AI Paper Atlas

Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes

What is Flagged in Uncertainty Quantification? Latent Density Models for Uncertainty Categorization

Hao Sun, Boris van Breugel, Jonathan Crabbé, Nabeel Seedat, Mihaela van der Schaar

NeurIPS 2023 · Probabilistic and scientific modeling › Uncertainty and Gaussian processes · approach: Epistemic versus aleatoric decomposition · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use Deep Ensemble as the UQ method instead of MC-DropoutDAUC categorization · Dirty-MNIST, top 5% uncertainshare of flagged examples that are OODImproved
use MC-Dropout instead of Deep EnsembleUQ method · Dirty-MNISTOOD flaggingHurt
use Deep Ensemble rather than BNN or MCDUQ method · Dirty-CIFAR-10Bnd and IDM flaggingImproved
use BNN instead of Deep EnsembleUQ method · Dirty-CIFAR-10Bnd and IDM flaggingHurt
vary Bnd and IDM thresholds away from 0.8 (ablation)DAUC · Dirty-MNISTcategorization conclusionsNo clear effect
change kernel from Gaussian to Tophat or Exponential (ablation)DAUC · Dirty-MNISTscore correlation: all above 0.86No clear effect
replace DAUC OOD scoring with Outlier-AE (ablation)OOD detector · Dirty-MNISTF1-score: 1.0000 -> 1.0000No clear effect
replace DAUC OOD scoring with IForest (ablation)OOD detector · Dirty-MNISTF1-score: 1.0000 -> 0.9999No clear effect
retrain on training data filtered to resemble B&I test examples (increase discarded proportion q)linear models and ResNet-18 · Covtype, Digits, Spam, DMNISTaccuracy on B&I classImproved
discard too much training data (large q) (ablation)filtered retraining · Covtype, Digits, Spam, DMNISTaccuracy on B&I classHurt
filtered retraining on Bnd classlinear models · Covtype, Digits, Spamaccuracy on Bnd classImproved
filtered retraining on IDM classlinear models · Covtype, Digits, Spamaccuracy on IDM classImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy