. The design incorporates subcarrier-level channel-aware feature calibration, end-to-end training via knowledge distillation, and a customized gradient estimator to effectively preserve semantic information and enable system-wide co-optimization. Experimental results on ImageNet demonstrate that the proposed approach achieves a superior trade-off among communication efficiency, computational cost, and accuracy, significantly outperforming existing methods.

Abstract

Edge inference has emerged as a promising solution for the proliferation of artificial intelligence (AI) services by deploying models at the network edge to circumvent cloud-routing latency. Existing edge inference approaches mainly focused on either cooperative inference to reduce latency or lightweight model design to fit resource-constrained devices. These solutions often address the communication and computation challenges separately, and thus struggle to achieve a balanced trade-off among transmission efficiency, on-device processing cost, and inference accuracy. To bridge this gap, this paper proposes a hybrid-precision task-oriented communication framework for edge inference to holistically balance communication, on-device computation, and utility. In this framework, a binarized front-end is deployed on the edge device to extract and transmit binary features via orthogonal frequency-division multiplexing (OFDM) signals, while a full-precision back-end on the edge server performs the final inference. To ensure model consistency, we introduce an on-device binarization method tailored for split inference and develop an integrated channel-aware transmission scheme featuring subcarrier-based feature calibration. Furthermore, a knowledge distillation (KD)-based training strategy, supported by specialized gradient estimators, is developed to optimize the end-to-end system and inherit semantic knowledge from a full-precision teacher model. Extensive experiments on the large-scale ImageNet dataset demonstrate the superiority of the proposed hybrid system. Our analysis confirms that this design achieves an optimal trade-off among communication efficiency, on-device computational cost, and inference accuracy, outperforming existing edge inference solutions.

Problem

• edge inference
• communication-computation trade-off
• task-oriented communication
• hybrid-precision models
• resource-constrained devices

Innovation

• hybrid-precision
• task-oriented communication
• edge inference
• knowledge distillation
• binary neural networks