跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2607.05114 · v2

Localized LoRA-MoE: Mixture of Low-Rank Experts With Adaptive Routing

局部化LoRA-MoE:自适应路由的低秩专家混合

首次提交 2026-07-06 22:06
版本更新 2026-11-18 02:47
雷达观测 2026-11-18 12:34
arXiv ↗PDF ↗

摘要

Parameter-efficient fine-tuning (PEFT) methods such as LoRA adapt large models through a single low-rank update. Such updates are susceptible to gradient interference: when a training stream interleaves multiple operational contexts, conflicting updates to shared adapter weights collapse the adapter into an unspecialized average. While recent methods introduce block-wise isolation to remedy this issue, their topologies are static and cannot adapt to context switching or localized sensor degradation. In this work, we introduce Localized LoRA-MoE, a framework that combines localized spatial blocking with context-conditioned expert routing. We propose and evaluate two architectural variants: Block-Wise LoRA-MoE (centralized routing), which modulates the entire structural grid through a single context-conditioned gate, and Cell-Wise LoRA-MoE (decentralized routing), which assigns an independent gate to every coordinate cell of the matrix grid. Across three structured transformation benchmarks--a high-dimensional SVD matrix simulation, context-switched transformations over real tabular data, and spatial reconstruction under simulated sensor occlusion--both variants outperform static baselines under comparable adapter budgets. Both variants perform similarly when context shifts the grid uniformly, while cell-wise routing is strictly better when contexts are spatially heterogeneous. When fine-tuning a language model on six interleaved GLUE tasks, task-conditioned routing at a fixed adapter budget outperforms LoRA and static localized adapters, matching per-task adapters with 6x budgets. At decoder scale on Qwen2.5-1.5B, both routed variants outperform static baselines across eight commonsense benchmarks, with cell-wise routing achieving +3.9 points over LoRA at matched active capacity.

AI 论文解读

基于摘要 · 非全文审阅

提出Localized LoRA-MoE框架,结合局部空间分块与上下文条件专家路由,以缓解LoRA在交错多上下文训练中的梯度干扰问题。

正在读取账号与研究记录…

版本差异

本站首次收录为 v2,暂无更早版本可对照。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v2 · 2026-11-18 02:47 · 最新观测版本

Localized LoRA-MoE: Mixture of Low-Rank Experts With Adaptive Routing

Parameter-efficient fine-tuning (PEFT) methods such as LoRA adapt large models through a single low-rank update. Such updates are susceptible to gradient interference: when a training stream interleaves multiple operational contexts, conflicting updates to shared adapter weights collapse the adapter into an unspecialized average. While recent methods introduce block-wise isolation to remedy this issue, their topologies are static and cannot adapt to context switching or localized sensor degradation. In this work, we introduce Localized LoRA-MoE, a framework that combines localized spatial blocking with context-conditioned expert routing. We propose and evaluate two architectural variants: Block-Wise LoRA-MoE (centralized routing), which modulates the entire structural grid through a single context-conditioned gate, and Cell-Wise LoRA-MoE (decentralized routing), which assigns an independent gate to every coordinate cell of the matrix grid. Across three structured transformation benchmarks--a high-dimensional SVD matrix simulation, context-switched transformations over real tabular data, and spatial reconstruction under simulated sensor occlusion--both variants outperform static baselines under comparable adapter budgets. Both variants perform similarly when context shifts the grid uniformly, while cell-wise routing is strictly better when contexts are spatially heterogeneous. When fine-tuning a language model on six interleaved GLUE tasks, task-conditioned routing at a fixed adapter budget outperforms LoRA and static localized adapters, matching per-task adapters with 6x budgets. At decoder scale on Qwen2.5-1.5B, both routed variants outperform static baselines across eight commonsense benchmarks, with cell-wise routing achieving +3.9 points over LoRA at matched active capacity.

查看此版本解读 →查看此版本原文 ↗