Long-tailed classifiers can be rebalanced after training by dividing class probabilities by the model's own average predicted prior, yielding small accuracy gains over using class frequencies.
Global and local mixture consistency cumulative learning for long-tailed visual recognitions
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2024 1verdicts
REJECT 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Prior2Posterior: Model Prior Correction for Long-Tailed Learning
Long-tailed classifiers can be rebalanced after training by dividing class probabilities by the model's own average predicted prior, yielding small accuracy gains over using class frequencies.