用记忆检索管理参数子空间，让模型持续学习时能快速恢复旧知识。
Neural Subspace Reallocation: Continual Learning as Retrieval-Based Subspace Memory Management
- 将LoRA模块视为可压缩、可检索的记忆单元，通过相似性召回旧知识。
- 在循环任务中恢复速度提升10倍，遗忘率比无记忆方法低9倍。
- 简单相似性检索已胜过复杂强化学习策略，证明记忆机制是关键。
我们提出神经子空间重分配（NSR），将持续学习重构为对参数子空间的内存管理问题。不同于将低秩适配（LoRA）模块作为一次性适配器，NSR将其视为可在冻结主干上压缩、检索的记忆单元，通过周期性流程：(1) 用SVD压缩已学的LoRA，(2) 存入任务知识库（TaskKnowledgeBank），(3) 通过嵌入相似性召回相关历史LoRA以热启动新或回归任务，(4) 相应重分配活跃子空间，同时通过蒸馏保护先验任务。我们证明，在循环环境中，任何无记忆的分配策略相对于有知识库支持的历史感知策略，累积遗憾至少为Ω(T(M−1)Δ_switch)（定理1）。实验表明，在Split-CIFAR-100上，知识库使循环恢复时间减少10倍，与预测一致；在异构5数据集基准上，NSR达到最高准确率且遗忘最少，其反向迁移接近零的程度比无记忆启发式方法快约9倍。关键控制实验显示：固定知识库，仅改变分配规则，基于相似性的检索规则在恢复循环任务上（0步对比1.8步）和精度上均优于或持平于学习型强化学习控制器。核心发现是：在固定容量下，性能提升主要来自记忆机制（压缩与相似性检索），而非学习型分配策略。内存预算分析证实，压缩后的知识库仅需每任务0.29 MB参数存储，通过保留前K项可控制总开销，同时保证被保留任务的快速恢复。
原文摘要 · Abstract (English)
We introduce Neural Subspace Reallocation (NSR), which reframes continual learning as memory management over parameter subspaces. Instead of treating Low-Rank Adaptation (LoRA) modules as disposable per-task adapters, NSR manages them as compressible, retrievable memory units on a frozen backbone through a recurring cycle: (1) compress learned LoRAs via SVD, (2) reserve them in a TaskKnowledgeBank, (3) recall related past LoRAs by embedding similarity to warm-start new or returning tasks, and (4) reallocate the active subspace accordingly, with distillation protecting prior tasks. We prove that in cyclic environments any memoryless allocation policy incurs cumulative regret Omega(T(M-1)Delta_switch) relative to a history-aware policy backed by the Bank (Theorem 1). Empirically, on Split-CIFAR-100 the Bank reduces cyclic recovery time by 10x, exactly as predicted, and on the heterogeneous 5-Datasets benchmark NSR achieves the highest accuracy and the least forgetting, about 9x closer to zero backward transfer than the memoryless heuristics. Crucially, we run a controlled study that isolates which component matters: holding the Bank fixed and varying only the allocation rule, we find that a simple similarity-based retrieval rule matches or beats a learned reinforcement-learning controller (recovering recurring tasks in 0 vs 1.8 steps and reaching equal accuracy). Our central, honest finding is therefore that the memory mechanism -- compression and similarity retrieval -- rather than a learned allocation policy, drives continual-learning performance under fixed capacity. A memory-budget analysis confirms the compressed Bank stays small -- 0.29 MB of parameter memory per task -- so a top-K retention cap bounds the total footprint while preserving fast recovery for retained tasks.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。