DGS：用于类增量学习的双梯度与语义漂移引导的低秩适配
DGS: Dual Gradient and Semantic-Shift Guided Low-Rank Adaptation for Class Incremental Learning
对视觉研究者而言，这篇论文给出了在预训练视觉模型上做类增量学习时，用双梯度融合（基子空间投影 + 任务 LoRA 梯度插值融合）替代刚性正交约束的思路，并配合分类器对齐与 patch 级对齐损失，可直接迁移到 ViT 类骨干的持续学习/多任务微调流程中。
原文依据 fuses task-specific LoRA gradients with their aligned counterparts through interpolated combination
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论在类增量学习（CIL）中，应用于预训练模型（PTM）的参数高效微调在适应新任务时仍然容易遭受灾难性遗忘。缓解灾难性遗忘的主流策略是将梯度约束在过去任务的正交子空间内，但刚性的梯度约束会损害可塑性。本文提出了一种新的 CIL 框架——双梯度与语义漂移引导的低秩适配（DGS），它通过梯度融合来平衡稳定性与可塑性，并通过分类器与 patch token 对齐来保持表征一致性。具体而言，该方法引入了双梯度更新策略：首先从预训练模型中导出基子空间投影，然后通过插值组合将任务特定的低秩适配梯度与其对齐后的对应梯度进行融合。这一设计在不牺牲任务特定表达能力的前提下促进了知识保持。此外，我们采用带有语义漂移估计的分类器对齐机制，该机制基于校准后的原型统计量来缓解分类器漂移，并引入一种新的 patch 级对齐损失以保持跨任务的特征一致性。在六个标准基准上的大量实验表明，我们的方法持续优于现有的 CIL 方法，凸显了其在持续学习场景中的有效性与泛化能力。
英文原文摘要 1441 字符
In Class-Incremental Learning (CIL), parameter efficient fine-tuning applied to Pre-trained Models (PTMs) remain vulnerable to catastrophic forgetting as they adapt to new tasks. The prevalent strategy to mitigate catastrophic forgetting is to constrain gradients within the orthogonal subspaces of past tasks, while rigid gradient constraints hinder plasticity. In this paper, we propose a novel CIL framework, Dual Gradient and Semantic-Shift Guided Low-Rank Adaptation (DGS), that balances stability and plasticity via gradient fusion and maintains representation consistency through classifier and patch-token alignment. Specifically, our method introduces the Dual Gradient update strategy that first derives a base subspace projection from the PTMs and then fuses task-specific LoRA gradients with their aligned counterparts through interpolated combination. This design promotes knowledge retention without sacrificing task-specific expressiveness. Furthermore, we employ a Classifier Alignment mechanism with Semantic shift estimation which is based on the calibrated prototype statistics to mitigate classifier shift, and introduce a novel Patch-level Alignment loss to preserve feature consistency across tasks. Extensive experiments on six standard benchmarks demonstrate that our approach consistently outperforms existing CIL methods, highlighting its effectiveness and generalization capability in continual learning scenarios.