AI RESEARCH
Uncertainty Decomposition via Cyclical SG-MCMC and Soft-label Learning for Subjective NLP
arXiv CS.AI
•
ArXi:2605.24773v1 Announce Type: new Annotator disagreement in emotion classification reflects ambiguity intrinsic to emotion concepts and is essential for predictor-quality assessment in subjective NLP. Yet no prior work integrates soft-label learning with Bayesian deep learning to evaluate uncertainty along axes including annotator-distribution fidelity. We train a linear head on a frozen RoBERTa via cyclical stochastic gradient Marko chain Monte Carlo (cSG-MCMC), targeting the empirical annotator distribution with a soft-label objective under a five-axis evaluation.