SNN-SC: A Spiking Semantic Communication Framework for Collaborative Intelligence

TL;DR

This paper introduces SNN-SC, a spiking neural network-based semantic communication framework for collaborative AI inference on edge devices, reducing transmission overhead and improving robustness over noisy channels.

Contribution

The paper proposes a novel SNN-based semantic communication model that directly transmits binary semantic features, enhancing efficiency and robustness compared to traditional DNN-based methods.

Findings

Higher compression ratio achieved.

Overcomes the cliff effect in noisy channels.

Lower computational complexity than DNN-based models.

Abstract

Collaborative Intelligence (CI) has emerged as a promising framework for deploying Artificial Intelligence (AI) models on resource-constrained edge devices. In CI, the AI model is partitioned between the edge device and the cloud, with intermediate features transmitted from the edge sub-model to the cloud sub-model to complete the inference task. However, reducing feature transmission overhead while maintaining task performance remains a challenge, particularly in the case of noisy wireless channels. In this paper, we propose a Spiking Neural Network (SNN)-based Semantic Communication (SC) model, SNN-SC, which extracts compact semantic information from features and transmits it through digital binary channels. Compared to the Deep Neural Network (DNN)-based SC model, whose output is floating-point, the binary output of SNN makes SNN-SC directly applicable to digital binary channels without the need for extra quantization. Moreover, we introduce a novel spiking neuron called IHF to enhance the reconstruction capability of the SNN-SC decoder. Finally, we enhance the performance of SNN-SC by maximizing the entropy of semantic information. SNN-SC achieves a higher compression ratio and overcomes the `cliff effect' compared to the traditional separate source and channel coding method. In addition, SNN-SC has lower computational complexity than the DNN-based SC model and maintains higher task performance under poor channel conditions.