arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.13628cs.NI
收藏

紧凑型大语言模型部署与世界模型辅助的卸载在移动边缘计算中

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

Nanyang Technological University(南洋理工大学);Guangdong University of Technology(广东工业大学);Queen’s University Belfast(贝尔法斯特女王大学);Auburn University(奥本大学)

Ruichen Zhang, Xiaofeng Luo, Jiayi He, Jiawen Kang, Zehui Xiong, Shiwen Mao

AI总结:

本文研究了移动边缘计算网络中紧凑型大语言模型(LLM)部署和世界模型辅助的推断卸载问题,提出了一种边缘紧凑型LLM部署(ECLD)框架,结合结构化剪枝、低比特量化和知识蒸馏来构建可部署于边缘的LLM变种,并通过四个互补指标评估这些模型。基于这些紧凑模型,提出了一个MEC卸载优化问题,旨在最小化长期平均推断延迟,同时满足设备能耗预算和LLM特定的服务质量约束。为了解决未知且时间变化的网络动态问题,开发了一种世界模型-近端策略优化(PPO)算法,该算法在原策略优化算法中加入了学习的递归世界模型,以提供改进的价值目标和短时间想象回放。

摘要:

本文研究了移动边缘计算网络中紧凑型大语言模型(LLM)部署和世界模型辅助的推断卸载问题。我们首先提出了一种边缘紧凑型LLM部署(ECLD)框架,该框架联合应用结构化剪枝、低比特量化和知识蒸馏,以构建可部署于边缘的LLM变种,并通过四个互补指标评估这些模型:可访问性、能耗、幻觉率和泛化准确性。基于这些紧凑模型,我们提出了一个MEC卸载优化问题,该问题旨在最小化长期平均推断延迟,同时满足每设备能耗预算和LLM特定的服务质量约束(即有效准确性和幻觉率)。为了解决未知且时间变化的网络动态问题,我们开发了一种世界模型-近端策略优化(PPO)算法,该算法在原策略优化算法中加入了学习的递归世界模型,以提供改进的价值目标和短时间想象回放。在Llama-3.1-8B、Qwen3-8B和Mistral-12B上的大量实验表明,ECLD将基础模型的存储压缩了约70-80%(例如,Llama-3.1-8B从15.3 GB压缩到3.3 GB),并将每个查询的能耗减少了高达50%,同时在很大程度上保持了准确性,经常降低了幻觉率,与仅量化或仅剪枝的基线相比。此外,这些实验还表明,世界模型-PPO在收敛速度上加快了约50%,在与原生PPO相比,最终奖励提高了15.8%,在不同用户群体中平均推断延迟减少了12-30%,同时满足准确性和幻觉约束,并接近始终卸载的质量,同时具有局部执行的大部分效率。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑