AcTTA: Rethinking Test-Time Adaptation via Dynamic Activation
This work addresses the problem of domain-shift robustness in machine learning models for researchers and practitioners, offering a novel perspective beyond affine modulation, though it is incremental in building on existing TTA approaches.
The paper tackled performance degradation under distribution shifts in test-time adaptation by proposing AcTTA, a framework that adaptively updates activation functions instead of focusing on normalization layers, achieving consistent improvements over existing methods on benchmarks like CIFAR10-C, CIFAR100-C, and ImageNet-C.
Test-time adaptation (TTA) aims to mitigate performance degradation under distribution shifts by updating model parameters during inference. Existing approaches have primarily framed adaptation around affine modulation, focusing on recalibrating normalization layers. This perspective, while effective, overlooks another influential component in representation dynamics: the activation function. We revisit this overlooked space and propose AcTTA, an activation-aware framework that reinterprets conventional activation functions from a learnable perspective and updates them adaptively at test time. AcTTA reformulates conventional activation functions (e.g., ReLU, GELU) into parameterized forms that shift their response threshold and modulate gradient sensitivity, enabling the network to adjust activation behavior under domain shifts. This functional reparameterization enables continuous adjustment of activation behavior without modifying network weights or requiring source data. Despite its simplicity, AcTTA achieves robust and stable adaptation across diverse corruptions. Across CIFAR10-C, CIFAR100-C, and ImageNet-C, AcTTA consistently surpasses normalization-based TTA methods. Our findings highlight activation adaptation as a compact and effective route toward domain-shift-robust test-time learning, broadening the prevailing affine-centric view of adaptation.