SLCA++：释放带预训练的持续学习中序列微调的潜力
SLCA++: Unleash the Power of Sequential Fine-tuning for Continual Learning with Pre-training
AI总结：
针对带预训练的持续学习中预训练知识渐进过拟合、提示方法容量有限的问题，提出SLCA++框架，通过慢学习器、分类器对齐等改进释放序列微调潜力，在图像分类持续学习场景上大幅超越现有最优方法。
摘要：
近年来，带预训练的持续学习（CLPT）受到了广泛关注，一改传统上从零开始训练的研究重心。使用强大的预训练模型（PTMs）能够极大地促进知识迁移并缓解灾难性遗忘，但也存在预训练知识逐渐过拟合到特定下游任务的问题。目前大多数研究通常会冻结预训练模型，加入任务特定提示来引导表征学习，同时在推理阶段搭配提示选择流程。然而，由于提示参数的容量有限，该策略在持续学习中仅能取得次优性能。相比之下，调整预训练模型的全部参数往往能为表征学习提供最大潜力，这使得序列微调（Seq FT）成为CLPT领域中被忽视的基础基线方法。为此，我们从序列微调的视角对渐进式过拟合问题展开了深入分析。考虑到表征学习速度过快以及分类层存在偏差是导致这一特定问题的原因，我们提出了先进的带分类器对齐的慢学习器（Slow Learner with Classifier Alignment，SLCA++）框架，以释放序列微调的潜力，作为CLPT的强基线方法。我们的方法包含一个慢学习器（Slow Learner）用于选择性降低骨干网络参数的学习率，以及一个分类器对齐模块以事后方式对齐互不相交的分类层。我们进一步通过对称交叉熵损失提升了慢学习器的效果，还采用了一种参数高效策略来实现搭载SLCA++的序列微调。在图像分类基准的多种持续学习场景下，我们的方法带来了大幅性能提升，且以较大优势超越了当前最优方法。代码链接：https://github.com/GengDavid/SLCA。
英文摘要：
In recent years, continual learning with pre-training (CLPT) has received widespread interest, instead of its traditional focus of training from scratch. The use of strong pre-trained models (PTMs) can greatly facilitate knowledge transfer and alleviate catastrophic forgetting, but also suffers from progressive overfitting of pre-trained knowledge into specific downstream tasks. A majority of current efforts often keep the PTMs frozen and incorporate task-specific prompts to instruct representation learning, coupled with a prompt selection process for inference. However, due to the limited capacity of prompt parameters, this strategy demonstrates only sub-optimal performance in continual learning. In comparison, tuning all parameters of PTMs often provides the greatest potential for representation learning, making sequential fine-tuning (Seq FT) a fundamental baseline that has been overlooked in CLPT. To this end, we present an in-depth analysis of the progressive overfitting problem from the lens of Seq FT. Considering that the overly fast representation learning and the biased classification layer constitute this particular problem, we introduce the advanced Slow Learner with Classifier Alignment (SLCA++) framework to unleash the power of Seq FT, serving as a strong baseline approach for CLPT. Our approach involves a Slow Learner to selectively reduce the learning rate of backbone parameters, and a Classifier Alignment to align the disjoint classification layers in a post-hoc fashion. We further enhance the efficacy of SL with a symmetric cross-entropy loss, as well as employ a parameter-efficient strategy to implement Seq FT with SLCA++. Across a variety of continual learning scenarios on image classification benchmarks, our approach provides substantial improvements and outperforms state-of-the-art methods by a large margin. Code: https://github.com/GengDavid/SLCA.
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。
