AI Paper Atlas

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

Task-Agnostic Undesirable Feature Deactivation Using Out-of-Distribution Data

Dongmin Park, Hwanjun Song, Minseok Kim, Jae-Gil Lee

NeurIPS 2021 · Label-efficient and robust learning › OOD and anomaly detection · approach: Uncertainty and misclassification detection · 2 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
replace softmax-level calibration (OAT) with feature-level calibrationOAT · CIFAR-10 N=500 (LSUN)accuracy: 40.73 -> 42.51Improved
add LSUN OOD data instead of SVHNTAUFE · CIFAR-10 N=500accuracy: 41.58 -> 42.51Improved
add Places365 instead of ImageNet-990 as OOD (small data)TAUFE · ImageNet-10 64x64 N=500accuracy: 42.80 -> 43.25Mixed
add TAUFE regularization to MixMatchMixMatch · CIFAR-10 semi-supervisedaccuracy: 88.32 -> 90.10Improved
add TAUFE regularization to MixMatchMixMatch · CIFAR-100 semi-supervisedaccuracy: 51.38 -> 52.58Improved
add TAUFE with LSUN OOD to PGD adversarial trainingPGD · CIFAR-10adversarial accuracy: 44.26 -> 45.48Improved
add TAUFE with LSUN OOD to PGD adversarial trainingPGD · CIFAR-10clean accuracy: 71.22 -> 72.31Improved
add TAUFE feature-level calibration for box regressionStandard ResNet-50 · CAR N=2,000 L1-IoUIoU: 83.52 -> 85.82Improved
use ImageNet instead of Places365 as OODTAUFE · CUB200 box regressionIoU: 77.24 -> 77.01Improved
replace softmax-level with feature-level calibration for WSOLOAT · CUB200 N=2,000 (ImageNet OOD)GT-known Loc: 55.24 -> 59.68Improved
increase image resolution of ImageNet-10 from 64 to 224TAUFE · ImageNet-10accuracyNo clear effect
add TAUFE to OOD detectionTAUFE · OOD detectiondetection performanceNo clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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