Test-time adaptation is stabilized by filtering unreliable samples, seeking flat entropy minima, and applying redundancy and inequity regularizers to pseudo-labeled class centroids.
On feature normalization and data augmentation,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Adapt in the Wild: Test-Time Entropy Minimization with Sharpness and Feature Regularization
Test-time adaptation is stabilized by filtering unreliable samples, seeking flat entropy minima, and applying redundancy and inequity regularizers to pseudo-labeled class centroids.