Finite ensemble sizes cause systematic slope attenuation in conditional reliability diagnostics for means, spreads, and probabilities; analytical expressions and practical estimators correct for this bias.
Murphy and Robert L
2 Pith papers cite this work, alongside 391 external citations. Polarity classification is still indexing.
2
Pith papers citing it
391
external citations · OpenAlex
years
2026 2verdicts
CONDITIONAL 2representative citing papers
There exists a non-trivial Lipschitz constant L* that makes a network intrinsically calibrated; LiST finds it automatically via temperature feedback and produces calibrated robust models on CIFAR and Tiny-ImageNet.
citing papers explorer
-
Ensemble size effects on conditional reliability estimates: slope attenuation bias and correction methods
Finite ensemble sizes cause systematic slope attenuation in conditional reliability diagnostics for means, spreads, and probabilities; analytical expressions and practical estimators correct for this bias.
-
LiST: Lipschitz Scaling Training for Robust and Calibrated Neural Networks
There exists a non-trivial Lipschitz constant L* that makes a network intrinsically calibrated; LiST finds it automatically via temperature feedback and produces calibrated robust models on CIFAR and Tiny-ImageNet.