AI Paper Atlas

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

ReAct: Out-of-distribution Detection With Rectified Activations

Yiyou Sun, Chuan Guo, Yixuan Li

NeurIPS 2021 · Label-efficient and robust learning › OOD and anomaly detection · approach: Spurious-aware OOD detection · 46 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
add activation clipping at 90th percentile (ReAct) to energy scoreResNet-50 energy score · ImageNet-1k OOD avgFPR95: 58.41 -> 31.43Improved
add ReAct to energy scoreResNet-18 CIFAR-10 · CIFAR-10 OOD avgFPR95: 35.60 -> 32.91Improved
add ReAct to energy scoreResNet-18 CIFAR-100 · CIFAR-100 OOD avgFPR95: 71.93 -> 59.61Improved
add ReAct to softmax scoreResNet-50 MSP · ImageNet-1k OOD avgFPR95: 66.95 -> 58.28Improved
add ReAct to ODINResNet-50 ODIN · ImageNet-1k OOD avgFPR95: 56.48 -> 44.10Improved
lower clipping percentile from 90 to 65 (ablation)ReAct ResNet-50 · ImageNet-1k OOD avgFPR95: 31.43 -> 74.62Hurt
raise clipping percentile from 90 to 99 (ablation)ReAct ResNet-50 · ImageNet-1k OOD avgFPR95: 31.43 -> 44.57Hurt
clip at p=90 instead of no clipping (ablation)ResNet-50 · ImageNet-1k valID accuracy: 75.08 -> 73.75Hurt
apply ReAct at penultimate layer instead of layer 3 (ablation)ResNet-50 · ImageNet-1k OOD avgFPR95: 63.87 -> 31.43Improved
add ReAct to GroupNorm model (ablation)ResNet-50 GroupNorm · ImageNet-1k OOD avgFPR95: 66.28 -> 51.58Improved
add ReAct to WeightNorm model (ablation)ResNet-50 WeightNorm · ImageNet-1k OOD avgFPR95: 50.34 -> 26.80Improved

Later papers that compared with this method

28 later accepted papers report ReAct as a baseline in their results tables; 28 of them report a better number on most of the results they share with it.

20 more not listed.

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