Resilient LLM-Empowered Semantic MAC Protocols via Zero-Shot Adaptation and Knowledge Distillation

TL;DR

This paper introduces resilient semantic MAC protocols powered by large language models, combining prompt-based adaptation, knowledge distillation, and phase switching to improve robustness and efficiency in dynamic network environments.

Contribution

It proposes three novel LLM-empowered MAC frameworks, including TPM, T2NPM, and T3NPM, with a new meta-resilience metric for environmental shift adaptation.

Findings

T3NPM achieves 20.56% higher meta-resilience.

T3NPM reduces computation cost by 19.8x compared to TPM.

Simulations validate improved robustness and efficiency.

Abstract

Neural network-based medium access control (MAC) protocol models (NPMs) improve goodput through site-specific operations but are vulnerable to shifts from their training network environments, such as changes in the number of user equipments (UEs) severely degrade goodput. To enhance resilience against such environmental shifts, we propose three novel semantic MAC protocol frameworks empowered by large language models (LLMs). First, we introduce a token-based protocol model (TPM), where an LLM generates MAC signaling messages. By editing LLM instruction prompts, TPM enables instant adaptation, which can be further enhanced by TextGrad, an LLM-based automated prompt optimizer. TPM inference is fast but coarse due to the lack of real interactions with the changed environment, and computationally intensive due to the large size of the LLM. To improve goodput and computation efficiency, we develop T2NPM, which transfers and augments TPM knowledge into an NPM via knowledge distillation (KD). Integrating TPM and T2NPM, we propose T3NPM, which employs TPM in the early phase and switches to T2NPM later. To optimize this phase switching, we design a novel metric of meta-resilience, which quantifies resilience to unknown target goodput after environmental shifts. Simulations corroborate that T3NPM achieves 20.56% higher meta-resilience than NPM with 19.8x lower computation cost than TPM in FLOPS.