English(EN) Multi-SPIN: Multi-Access Speculative Inference for Cooperative Token Generation at the Edge

Multi-SPIN 实现了边缘端的协同 LLM 令牌生成

作者 PulseAugur 编辑部 · [2 个来源] · 2026-06-03 08:16

研究人员开发了 Multi-SPIN，一种用于边缘端协同令牌生成的新型架构。该系统利用设备上较小的语言模型创建草稿令牌，然后由边缘服务器上较大的 LLM 并行验证。该方法旨在平衡资源受限设备和服务器之间的计算负载，优化草稿长度和带宽分配，以最大化整体令牌生成速度。 AI

影响优化了边缘设备的 LLM 推理，有望在协同生成场景中提高响应速度并降低服务器负载。

排序理由详细介绍新推理架构的学术论文。

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.AI TIER_1 English(EN) · Haotian Zheng, Zhanwei Wang, Mingyao Cui, Chang Cai, Hongyang Du, Kaibin Huang · 2026-06-04 04:00

Multi-SPIN：边缘协同生成令牌的多接入推测推理

arXiv:2606.04581v1 Announce Type: cross Abstract: Speculative inference (SPIN) was originally developed as an efficient architecture to accelerate Large Language Models (LLMs). In this work, we propose its distributed deployment to enable cooperative token generation in a multius…
arXiv cs.AI TIER_1 English(EN) · Kaibin Huang · 2026-06-03 08:16

Multi-SPIN：边缘协同生成令牌的多接入推测推理

Speculative inference (SPIN) was originally developed as an efficient architecture to accelerate Large Language Models (LLMs). In this work, we propose its distributed deployment to enable cooperative token generation in a multiuser edge system; its advantage is to effectively ba…