AI Paper Atlas

Research map / Label-efficient and robust learning / OOD and anomaly detection

A Call to Reflect on Evaluation Practices for Failure Detection in Image Classification

Paul F Jaeger, Carsten Tim Lüth, Lukas Klein, Till J. Bungert

ICLR 2023 · Label-efficient and robust learning › OOD and anomaly detection · approach: Uncertainty and misclassification detection · 16 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
Evaluate confidence scoring methods across varied datasets and realistic distribution shifts using AURCConfidence scoring functions for image classifiers · FD-Shifts benchmarkAURCImproved
Use MCD-MSR rather than ConfidNetCNN classifier · CIFAR-10 and CIFAR-100 iidAURCImproved
Use OoD-D methods rather than softmax baselinesConfidence scoring functions · Non-semantic new-class shiftsAURCImproved
Apply OoD-D methods across nuanced covariate shifts instead of their intended new-class scenariosOoD-D confidence scoring functions · Covariate shiftsAURCHurt
Use expected entropy rather than mutual informationMCD uncertainty measures · IID test setsAURCImproved
Expect mutual information to outperform expected entropy on distribution shiftsMCD uncertainty measures · Distribution shiftsAURCNo clear effect
Use 32-bit rather than 64-bit precision for MSR logitsViT classifier · CIFAR-10AURC * 10^3: 0.950 -> 3.480Hurt
Cast MSR logits to 64-bit rather than 32-bit precisionViT classifier · BREEDSAURC * 10^3: 4.559 -> 1.893Improved
Use ViT rather than CNNClassifier with confidence scoring · Most datasetsAURCImproved
Use ViT rather than CNNClassifier with confidence scoring · iWildCamAURCHurt

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