A semantic communication-based workload-adjustable transceiver for wireless AI-generated content (AIGC) delivery

TL;DR

This work addresses delivering high-quality AIGC content over wireless networks amid unstable channels and uneven resources. It introduces ROUTE, a semantic-communication-enabled, workload-adjustable transceiver that couples edge diffusion-based AIGC generation with cooperative edge-local decoding, allowing dynamic adjustment of semantic density and computing load. Key contributions include edge semantic extraction and diffusion-based generation, a modified receiver fine-tuning diffusion to mitigate channel noise, cloud-based pre-training with modular updates, and a learning-based policy for denoising-step allocation; results show improved latency and image quality versus baselines. The approach enables robust, low-latency AIGC delivery in dynamic wireless environments by leveraging SemCom for reduced bandwidth and diffusion models for flexible computation.

Abstract

With the significant advances in generative AI (GAI) and the proliferation of mobile devices, providing high-quality AI-generated content (AIGC) services via wireless networks is becoming the future direction. However, the primary challenges of AIGC service delivery in wireless networks lie in unstable channels, limited bandwidth resources, and unevenly distributed computational resources. In this paper, we employ semantic communication (SemCom) in diffusion-based GAI models to propose a Resource-aware wOrkload-adjUstable TransceivEr (ROUTE) for AIGC delivery in dynamic wireless networks. Specifically, to relieve the communication resource bottleneck, SemCom is utilized to prioritize semantic information of the generated content. Then, to improve computational resource utilization in both edge and local and reduce AIGC semantic distortion in transmission, modified diffusion-based models are applied to adjust the computing workload and semantic density in cooperative content generation. Simulations verify the superiority of our proposed ROUTE in terms of latency and content quality compared to conventional AIGC approaches.

Figures (4)

• Figure 1: The framework of proposed ROUTE.
• Figure 2: The diffusion and reverse diffusion processes.
• Figure 3: The delivered images of the three frameworks under different SNRs (text input: "A cute furry cat").
• Figure 4: The average latency for fetching one image.