双相持续学习：监督适应与无监督保留的结合
Dual-Phase Continual Learning: Supervised Adaptation Meets Unsupervised Retention
AI总结：
本文提出了一种基于无监督方法的双相持续学习框架，通过利用测试时间数据提升先前任务性能，无需存储或重播数据，有效缓解VLMs中的遗忘问题。
摘要：
基础视觉-语言模型（VLMs）在多种任务中表现出色，但将其适应到新领域而不遗忘先前知识仍是一个关键挑战。持续学习（CL）通过使模型能够从新数据中按顺序学习，同时减轻对先前信息的遗忘，通常在有监督的设置下进行，涉及标签转移。然而，突然的分布转移仍可能导致显著的遗忘，可能抵消有监督更新的好处，尤其是在存储或重播过去数据不可行时。在本文中，我们提出利用无标签的测试时间数据以无监督的方式加强先前任务的性能，而无需重播或存储示例。与传统的测试时间适应（TTA）不同，后者主要关注领域转移或损坏，我们的方法通过利用部署过程中遇到的代表性测试样本来提高早期任务的性能。我们引入了一个简单的教师-学生框架，通过基于梯度的稀疏参数更新，并展示该方法在VLMs中的持续学习中有效减轻了类别递增CL中的遗忘，提供了一种无记忆替代传统经验重播的方法，并具有强大的实证结果。
英文摘要：
Foundational Vision-Language Models (VLMs) excel across diverse tasks, but adapting them to new domains without forgetting prior knowledge remains a critical challenge. Continual Learning (CL) addresses this challenge by enabling models to learn sequentially from new data while mitigating the forgetting of prior information, typically under supervised settings involving label shift. Nonetheless, abrupt distribution shifts can still cause substantial forgetting, potentially nullifying the benefits of supervised updates, especially when storing or replaying past data is infeasible. In this work, we propose leveraging unlabeled testtime data in an unsupervised manner to reinforce prior task performance without requiring replay or stored examples. Unlike traditional Test Time Adaptation (TTA), which primarily focuses on domain shift or corruption, our method improves performance on earlier tasks by exploiting representative test samples encountered during deployment. We introduce a simple Teacher-Student framework with gradient-based sparse parameter updates, and show that it effectively mitigates forgetting in class-incremental CL for VLMs, offering a memory-free alternative to episodic replay with strong empirical results.
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。
