Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Posterior Network: Uncertainty Estimation without OOD Samples via Density-Based Pseudo-Counts
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 |
|---|---|---|---|
| add normalizing flow density in latent space (ablation) | PostNet · Sensorless Drive OOD epistemic | AUC-PR: 31.15 -> 88.73 | Helped |
| add Bayesian loss instead of cross-entropy (ablation) | PostNet · Sensorless Drive OOD epistemic | AUC-PR: 82.63 -> 88.73 | Helped |
| add batch norm before flow (sequential training) (ablation) | PostNet · Sensorless Drive accuracy | accuracy: 15.09 -> 98.42 | Helped |
| train encoder and flow jointly instead of sequentially (ablation) | PostNet · Sensorless Drive OOD epistemic | AUC-PR: 71.75 -> 88.73 | Helped |
| replace mixture of Gaussians with flows (ablation) | PostNet · Sensorless Drive OOD epistemic | AUC-PR: 65.35 -> 88.65 | Helped |
| increase latent dimension to 32 (ablation) | PostNet · Segment/Sensorless Drive | accuracy and OOD | Hurt |
| radial flow vs IAF (ablation) | PostNet · Sensorless Drive OOD epistemic | AUC-PR: 88.73 vs 88.65 | No clear effect |
| use PostNet instead of RKL-PN trained with uniform noise OOD | PostNet · CIFAR10 VGG16 OOD SVHN epistemic | AUC-PR: 49.37 -> 77.71 | Helped |
| use PostNet instead of RKL-PN w/ FashionMNIST OOD | PostNet · MNIST OOD FashionMNIST aleatoric | AUC-PR: 100.0 -> 97.78 | Hurt |
| use PostNet instead of RKL-PN w/ F. | PostNet · MNIST OODom FMNIST aleatoric | AUC-PR: 88.96 -> 100.0 | Helped |
| use PostNet instead of Dropout | PostNet · CIFAR10 VGG16 accuracy | accuracy: 82.84 -> 84.85 | Helped |
| stronger encoder (conv to ResNet) (ablation) | PostNet · CIFAR10 OOD SVHN epistemic | AUC-PR: 57.72 -> 72.25 | Helped |
Closest papers by meaning
- Predictive Uncertainty Estimation via Prior Networks (NeurIPS 2018 · similarity 0.90)
- One Step Closer to Unbiased Aleatoric Uncertainty Estimation (AAAI 2024 · similarity 0.86)
- Uncertainty Quantification with the Empirical Neural Tangent Kernel (NeurIPS 2025 · similarity 0.86)
- Estimating Epistemic and Aleatoric Uncertainty with a Single Model (NeurIPS 2024 · similarity 0.86)
- Epistemic Uncertainty Quantification For Pre-Trained Neural Networks (CVPR 2024 · similarity 0.85)
- Quantifying Uncertainty in the Presence of Distribution Shifts (NeurIPS 2025 · similarity 0.85)
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.