technique

soft-label learning

techniqueactivesoft-label-learning-7c44eb33·1 events·first seen 29d ago

Aliases: soft-label learning

Co-occurring entities

MNIST human uncertainty alignment model calibration dataset cartography

More like this (12)

Soft Label Supervision Soft Q-Learning Imitation Learning meta-learning neurosymbolic learning unsupervised learning in-context learning Learning with Opponent-Learning Awareness (LOLA)few-shot learning point cloud learning Q-learning Operator Learning

Recent events (1)

5arXiv · cs.CL·29d ago·source ↗

Controlled Audit of Human vs. Synthetic Soft-Labels for Calibration and Uncertainty Alignment

This paper presents a controlled study disentangling the effects of human soft-labels from label mode-shift corrections in soft-label learning, using MNIST and a synthetic variant. The authors find that human soft-labels primarily act as a regularizer improving calibration on difficult samples and promoting stable training convergence, rather than simply correcting mislabeled data. Dataset cartography analysis shows models trained on human soft-labels mirror human uncertainty patterns, while those trained on synthetic labels fail to align. The work provides a diagnostic testbed for evaluating human-AI uncertainty alignment.

Evaluation and Benchmarking AI Safety Research MNIST human uncertainty alignment model calibration +3 more