基于遗忘曲线设计记忆重放，让大模型持续学习更持久。
FOREVER: Forgetting Curve-Inspired Memory Replay for Language Model Continual Learning
- 用优化器更新量定义模型时间，动态调整重放时机。
- 在多个基准上显著减少灾难性遗忘，3个任务均提升10%以上。
- 适合需要长期增量学习的大模型研究与应用。
大语言模型的持续学习旨在实现顺序知识获取而不产生灾难性遗忘。记忆重放方法因实用性和有效性被广泛采用，但多数依赖固定的、基于训练步数的启发式策略，常与模型实际学习进度错配，因为相同训练步数可能引发不同程度的参数变化。受最新发现——大语言模型遗忘模式类似人类艾宾浩斯遗忘曲线——的启发，我们提出 FOREVER（FORgEtting curVe-inspired mEmory Replay）框架，将重放调度与以模型为中心的时间概念对齐。FOREVER 通过优化器更新幅度定义模型时间，使基于遗忘曲线的重放间隔与模型内部演化进程同步而非固定训练步数。在此基础上，该框架引入基于遗忘曲线的重放调度器以确定重放时机，并设计强度感知正则化机制以自适应控制重放强度。在三个持续学习基准及从 0.6B 到 13B 参数的多种模型上进行的大量实验表明，FOREVER 能稳定缓解灾难性遗忘。
原文摘要 · Abstract (English)
Continual learning (CL) for large language models (LLMs) aims to enable sequential knowledge acquisition without catastrophic forgetting. Memory replay methods are widely used for their practicality and effectiveness, but most rely on fixed, step-based heuristics that often misalign with the model's actual learning progress, since identical training steps can result in varying degrees of parameter change. Motivated by recent findings that LLM forgetting mirrors the Ebbinghaus human forgetting curve, we propose FOREVER (FORgEtting curVe-inspired mEmory Replay), a novel CL framework that aligns replay schedules with a model-centric notion of time. FOREVER defines model time using the magnitude of optimizer updates, allowing forgetting curve-inspired replay intervals to align with the model's internal evolution rather than raw training steps. Building on this approach, FOREVER incorporates a forgetting curve-based replay scheduler to determine when to replay and an intensity-aware regularization mechanism to adaptively control how to replay. Extensive experiments on three CL benchmarks and models ranging from 0.6B to 13B parameters demonstrate that FOREVER consistently mitigates catastrophic forgetting.
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。