Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Single Model Uncertainty Estimation via Stochastic Data Centering
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 |
|---|---|---|---|
| scale mean logits by anchor-based uncertainty (full Delta-UQ) instead of using the anchor-ensemble mean alone (ablation) | Delta-UQ ResNet-50 · ImageNet-C severity 5 | ECE mean: 0.110 -> 0.044 | Helped |
| use anchoring at inference to compute the mean over anchors (vs naive {0,x} inference) (ablation) | Delta-UQ ResNet-50 · ImageNet-C severity 5 | ECE mean: 0.130 -> 0.110 | Helped |
| train with anchoring even if not used at inference (naive {0,x}) vs vanilla (ablation) | ResNet-50 · ImageNet-C severity 5 | ECE mean: 0.194 -> 0.130 | Helped |
| use anchor-based uncertainty to calibrate vs deep ensembles | ResNet-50 · ImageNet-C severity 5 | ECE mean: 0.088 -> 0.044 | Helped |
| replace GP/MCD/BNN/DEns surrogate with Delta-UQ in Bayesian optimization | sequential optimization · black-box function suite | AUC: avg rank 1.0 vs 2.95 (DEns) | Helped |
| use Delta-UQ single model instead of deep ensemble for accuracy | ImageNet-C ResNet-50 · ImageNet-C average accuracy | accuracy: 50.05 -> 48.95 | Hurt |
| use Delta-UQ single model instead of deep ensemble | CIFAR10-C ResNet-20 · CIFAR10-C average accuracy | accuracy: 72.9 -> 74.25 | Helped |
| add anchor training on ImageNet | ResNet-50 · ImageNet val | top-1 accuracy: 76.1 -> 76.1 | No clear effect |
Closest papers by meaning
- Epistemic Uncertainty Quantification For Pre-Trained Neural Networks (CVPR 2024 · similarity 0.90)
- Uncertainty Quantification with the Empirical Neural Tangent Kernel (NeurIPS 2025 · similarity 0.90)
- General Uncertainty Estimation with Delta Variances (AAAI 2025 · similarity 0.89)
- One Step Closer to Unbiased Aleatoric Uncertainty Estimation (AAAI 2024 · similarity 0.88)
- Accurate and Scalable Estimation of Epistemic Uncertainty for Graph Neural Networks (ICLR 2024 · similarity 0.88)
- Transitional Uncertainty with Layered Intermediate Predictions (ICML 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.