Research map / Label-efficient and robust learning / OOD and anomaly detection
Towards neural networks that provably know when they don't know
Read the paperCheck 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 |
|---|---|---|---|
| integrate in- and out-distribution GMM densities into confidence calibration | CCU · adversarial noise neighborhoods on MNIST, FMNIST, SVHN, CIFAR10, CIFAR100 | AUC: 100.0 on all five datasets; ACET: 100.0, 100.0, 96.5, 99.9, 95.8 | Helped |
| integrate in- and out-distribution GMM densities into confidence calibration | CCU · adversarial noise neighborhoods on MNIST, FMNIST, SVHN, CIFAR10, CIFAR100 | success rate: 0.0 on all five datasets | Helped |
| train ACET around uniform noise instead of tiny images (ablation) | ACET · SVHN adversarial noise | AUC: 96.5 -> 5.4 | Hurt |
| train ACET around uniform noise instead of tiny images (ablation) | ACET · CIFAR10 adversarial noise | AUC: 99.9 -> 35.9 | Hurt |
| train ACET around uniform noise instead of tiny images (ablation) | ACET · CIFAR100 adversarial noise | AUC: 95.8 -> 14.6 | Hurt |
| train ACET around uniform noise instead of tiny images (ablation) | ACET · MNIST adversarial noise | AUC: 100.0 -> 100.0 | No clear effect |
| train ACET around uniform noise instead of tiny images (ablation) | ACET · FMNIST adversarial noise | AUC: 100.0 -> 100.0 | No clear effect |
| use tiny images rather than uniform noise as ACET's training out-distribution (ablation) | ACET · standard OOD detection | AUC | Helped |
| use tiny-image out-distribution data and certified density-based confidence | CCU · standard OOD detection | AUC | Mixed |
Closest papers by meaning
- Leveraging Perturbation Robustness to Enhance Out-of-Distribution Detection (CVPR 2025 · similarity 0.86)
- NECO: NEural Collapse Based Out-of-distribution detection (ICLR 2024 · similarity 0.86)
- Expecting The Unexpected: Towards Broad Out-Of-Distribution Detection (NeurIPS 2024 · similarity 0.86)
- Test-Time Linear Out-of-Distribution Detection (CVPR 2024 · similarity 0.86)
- Can We Ignore Labels in Out of Distribution Detection? (ICLR 2025 · similarity 0.86)
- Exploiting Discrepancy in Feature Statistic for Out-of-Distribution Detection (AAAI 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.