Research map / Label-efficient and robust learning / OOD and anomaly detection
Generalized ODIN: Detecting Out-of-Distribution Image Without Learning From Out-of-Distribution Data
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 DeConf-C* instead of Mahalanobis* | DenseNet OoD detector · CIFAR-100 vs ImageNet(r) | AUROC: 96.4 -> 98.6 | Helped |
| Use DeConf-C* instead of Mahalanobis* | DenseNet OoD detector · CIFAR-100 vs LSUN(r) | TNR@TPR95: 82.6 -> 93.8 | Helped |
| Use DeConf-C* instead of Mahalanobis* | DenseNet OoD detector · CIFAR-100 vs SVHN | TNR@TPR95: 43.3 -> 77.0 | Helped |
| Use DeConf-C* instead of Mahalanobis* | DenseNet OoD detector · CIFAR-100 vs Gaussian | AUROC: 100. -> 99.9 | Hurt |
| Use DeConf-C* instead of Mahalanobis* | DenseNet OoD detector · CIFAR-10 vs ImageNet(r) | AUROC: 98.2 -> 99.1 | Helped |
| Use DeConf-C* instead of Mahalanobis* | DenseNet OoD detector · CIFAR-10 vs LSUN(r) | TNR@TPR95: 91.7 -> 97.6 | Helped |
| Use DeConf-C* instead of Mahalanobis* | DenseNet OoD detector · CIFAR-10 vs SVHN | AUROC: 98.0 -> 98.8 | Helped |
| Use DeConf-C* instead of Mahalanobis* | DenseNet OoD detector · CIFAR-10 vs Gaussian | AUROC: 100. -> 99.9 | Hurt |
| Add the decomposed confidence structure (ablation) | CIFAR-10 and SVHN OoD detectors · Eight OoD benchmark datasets | AUROC | Helped |
| Add the decomposed confidence structure (ablation) | CIFAR-100 DeConf-I and g(x) detectors · Eight OoD benchmark datasets | AUROC | Hurt |
| Add input preprocessing with epsilon selected using in-distribution validation data (ablation) | Baseline, ODIN, Mahalanobis, and DeConf scoring functions · CIFAR-100 and eight OoD datasets | OoD detection performance | Helped |
| Add dropout regularization with rate 0.7 (ablation) | DeConf-I and g(x) branches · CIFAR-100 OoD detection | AUROC | Helped |
Closest papers by meaning
- Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networks (ICLR 2018 · similarity 0.92)
- Splitting & Integrating: Out-of-Distribution Detection via Adversarial Gradient Attribution (ICML 2025 · similarity 0.88)
- Image Background Serves as Good Proxy for Out-of-distribution Data (ICLR 2024 · similarity 0.88)
- A Systematic Analysis of Out-of-Distribution Detection Under Representation and Training Paradigm Shifts (NeurIPS 2026 · similarity 0.88)
- Mysteries of the Deep: Role of Intermediate Representations in Out of Distribution Detection (NeurIPS 2025 · similarity 0.87)
- Out-of-Distribution Detection with Prototypical Outlier Proxy (AAAI 2025 · 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.