search

DequanWang / tent

ICLR21 Tent: Fully Test-Time Adaptation by Entropy Minimization
https://arxiv.org/abs/2006.10726
MIT License
364 stars 47 forks source link

how does minimizing the shannon entropy increase the accuracy of classification? #15

Open XYHhhhhhhhh opened 2 years ago

XYHhhhhhhhh commented 2 years ago

all i know is minimizing the shannon entropy can lower the uncertainty of the model, but how can i assure the model is becoming more accurate instead of becoming less accurate after minimizing the entropy?looking forward to an answer

ArlenCHEN commented 1 year ago

I think minimizing entropy = increasing accuracy makes sense ONLY when the pre-adaptation probability is well-calibrated. But I cannot find any discussion on calibration in the paper...