AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

Category-Extensible Out-of-Distribution Detection via Hierarchical Context Descriptions

Kai Liu, Zhihang Fu, Chao Chen, Sheng Jin, Ze Chen, Mingyuan Tao, Rongxin Jiang, Jieping Ye

NeurIPS 2023 · Vision-language and multimodal › CLIP and zero-shot · approach: Feature calibration for unseen-class 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
Learn hierarchical perceptual and spurious contexts instead of using a perceptual-only baseline (ablation)CATEX on ImageNet-100 · ImageNet-100FPR95: 13.07 -> 11.39Improved
Learn hierarchical contexts using perturbation-guided synthetic OOD samples (ablation)CATEX on ImageNet-100 · ImageNet-100FPR95: 11.39 -> 10.59Improved
Regularize inference similarities with the spurious-context score (ablation)CATEX on ImageNet-100 · ImageNet-100FPR95: 10.59 -> 10.31Improved
Use a larger perturbing ratio, especially perturbing all context tokens (ablation)CATEX · ImageNet-100FPR95: 10.31 -> 11.70Hurt
Choose the most distant token for perturbation rather than a random token (ablation)CATEX · ImageNet-100FPR95: 10.31 -> 10.42Hurt
Choose the closest token for perturbation rather than a random token (ablation)CATEX · ImageNet-100FPR95: 10.31 -> 10.86Hurt
Add an orthogonal constraint when using multiple spurious contexts (ablation)CATEX with 4 spurious contexts · ImageNet-100FPR95: 10.27 -> 9.89Improved
Increase the number of unconstrained spurious contexts (ablation)CATEX · ImageNet-100FPR95: 10.31 -> 10.21No clear effect
Use the regularized score with simulated spurious contexts rather than visual descriptions aloneZero-shot classification with CLIP and VisDesc · ImageNet-1K zero-shot classificationACC-Top1: 65.47 -> 65.84Improved
Use CATEX rather than NPOS for standard OOD detectionImageNet-1K ID, average over OOD datasets · ImageNet-1K ID; OOD averageFPR95: 37.93 -> 29.66Improved
Merge independently learned contexts for the two category subsetsCATEX · ImageNet-200 unionACC: 86.23 -> 89.61Improved
Use CATEX instead of MCM under a severe domain shiftImageNet-1K-trained model · ImageNet-SketchACC: 47.42 -> 48.59Hurt

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