Replaces the single latent true label assumption with a per-object latent label distribution for unbiased and consistent aggregation of noisy labels, performing better on ambiguous tasks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.HC 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Latent Distribution Assumption for Unbiased and Consistent Consensus Modelling
Replaces the single latent true label assumption with a per-object latent label distribution for unbiased and consistent aggregation of noisy labels, performing better on ambiguous tasks.