面向基础模型的参数重要性驱动持续学习
Parameter Importance-Driven Continual Learning for Foundation Models
AI总结：
针对基础模型领域后训练易发生灾难性遗忘的问题，提出PIECE持续学习方法，仅选择性更新0.1%核心参数，无需历史数据、不增参，在多类模型上实现了最优持续学习性能。
摘要：
特定领域后训练往往会引发灾难性遗忘，导致基础模型丧失通用推理能力，限制其对动态真实环境的适应性。在获取下游领域知识的同时保留通用能力，是大语言模型与多模态模型面临的核心挑战。传统持续学习方法如正则化、回放、架构隔离等存在下游性能差、依赖无法获取的历史数据或额外参数开销的问题。近期的参数高效微调（PET）方法虽能缓解遗忘，但其效果高度依赖参数选择与更新策略。本文提出PIECE，即基于参数重要性估计的持续增强方法，该方法无需访问先前训练数据、不增加模型参数，即可在高效学习领域知识的同时保留通用能力。PIECE在两个重要性估计器的引导下，仅选择性更新与新任务最相关的0.1%核心参数：一是基于Fisher信息的PIECE-F，二是基于融合梯度与曲率信息的二阶归一化的PIECE-S。在3个语言模型和2个多模态模型上开展的实验表明，PIECE可保留通用能力，在各类下游任务上取得了当前最优的持续学习性能。研究结果为构建无灾难性遗忘、可扩展的领域自适应基础模型提供了可行路径。
英文摘要：
Domain-specific post-training often causes catastrophic forgetting, making foundation models lose their general reasoning ability and limiting their adaptability to dynamic real-world environments. Preserving general capabilities while acquiring downstream domain knowledge is a central challenge for large language and multimodal models. Traditional continual learning methods, such as regularization, replay and architectural isolation, suffer from poor downstream performance, reliance on inaccessible historical data, or additional parameter overhead. While recent parameter-efficient tuning (PET) methods can alleviate forgetting, their effectiveness strongly depends on the choice of parameters and update strategies. In this paper, we introduce PIECE, a Parameter Importance Estimation-based Continual Enhancement method that preserves general ability while efficiently learning domain knowledge without accessing prior training data or increasing model parameters. PIECE selectively updates only 0.1% of core parameters most relevant to new tasks, guided by two importance estimators: PIECE-F based on Fisher Information, and PIECE-S based on a second-order normalization that combines gradient and curvature information. Experiments across three language models and two multimodal models show that PIECE maintains general capabilities and achieves state-of-the-art continual learning performance across diverse downstream tasks. Our results highlight a practical path to scalable, domain-adaptive foundation models without catastrophic forgetting.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
