CODA uses consensus-based priors and Bayesian updating to select the best candidate model with far fewer labels than prior active model selection methods, beating them on 18 of 26 benchmark tasks.
A general model for aggregating annota- tions across simple, complex, and multi-object annotation tasks
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
-
Consensus-Driven Active Model Selection
CODA uses consensus-based priors and Bayesian updating to select the best candidate model with far fewer labels than prior active model selection methods, beating them on 18 of 26 benchmark tasks.