Evaluation of Test-Time Adaptation Under Computational Time Constraints
This addresses the need for more realistic evaluation in TTA for researchers and practitioners, highlighting the trade-off between accuracy and efficiency, though it is incremental as it focuses on benchmarking rather than introducing a new method.
The paper tackles the problem of evaluating test-time adaptation (TTA) methods by proposing a new online protocol that penalizes slower methods with fewer adaptation samples, revealing that simpler, faster methods like SHOT can outperform state-of-the-art but slower methods like SAR when accounting for computational constraints.
This paper proposes a novel online evaluation protocol for Test Time Adaptation (TTA) methods, which penalizes slower methods by providing them with fewer samples for adaptation. TTA methods leverage unlabeled data at test time to adapt to distribution shifts. Although many effective methods have been proposed, their impressive performance usually comes at the cost of significantly increased computation budgets. Current evaluation protocols overlook the effect of this extra computation cost, affecting their real-world applicability. To address this issue, we propose a more realistic evaluation protocol for TTA methods, where data is received in an online fashion from a constant-speed data stream, thereby accounting for the method's adaptation speed. We apply our proposed protocol to benchmark several TTA methods on multiple datasets and scenarios. Extensive experiments show that, when accounting for inference speed, simple and fast approaches can outperform more sophisticated but slower methods. For example, SHOT from 2020, outperforms the state-of-the-art method SAR from 2023 in this setting. Our results reveal the importance of developing practical TTA methods that are both accurate and efficient.