Ranked Entropy Minimization for Continual Test-Time Adaptation
This work addresses stability issues in continual test-time adaptation for machine learning models, representing an incremental improvement over existing entropy minimization methods.
The paper tackles the problem of model collapse in continual test-time adaptation, where entropy minimization methods can converge to trivial solutions, and proposes ranked entropy minimization to improve stability, achieving effective results across various benchmarks.
Test-time adaptation aims to adapt to realistic environments in an online manner by learning during test time. Entropy minimization has emerged as a principal strategy for test-time adaptation due to its efficiency and adaptability. Nevertheless, it remains underexplored in continual test-time adaptation, where stability is more important. We observe that the entropy minimization method often suffers from model collapse, where the model converges to predicting a single class for all images due to a trivial solution. We propose ranked entropy minimization to mitigate the stability problem of the entropy minimization method and extend its applicability to continuous scenarios. Our approach explicitly structures the prediction difficulty through a progressive masking strategy. Specifically, it gradually aligns the model's probability distributions across different levels of prediction difficulty while preserving the rank order of entropy. The proposed method is extensively evaluated across various benchmarks, demonstrating its effectiveness through empirical results. Our code is available at https://github.com/pilsHan/rem