用冻结模型+适配器，少样本持续学习不遗忘。
Few-Shot Continual Learning for 3D Brain MRI with Frozen Foundation Models
- 冻结预训练主干，仅训练任务专属低秩适配器
- 双任务连续学习下保持零遗忘，参数增量低于0.1%
- 适合医疗影像中需长期维护多个任务的场景
在大规模3D医学影像数据上预训练的基座模型，在少样本持续学习环境下适应多个下游任务时面临挑战。本文针对3D脑部MRI的少样本持续学习问题，采用冻结的预训练主干与任务专用的低秩适配器（LoRA）相结合的方法。任务按顺序输入——肿瘤分割（BraTS）和脑龄估计（IXI），且不重放旧任务数据。每个任务配备独立的LoRA适配器；仅训练适配器和任务特定头，主干保持冻结，从而设计上杜绝灾难性遗忘（后向转移误差BWT=0）。连续学习中，顺序全微调导致严重遗忘（T1 Dice从0.80降至0.16），而顺序线性探针虽在T1表现良好（Dice 0.79），但在T2失败（MAE 1.45）。本文方法在两任务间取得最佳平衡：T1 Dice为0.62±0.07，T2 MAE为0.16±0.05，实现零遗忘且每任务可训练参数占比低于0.1%，尽管在T2存在显著系统性年龄低估（Wilcoxon p<0.001）。结果表明，冻结基座模型结合任务专属LoRA适配器，是少样本持续学习中同时维持多任务性能的有效方案。
原文摘要 · Abstract (English)
Foundation models pretrained on large-scale 3D medical imaging data face challenges when adapted to multiple downstream tasks under continual learning with limited labeled data. We address few-shot continual learning for 3D brain MRI by combining a frozen pretrained backbone with task-specific Low-Rank Adaptation (LoRA) modules. Tasks arrive sequentially -- tumor segmentation (BraTS) and brain age estimation (IXI) -- with no replay of previous task data. Each task receives a dedicated LoRA adapter; only the adapter and task-specific head are trained while the backbone remains frozen, thereby eliminating catastrophic forgetting by design (BWT=0). In continual learning, sequential full fine-tuning suffers severe forgetting (T1 Dice drops from 0.80 to 0.16 after T2), while sequential linear probing achieves strong T1 (Dice 0.79) but fails on T2 (MAE 1.45). Our LoRA approach achieves the best balanced performance across both tasks: T1 Dice 0.62$\pm$0.07, T2 MAE 0.16$\pm$0.05, with zero forgetting and $<$0.1\% trainable parameters per task, though with noted systematic age underestimation in T2 (Wilcoxon $p<0.001$). Frozen foundation models with task-specific LoRA adapters thus offer a practical solution when both tasks must be maintained under few-shot continual learning.
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。