面向野外世界稳定联邦持续测试时自适应
Towards Stable Federated Continual Test-Time Adaptation in Wild World
为视觉模型在无标签、客户端顺序到达的野外部署场景提供了一套稳定的在线自适应与聚合机制：用最大后验估计把贝叶斯先验注入本地 TTA 以防漂移，再用不确定性门控决定是否把单客户端更新并入全局模型，可直接迁移到跨域分类/分割的持续部署流程中。
原文依据 we consider the Federated Continual Test-Time Adaptation (FedCTTA) setting, where unlabeled clients arrive sequentially
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论联邦学习（FL）能够在保护隐私的前提下实现协同模型训练，但在部署阶段面临客户端数据异构与域偏移的挑战。尽管个性化联邦学习（PFL）缓解了异构性问题，它通常需要来自目标客户端的有标签数据，而这一假设并不现实。测试时自适应（TTA）提供了无标签自适应，但将其直接用于持续联邦场景可能破坏全局模型的稳定性并导致灾难性遗忘。为此，我们考虑联邦持续测试时自适应（FedCTTA）设定：无标签客户端依次到达，需要在线自适应并持续更新全局模型。我们提出 BPFedCTTA 框架，采用贝叶斯先验引导自适应（BPA）通过最大后验估计实现稳定的本地自适应，并采用不确定性门控的单客户端聚合（UGSA）依据客户端不确定性选择性地整合更新。该方法在自适应与知识保持之间取得平衡，从而缓解遗忘。在跨域分类与分割上的大量实验表明，BPFedCTTA 在顺序自适应与全局模型提升方面优于现有的 FL、PFL 与 TTA 方法。源代码将在论文录用后公开。
英文原文摘要 1324 字符
Federated Learning (FL) enables collaborative model training while preserving privacy, but faces challenges with client data heterogeneity and domain shifts during deployment. Although Personalized Federated Learning (PFL) mitigates heterogeneity, it typically requires labelled data from target clients, which is an impractical assumption. Test-Time Adaptation (TTA) offers label-free adaptation, yet its direct use in a continual federated setting risks destabilizing the global model and causing catastrophic forgetting. To address this, we consider the Federated Continual Test-Time Adaptation (FedCTTA) setting, where unlabeled clients arrive sequentially, requiring online adaptation and continuous global model updates. We propose BPFedCTTA, a framework that employs Bayesian Prior-guided Adaptation (BPA) for stable local adaptation via Maximum a Posteriori estimation, and Uncertainty-Gated Single-client Aggregation (UGSA) to selectively integrate updates based on client uncertainty. This approach balances adaptation with knowledge retention, thereby mitigating forgetting. Extensive experiments on cross-domain classification and segmentation show BPFedCTTA outperforms existing FL, PFL, and TTA methods in sequential adaptation and global model improvement. The source code will be made public upon acceptance.