AI Paper Atlas

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

Rethinking Reconstruction Autoencoder-Based Out-of-Distribution Detection

Yibo Zhou

CVPR 2022 · Label-efficient and robust learning › OOD and anomaly detection · approach: Spurious-aware OOD detection · 77 citations (OpenAlex)

Check 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
Replace image reconstruction targets with classifier activation-vector features (ablation)LSA OoD detector · CIFAR-100 vs. TinyImagenet-cropFPR@95%TPR: 42.0 -> 99.5Hurt
Replace L2 distance with normalized L2 distance (ablation)LSA with AV feature reconstruction · CIFAR-100 vs. TinyImagenet-cropFPR@95%TPR: 99.5 -> 5.9Improved
Replace autoregressive latent-space regularization with cross-entropy regularization (ablation)LSA with normalized L2 · CIFAR-100 vs. TinyImagenet-cropFPR@95%TPR: 5.9 -> 3.0Improved
Use layerwise feature reconstruction instead of basic input reconstruction scoring (ablation)LSA with cross-entropy regularizer · CIFAR-100 vs. TinyImagenet-cropAUROC: 99.1 -> 91.3Hurt
Add epsilon terms to the layerwise certainty score (ablation)Layerwise semantic reconstruction · CIFAR-100 vs. TinyImagenet-cropAUROC: 91.3 -> 99.4Improved
Add epsilon terms to the layerwise certainty score (ablation)Layerwise semantic reconstruction · CIFAR-100 vs. TinyImagenet-cropFPR@95%TPR: 0.2 -> 1.5Hurt
Remove weight decay when training the classifier used for activation features (ablation)Layerwise semantic reconstruction with Dense-BC · CIFAR-100 vs. TinyImagenet-cropAUROC: 96.1 / 97.1Improved
Remove weight decay from Dense-BC classifier training (ablation)Dense-BC classifier · CIFAR-100 classificationaccuracy: 77% to 72%Hurt
Remove weight decay from Wide-ResNet classifier training (ablation)Wide-ResNet classifier · CIFAR-100 classificationaccuracy: 78% to 77%Hurt

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