通过保留主子空间，实现视觉语言模型持续学习中的知识稳定与更新平衡。
KeepLoRA: Continual Learning with Residual Gradient Adaptation
- 在残差子空间中限制LoRA参数更新，避免干扰已有知识。
- 新任务学习时投影梯度至与历史特征正交的子空间，提升知识保留能力。
- 适用于需要长期增量学习的预训练视觉语言模型场景。
预训练视觉语言模型的持续学习需兼顾三重目标：保留预训练知识、维持已学任务知识，同时保持学习新知识的灵活性。本文提出一种简单有效的方法KeepLoRA，以平衡这三者。我们分析了模型参数空间中的知识保留机制，发现通用知识主要编码于主子空间，而任务特定知识则编码于残差子空间。受此启发，KeepLoRA通过将新任务的梯度投影到与预训练模型主子空间及前序任务主导方向均正交的子空间，仅在残差子空间更新LoRA参数，从而避免对已有能力的干扰。理论与实证分析表明，KeepLoRA能有效平衡三重目标，并取得当前最优性能。代码已开源：https://github.com/MaolinLuo/KeepLoRA。
原文摘要 · Abstract (English)
Continual learning for pre-trained vision-language models requires balancing three competing objectives: retaining pre-trained knowledge, preserving knowledge from a sequence of learned tasks, and maintaining the plasticity to acquire new knowledge. This paper presents a simple but effective approach called KeepLoRA to effectively balance these objectives. We first analyze the knowledge retention mechanism within the model parameter space and find that general knowledge is mainly encoded in the principal subspace, while task-specific knowledge is encoded in the residual subspace. Motivated by this finding, KeepLoRA learns new tasks by restricting LoRA parameter updates in the residual subspace to prevent interfering with previously learned capabilities. Specifically, we infuse knowledge for a new task by projecting its gradient onto a subspace orthogonal to both the principal subspace of pre-trained model and the dominant directions of previous task features. Our theoretical and empirical analyses confirm that KeepLoRA balances the three objectives and achieves state-of-the-art performance. The implementation code is available at https://github.com/MaolinLuo/KeepLoRA.
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。