Research map / Label-efficient and robust learning / OOD and anomaly detection
MOOD: Multi-Level Out-of-Distribution Detection
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 dynamic complexity-based exits instead of the final classifier only | MSDNet OOD detection · CIFAR-10 ID; 10 OOD datasets (average) | AUROC: 0.9048 -> 0.9126 | Helped |
| Use dynamic complexity-based exits instead of the final classifier only | MSDNet OOD detection · CIFAR-100 ID; 10 OOD datasets (average) | FLOPs (× 10^8): 1.05 -> 0.79 | Helped |
| Use adjusted energy rather than ordinary energy scoring (ablation) | MOOD · CIFAR-10 ID; 10 OOD datasets (average) | AUROC: 0.9070 -> 0.9129 | Helped |
| Use adjusted energy rather than ordinary energy scoring (ablation) | MOOD · CIFAR-100 ID; 10 OOD datasets (average) | AUROC: 0.8433 -> 0.8507 | Helped |
| Replace adjusted energy with MSP scoring (ablation) | MOOD · CIFAR-10 ID; 10 OOD datasets (average) | AUROC: 0.9129 -> 0.8984 | Hurt |
| Replace adjusted energy with ODIN scoring (ablation) | MOOD · CIFAR-100 ID; 10 OOD datasets (average) | AUROC: 0.8507 -> 0.8548 | Helped |
| Replace complexity-based exits with greedy score-threshold exits (ablation) | MOOD · CIFAR-100 ID; 10 OOD datasets (average) | FPR95: 0.5709 -> 0.6635 | Hurt |
| Replace complexity-based exits with random exit selection (ablation) | MOOD · CIFAR-100 ID; 10 OOD datasets (average) | AUROC: 0.8507 -> 0.8298 | Hurt |
| Use complexity-based dynamic exits rather than the best fixed exit (ablation) | MOOD · CIFAR-100 ID; 10 OOD datasets (average) | AUROC: 0.8531 -> 0.8507 | No clear effect |
| Use JPEG2000 rather than PNG for complexity estimation (ablation) | MOOD · CIFAR-10 ID; 10 OOD datasets (average) | AUROC: 0.9129 -> 0.9123 | No clear effect |
| Use JPEG2000 rather than PNG for complexity estimation (ablation) | MOOD · CIFAR-100 ID; 10 OOD datasets (average) | AUROC: 0.8507 -> 0.8558 | Helped |
Closest papers by meaning
- Expecting The Unexpected: Towards Broad Out-Of-Distribution Detection (NeurIPS 2024 · similarity 0.91)
- The Best of Both Worlds: On the Dilemma of Out-of-distribution Detection (NeurIPS 2024 · similarity 0.91)
- Splitting & Integrating: Out-of-Distribution Detection via Adversarial Gradient Attribution (ICML 2025 · similarity 0.90)
- ImageNet-OOD: Deciphering Modern Out-of-Distribution Detection Algorithms (ICLR 2024 · similarity 0.90)
- MultiOOD: Scaling Out-of-Distribution Detection for Multiple Modalities (NeurIPS 2024 · similarity 0.90)
- Mysteries of the Deep: Role of Intermediate Representations in Out of Distribution Detection (NeurIPS 2025 · similarity 0.90)
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.