Compact LLM Deployment and World Model Assisted Offloading in Mobile Edge Computing

TL;DR

This paper presents a framework for deploying compact LLMs on mobile edge devices using pruning, quantization, and knowledge distillation, and introduces a world model-based offloading algorithm to optimize inference latency and quality.

Contribution

It introduces an integrated approach combining model compression and a novel world model-PPO algorithm for efficient LLM inference offloading in MEC networks.

Findings

Model compression reduces storage by 70-80% and energy use by 50%.

World model-PPO accelerates convergence and improves reward by 15.8%.

Inference latency decreases by 12-30%, maintaining accuracy and reducing hallucinations.

Abstract

This paper investigates compact large language model (LLM) deployment and world-model-assisted inference offloading in mobile edge computing (MEC) networks. We first propose an edge compact LLM deployment (ECLD) framework that jointly applies structured pruning, low-bit quantization, and knowledge distillation to construct edge-deployable LLM variants, and we evaluate these models using four complementary metrics: accessibility, energy consumption, hallucination rate, and generalization accuracy. Building on the resulting compact models, we formulate an MEC offloading optimization problem that minimizes the long-term average inference latency subject to per-device energy budgets and LLM-specific quality-of-service constraints on effective accuracy and hallucination. To solve this problem under unknown and time-varying network dynamics, we develop a world model-proximal policy optimization (PPO) algorithm, which augments an on-policy PPO algorithm with a learned recurrent world model that provides improved value targets and short imagination rollouts. Extensive experiments on Llama-3.1-8B, Qwen3-8B, and Mistral-12B show that ECLD compresses base models by about 70-80% in storage (i.e., from 15.3 GB to 3.3 GB for Llama-3.1-8B) and reduces per-query energy consumption by up to 50%, while largely preserving accuracy and often lowering hallucination compared with quantization-only or pruning-only baselines. Moreover, they also show that world model-PPO speeds up convergence by about 50%, improves the final reward by 15.8% over vanilla PPO, and reduces average inference latency by 12-30% across different user populations, while satisfying the accuracy and hallucination constraints and approaching the generation quality of always-offloading with much of the efficiency of local execution.