On ChaosNLI, POPQUORN, and a CIFAR-10H subset, training on full annotation distributions instead of majority labels preserves accuracy while better matching human uncertainty.
Learning from multi-annotator data: A noise-aware classification framework.ACM Trans
1 Pith paper cite this work, alongside 15 external citations. Polarity classification is still indexing.
1
Pith paper citing it
15
external citations · OpenAlex
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Distributions In, Distributions Out: The Case for Soft-Label Training
On ChaosNLI, POPQUORN, and a CIFAR-10H subset, training on full annotation distributions instead of majority labels preserves accuracy while better matching human uncertainty.