Continual Test-Time Adaptation (CTTA) aims to enable models to adapt on-the-fly to a stream of unlabeled data under evolving distribution shifts. However, existing CTTA methods typically rely on shared model parameters across all domains, making them vulnerable to feature entanglement and catastrophic forgetting in the presence of large or non-stationary domain shifts. To address this limitation, we propose \textbf{ExPaMoE}, a novel framework based on an \emph{Expandable Parallel Mixture-of-Experts} architecture. ExPaMoE decouples domain-general and domain-specific knowledge via a dual-branch expert design with token-guided feature separation, and dynamically expands its expert pool based on a \emph{Spectral-Aware Online Domain Discriminator} (SODD) that detects distribution changes in real-time using frequency-domain cues. Extensive experiments demonstrate the superiority of ExPaMoE across diverse CTTA scenarios. We evaluate our method on standard benchmarks including CIFAR-10C, CIFAR-100C, ImageNet-C, and Cityscapes-to-ACDC for semantic segmentation. Additionally, we introduce \textbf{ImageNet++}, a large-scale and realistic CTTA benchmark built from multiple ImageNet-derived datasets, to better reflect long-term adaptation under complex domain evolution. ExPaMoE consistently outperforms prior arts, showing strong robustness, scalability, and resistance to forgetting.
View on arXiv