arXiv:2608.17530cs.AIcs.LG2026-08综述

用记忆规律重训语言模型,让旧知识不丢、新知识快学。

When to Review: Spaced Repetition for Continual Pre-Training of Language Models

  • 基于记忆科学设计复习时机,按例题遗忘速度动态安排重练
  • 在不同规模模型上恢复5至37个百分点的旧知识准确率损失
  • 适合需要长期学习新数据又不丢旧知识的场景

持续预训练大模型需在获取新信息时不丢失旧知识。现有重播方法常采用全局固定的新旧样本混合并均匀采样,忽略了样本遗忘速度的差异。本文将持续预训练建模为自适应复习调度:训练循环不仅决定回放历史数据量,还判断哪些样本应在何时重练。提出基于认知科学的间隔重复训练(SRT)框架,采用SuperMemo-2(SM-2)算法调度样本重练。SRT维护每个样本的复习状态,将困惑度映射为回忆质量信号,同时对历史样本进行保留调度、对新样本进行巩固调度,模型、目标函数和优化器保持不变。在时间分离的Wikipedia和代码语料上,SRT显著改善稳定-可塑性权衡,相较朴素持续预训练恢复5至37个百分点的旧知识准确率损失,同时保持或提升新知识学习能力。大规模实验表明,SRT能维持广泛基准性能,而朴素持续预训练与均匀重播则严重退化。视觉和表格数据实验进一步显示,该调度原则在配合合适回忆信号时可推广至非语言任务。

原文摘要 · Abstract (English)

Continual pre-training of large language models must acquire new information without erasing old knowledge. Existing replay methods often choose a global old/new mixture and sample uniformly, ignoring that examples differ in how quickly they are forgotten. We formulate continual pre-training as adaptive review scheduling: the training loop should decide not only how much history to replay, but which examples should return at each step. We introduce Spaced Repetition Training (SRT), a continual learning framework inspired by cognitive science, which schedules sample-rehearsal using the SuperMemo-2 (SM-2) algorithm. SRT maintains per-example review state, maps per-example perplexity to a recall-quality signal, and schedules historical examples for retention and new examples for consolidation while leaving the model, objective, and optimizer unchanged. On temporally separated Wikipedia and code corpora, SRT improves the stability-plasticity trade-off, recovering 5 to 37 percentage points of old-knowledge accuracy lost by naive continual pre-training across model scales while preserving or improving new-knowledge acquisition. At larger scale, SRT preserves broad benchmark performance that naive continual pre-training and uniform replay substantially degrade. Experiments with vision and tabular data further suggest that the scheduling principle extends beyond language when paired with an appropriate recall signal.

持续学习记忆机制语言模型复习调度

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。