Research map / Label-efficient and robust learning / OOD and anomaly detection
SCOPED: Score–Curvature Out-of-distribution Proximity Evaluator for Diffusion
Brett Barkley, Preston Culbertson, David Fridovich-Keil
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 |
|---|---|---|---|
| Use SCOPED (2F+2J) instead of DiffPath on average vision AUROC | Diffusion OOD detection · Vision OOD average | AUROC: 0.918 -> 0.892 | Hurt |
| Use SCOPED instead of DiffPath on C10 vs C100 | Diffusion OOD detection · C10 vs C100 | AUROC: 0.590 -> 0.477 | Hurt |
| Add global sign correction to the typicality ratio (ablation) | SCOPED two-step · Vision OOD average | AUROC: 0.81 -> 0.84 | Improved |
| Evaluate at mid-level timestep t=300 instead of early t=1 only (ablation) | SCOPED single step · Vision OOD average | AUROC: 0.7985 -> 0.8835 | Improved |
| Combine early and mid timesteps instead of single t=300 (ablation) | SCOPED · Vision OOD average | AUROC: 0.884 -> 0.892 | Improved |
| Choose the best timestep per dataset pair with oracle knowledge (ablation) | SCOPED single step · Vision OOD average | AUROC: 0.884 -> 0.944 | Improved |
| Train detector on most diverse buffer (medium-replay) in D4RL (ablation) | SCOPED · D4RL Hopper/Walker2d/Ant | AUROC | Hurt |
Closest papers by meaning
- Detecting Out-of-Distribution Through the Lens of Neural Collapse (CVPR 2025 · similarity 0.87)
- GOOD: Training-Free Guided Diffusion Sampling for Out-of-Distribution Detection (NeurIPS 2025 · similarity 0.87)
- Revisiting Logit Distributions for Reliable Out-of-Distribution Detection (NeurIPS 2025 · similarity 0.86)
- Fast Decision Boundary based Out-of-Distribution Detector (ICML 2024 · similarity 0.86)
- Understanding the Generalization of Pretrained Diffusion Models on Out-of-Distribution Data (AAAI 2024 · similarity 0.86)
- MultiOOD: Scaling Out-of-Distribution Detection for Multiple Modalities (NeurIPS 2024 · similarity 0.86)
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.
