Research map / Label-efficient and robust learning / OOD and anomaly detection
OOD-MAML: Meta-Learning for Few-Shot Out-of-Distribution Detection and Classification
Taewon Jeong, Heeyoung Kim
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 learned fake (adversarial OOD) samples to (N+1)-class MAML (ablation) | (N+1)-MAML · Omniglot | detection accuracy (alpha=95): 0.9142 -> 0.9524 | Improved |
| add learned fake samples to (N+1)-class MAML (ablation) | (N+1)-MAML · CIFAR-FS | TNR (alpha=95): 0.1051 -> 0.2017 | Improved |
| add learned fake samples to (N+1)-class MAML (ablation) | (N+1)-MAML · miniImageNet | TNR (alpha=95): 0.0749 -> 0.1009 | Improved |
| use OOD-MAML one-class sub-tasks instead of (N+1)-MAML* with fake samples (ablation) | OOD-MAML · miniImageNet | TNR (alpha=95): 0.1009 -> 0.6770 | Improved |
| use OOD-MAML instead of ODIN on MAML | few-shot OOD detection · CIFAR-FS | TNR (alpha=95): 0.2311 -> 0.5492 | Improved |
| use task-specific adapted fake samples (Eq. 2) in addition to task-agnostic fake samples (ablation) | OOD-MAML · CIFAR-FS feature visualization | decision boundary tightness / detection | Improved |
| increase number of fake samples M from 3 to 5 (ablation) | OOD-MAML · Omniglot | detection accuracy (alpha=95): 0.9712 -> 0.9701 | No clear effect |
| replace MAML with OOD-MAML for classification | 5-shot 5-way classification · miniImageNet | classification accuracy: 0.5926 -> 0.6044 | No clear effect |
| change test-time number of classes N from 5 to 7 without retraining | OOD-MAML · CIFAR-FS | classification accuracy: 0.7158 -> 0.6964 | Hurt |
Closest papers by meaning
- Learning with Mixture of Prototypes for Out-of-Distribution Detection (ICLR 2024 · similarity 0.89)
- Expecting The Unexpected: Towards Broad Out-Of-Distribution Detection (NeurIPS 2024 · similarity 0.88)
- MetaOOD: Automatic Selection of OOD Detection Models (ICLR 2025 · similarity 0.88)
- Can We Ignore Labels in Out of Distribution Detection? (ICLR 2025 · similarity 0.88)
- Beyond Clean Training Data: A Versatile and Model-Agnostic Framework for Out-of-Distribution Detection with Contaminated Training Data (CVPR 2025 · similarity 0.88)
- The Best of Both Worlds: On the Dilemma of Out-of-distribution Detection (NeurIPS 2024 · similarity 0.88)
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.
