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

TL;DR

This work tackles resilience of neural MAC protocols (NPM) to environmental shifts that alter network conditions, such as more active UEs. It introduces LLM-based semantic MAC frameworks: TPM for immediate, prompt-driven signaling; TextGrad-aided TPM to automatically refine instructions; T2NPM to distill TPM knowledge into a smaller NPM and accelerate re-training; and T3NPM to combine TPM and T2NPM with MixSwitch to optimize the transition. A new meta-resilience metric is defined to evaluate resilience across a range of unknown target goodputs, enabling fair comparisons and transition optimization. Results show that T3NPM achieves the highest meta-resilience, surpassing NPM, TPM, T2NPM, and S-ALOHA, while dramatically reducing inference costs compared to TPM. These findings support the practical value of integrating semantic reasoning with data-driven MAC and KD-based transfer for robust, real-time network control.

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.

Figures (12)

• Figure 1: LLM-empowered semantic MAC protocols.
• Figure 2: Network model under environmental shifts.
• Figure 3: A schematic illustration of NPM.
• Figure 4: A schematic illustration of TPM.
• Figure 5: TPM example.