English(EN) FED-FSTQ: Fisher-Guided Token Quantization for Communication-Efficient Federated Fine-Tuning of LLMs on Edge Devices

Fed-FSTQ 将边缘设备上的LLM微调流量减少了46倍

作者 PulseAugur 编辑部 · [2 个来源] · 2026-04-28 09:29

研究人员开发了Fed-FSTQ，一种用于在边缘设备上高效进行大型语言模型（LLM）联邦微调的新系统。该方法使用Fisher代理来指导令牌量化，优先处理重要信息并减少冗余传输。Fed-FSTQ旨在做到模型无关，并兼容现有的联邦学习管道（如LoRA），支持带宽异构的客户端。实验表明，上行流量显著减少，达到准确度的时间缩短，并在边缘硬件上具有潜在的速度提升。 AI

影响降低了边缘设备上联邦LLM微调的通信开销，实现了更高效的设备端适应。

排序理由介绍LLM微调新方法的学术论文。

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 2 个来源。我们如何撰写摘要 →

报道来源 [2]

arXiv cs.LG TIER_1 English(EN) · Changyu Li, Shuanghong Huang, Jiashen Liu, Ming Lei, Jidu Xing, Kaishun Wu, Lu Wang, Fei Luo · 2026-04-29 04:00

FED-FSTQ: Fisher-Guided Token Quantization for Communication-Efficient Federated Fine-Tuning of LLMs on Edge Devices

arXiv:2604.25421v1 Announce Type: new Abstract: Federated fine-tuning provides a practical route to adapt large language models (LLMs) on edge devices without centralizing private data, yet in mobile deployments the training wall-clock is often bottlenecked by straggler-limited u…
arXiv cs.AI TIER_1 English(EN) · Fei Luo · 2026-04-28 09:29

FED-FSTQ: Fisher-Guided Token Quantization for Communication-Efficient Federated Fine-Tuning of LLMs on Edge Devices

Federated fine-tuning provides a practical route to adapt large language models (LLMs) on edge devices without centralizing private data, yet in mobile deployments the training wall-clock is often bottlenecked by straggler-limited uplink communication under heterogeneous bandwidt…

报道来源 [2]

FED-FSTQ: Fisher-Guided Token Quantization for Communication-Efficient Federated Fine-Tuning of LLMs on Edge Devices

FED-FSTQ: Fisher-Guided Token Quantization for Communication-Efficient Federated Fine-Tuning of LLMs on Edge Devices

相关实体

相关话题