Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Single-Model Uncertainties for Deep Learning
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 |
|---|---|---|---|
| Use SQR rather than ConditionalGaussian for interval width | SQR · concrete | MPIW: 0.32 ±0.09 -> 0.31 ±0.06 | Helped |
| Use SQR rather than ConditionalGaussian for interval width | SQR · power | MPIW: 0.18 ±0.00 -> 0.18 ±0.01 | No clear effect |
| Use SQR rather than ConditionalGaussian for interval width | SQR · wine | MPIW: 0.49 ±0.03 -> 0.45 ±0.04 | Helped |
| Use SQR rather than ConditionalGaussian for interval width | SQR · yacht | MPIW: 0.03 ±0.01 -> 0.06 ±0.04 | Hurt |
| Use SQR rather than ConditionalGaussian for interval width | SQR · naval | MPIW: 0.15 ±0.25 -> 0.12 ±0.09 | Helped |
| Use SQR rather than ConditionalGaussian for interval width | SQR · energy | MPIW: 0.12 ±0.18 -> 0.08 ±0.03 | Helped |
| Use SQR rather than ConditionalGaussian for interval width | SQR · boston | MPIW: 0.55 ±0.20 -> 0.36 ±0.09 | Helped |
| Use SQR rather than ConditionalGaussian for interval width | SQR · kin8nm | MPIW: 0.20 ±0.01 -> 0.23 ±0.02 | Hurt |
| Use OCs instead of entropy for OOD detection | OCs · CIFAR-10 | ROC AUC: 0.80 ± 0.01 -> 0.83 ± 0.01 | Helped |
| Use OCs instead of BALD for OOD detection | OCs · Fashion-MNIST | ROC AUC: 0.95 ± 0.02 -> 0.92 ± 0.00 | Hurt |
| Use OCs instead of BALD for OOD detection | OCs · MNIST | ROC AUC: 0.95 ± 0.02 -> 0.95 ± 0.00 | No clear effect |
| Use OCs instead of entropy for OOD detection | OCs · SVHN | ROC AUC: 0.93 ± 0.01 -> 0.91 ± 0.00 | Hurt |
Closest papers by meaning
- Single Model Uncertainty Estimation via Stochastic Data Centering (NeurIPS 2022 · similarity 0.89)
- Possibilistic Predictive Uncertainty for Deep Learning (ICML 2026 · similarity 0.89)
- Is Epistemic Uncertainty Faithfully Represented by Evidential Deep Learning Methods? (ICML 2024 · similarity 0.88)
- Epistemic Uncertainty Quantification For Pre-Trained Neural Networks (CVPR 2024 · similarity 0.88)
- One Step Closer to Unbiased Aleatoric Uncertainty Estimation (AAAI 2024 · similarity 0.87)
- Are Uncertainty Quantification Capabilities of Evidential Deep Learning a Mirage? (NeurIPS 2024 · 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.