Research map / Label-efficient and robust learning / OOD and anomaly detection
Near OOD Detection for Vision-Language Prompt Learning with Contrastive Logit Score
Myong Chol Jung, Joanna F. Dipnall, Belinda Gabbe, He Zhao
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add CLS-M scoring (subtract scaled null-class similarity) to the MaxLogit score | CoOp · Near OOD, 13 datasets average (16-shot) | AUROC: 80.74 -> 81.84 | Improved |
| add CLS-M scoring (subtract scaled null-class similarity) to the MaxLogit score | CoCoOp · Near OOD, 13 datasets average (16-shot) | AUROC: 81.09 -> 82.74 | Improved |
| add CLS-M scoring (subtract scaled null-class similarity) to the MaxLogit score | IVLP · Near OOD, 13 datasets average (16-shot) | AUROC: 81.12 -> 84.34 | Improved |
| add CLS-M scoring (subtract scaled null-class similarity) to the MaxLogit score | KgCoOp · Near OOD, 13 datasets average (16-shot) | AUROC: 80.84 -> 83.12 | Improved |
| add CLS-M scoring (subtract scaled null-class similarity) to the MaxLogit score | ProGrad · Near OOD, 13 datasets average (16-shot) | AUROC: 79.77 -> 82.35 | Improved |
| add CLS-M scoring (subtract scaled null-class similarity) to the MaxLogit score | MaPLe · Near OOD, 13 datasets average (16-shot) | AUROC: 81.06 -> 83.94 | Improved |
| add CLS-M scoring (subtract scaled null-class similarity) to the MaxLogit score | PromptSRC · Near OOD, 13 datasets average (16-shot) | AUROC: 83.85 -> 85.77 | Improved |
| add CLS-M scoring (subtract scaled null-class similarity) to the MaxLogit score | LoCoOp · Near OOD, 13 datasets average (16-shot) | AUROC: 77.55 -> 81.74 | Improved |
| add CLS-E scoring to the Energy score | PromptSRC 16-shot · Near OOD, 13 datasets average | AUROC: 84.60 -> 87.10 | Improved |
| use CLS-M instead of LoCoOp GL-MCM | PromptSRC 16-shot · Near OOD, 13 datasets average | AUROC: 80.87 -> 87.07 | Improved |
| use CLS-M instead of MaxLogit | LoCoOp · Far OOD, ImageNet ID average | AUROC: 96.92 -> 97.72 | Improved |
| use CLS-E instead of Energy | LoCoOp · Far OOD, ImageNet ID average | AUROC: 94.97 -> 96.91 | Improved |
Closest papers by meaning
- Local-Prompt: Extensible Local Prompts for Few-Shot Out-of-Distribution Detection (ICLR 2025 · similarity 0.88)
- Adaptive Multi-prompt Contrastive Network for Few-shot Out-of-distribution Detection (ICML 2025 · similarity 0.88)
- Overcoming the Pitfalls of Vision-Language Model Finetuning for OOD Generalization (ICLR 2024 · similarity 0.87)
- Reflexive Guidance: Improving OoDD in Vision-Language Models via Self-Guided Image-Adaptive Concept Generation (ICLR 2025 · similarity 0.87)
- DeCoOp: Robust Prompt Tuning with Out-of-Distribution Detection (ICML 2024 · similarity 0.87)
- Self-Calibrated Tuning of Vision-Language Models for Out-of-Distribution Detection (NeurIPS 2024 · similarity 0.87)
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.
