Research map / Probabilistic and scientific modeling / Variational and generative sampling
Locally Most Powerful Bayesian Test for Out-of-Distribution Detection using Deep Generative Models
Keunseo Kim, JunCheol Shin, 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 |
|---|---|---|---|
| use LMPBT instead of input complexity (IC) | VAE · CIFAR-10 -> SVHN | AUROC: 0.912 -> 0.931 | Improved |
| use LMPBT instead of likelihood regret (LR) | VAE · CIFAR-10 -> CelebA | AUROC: 0.714 -> 0.783 | Improved |
| use LMPBT instead of likelihood regret (LR) | VAE · CIFAR-100 -> CelebA | AUROC: 0.598 -> 0.683 | Improved |
| use LMPBT instead of input complexity (IC) | VAE · CIFAR-100 -> SVHN | AUROC: 0.912 -> 0.956 | Improved |
| use LMPBT instead of likelihood ratio (LLR) | VAE · CIFAR-10 -> MNIST | AUROC: 0.008 -> 0.998 | Improved |
| use LMPBT instead of likelihood regret (LR) | VAE · CIFAR-10 -> MNIST | AUROC: 0.998 -> 0.998 | No clear effect |
| use top 30 Hessian eigenvalues instead of fewer than 10 (ablation) | VAE · CIFAR-10 -> SVHN | AUROC: 0.913 | Improved |
| increase number of Hessian eigenvalues beyond the optimum (ablation) | VAE · CIFAR-10 -> SVHN | AUROC | Mixed |
| use LMPBT instead of likelihood regret (LR) for cost | VAE · CIFAR-10 | seconds per sample: 0.30s -> 0.11s | Improved |
| use LMPBT instead of LLR or IC for cost | VAE · CIFAR-10 | seconds per sample: 0.0125 s -> 0.11s | Hurt |
| scale Hessian eigendecomposition with parameters (ablation) | VAE · FMNIST/CIFAR sizes | time per parameter: 0.0028 s/dim -> 0.0025 s/dim | No clear effect |
| Hessian eigendecomposition cost relative to training | VAE · FMNIST | time cost: 2543 s vs 13448 s | No clear effect |
Closest papers by meaning
- A Geometric Explanation of the Likelihood OOD Detection Paradox (ICML 2024 · similarity 0.88)
- The Best of Both Worlds: On the Dilemma of Out-of-distribution Detection (NeurIPS 2024 · similarity 0.88)
- Enhancing the Power of OOD Detection via Sample-Aware Model Selection (CVPR 2024 · similarity 0.86)
- Out-of-Distribution Detection with Prototypical Outlier Proxy (AAAI 2025 · similarity 0.86)
- Outlier Synthesis via Hamiltonian Monte Carlo for Out-of-Distribution Detection (ICLR 2025 · similarity 0.86)
- OODD: Test-time Out-of-Distribution Detection with Dynamic Dictionary (CVPR 2025 · 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.
