Research map / Label-efficient and robust learning / OOD and anomaly detection
Out-of-distribution Detection Learning with Unreliable Out-of-distribution Sources
Haotian Zheng, Qizhou Wang, Zhen Fang, Xiaobo Xia, Feng Liu, Tongliang Liu, Bo Han
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 ATOL instead of ManifoldGAN | data generation-based OOD detection · CIFAR-10 | FPR95: 26.68 -> 14.66 | Improved |
| use ATOL instead of ManifoldGAN | data generation-based OOD detection · ImageNet | FPR95: 72.50 -> 51.80 | Improved |
| add threshold tau to separate auxiliary ID and OOD latent distributions (ablation) | ATOL · CIFAR-100 | FPR95: 92.90 -> 55.06 | Improved |
| add distance-preserving regularization l_reg on the generator (ablation) | ATOL · CIFAR-10 | FPR95: 41.13 -> 22.29 | Improved |
| add ID distribution alignment loss l_align (ablation) | ATOL · CIFAR-100 | FPR95: 74.20 -> 55.06 | Improved |
| replace DCGAN with BigGAN generator (ablation) | ATOL · CIFAR-100 | FPR95: 55.52 -> 36.72 | Improved |
| replace trained DCGAN with random-parameterized DCGAN (ablation) | ATOL · CIFAR-100 | FPR95: 55.52 -> 65.69 | Hurt |
| use class-conditional instead of class-agnostic auxiliary ID data (ablation) | ATOL · CIFAR-100 | FPR95: 61.54 -> 55.52 | Improved |
| use Euclidean instead of cosine-based centralized distance in l_reg (ablation) | ATOL · CIFAR-10 | FPR95: 19.73 -> 14.62 | Improved |
| increase alignment weight alpha from 0.01 to 1 (ablation) | ATOL · CIFAR-10 | FPR95: 22.99 -> 14.11 | Improved |
| increase alpha from 1 to 10 (ablation) | ATOL · CIFAR-100 | FPR95: 55.38 -> 61.82 | Mixed |
| increase latent space size u from 0.5 to 8 (ablation) | ATOL · CIFAR-10 | FPR95: 18.15 -> 13.90 | Improved |
Closest papers by meaning
- Can We Ignore Labels in Out of Distribution Detection? (ICLR 2025 · similarity 0.90)
- Mining In-distribution Attributes in Outliers for Out-of-distribution Detection (AAAI 2025 · similarity 0.90)
- Out-of-Distribution Detection with Prototypical Outlier Proxy (AAAI 2025 · similarity 0.89)
- Out-Of-Distribution Detection with Diversification (Provably) (NeurIPS 2024 · similarity 0.89)
- Learning to Shape In-distribution Feature Space for Out-of-distribution Detection (NeurIPS 2024 · similarity 0.89)
- What If the Input is Expanded in OOD Detection? (NeurIPS 2024 · similarity 0.89)
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.
