Evaluation of Test-Time Adaptation Under Computational Time Constraints

Abstract

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.

Resources

arXiv: 2304.04795

Citation

@inproceedings{alfarra2024tta,
  title     = {Evaluation of Test-Time Adaptation Under Computational Time Constraints},
  author    = {Alfarra, Motasem and Itani, Hani and Pardo, Alejandro and Alhuwaider, Shyma and Ramazanova, Merey and P{\'{e}}rez, Juan C. and Cai, Zhipeng and M{\"{u}}ller, Matthias and Ghanem, Bernard},
  booktitle = {International Conference on Machine Learning (ICML)},
  year      = {2024}
}