Research map / Learning theory and algorithms / Prediction and decision losses
PAC Prediction Sets Under Covariate Shift
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 PS-W with calibrated uncertainty intervals for importance weights instead of PS-R's heuristic point weights (ablation) | PS-W · ImageNet-PGD | error: 0.074 -> 0.049 | Helped |
| Use PS-W with calibrated uncertainty intervals for importance weights instead of PS-R's heuristic point weights (ablation) | PS-W · ImageNet-C13 | error: 0.000 -> 0.061 | Helped |
| Replace calibrated IW uncertainty intervals with histogram density-based point estimates (ablation) | PS-W · ImageNet-PGD | size: 13.9 -> 7.8 | Helped |
| Replace calibrated IW uncertainty intervals with histogram density-based point estimates (ablation) | PS-W · ImageNet-C13 | error: 0.061 -> 0.095 | Hurt |
| Replace calibrated IW uncertainty intervals with histogram density-based point estimates (ablation) | PS-W · ImageNet-C13 | size: 35.8 -> 15.9 | Helped |
| Use calibrated importance-weight uncertainty intervals rather than uncalibrated point estimates (ablation) | PS-W · Support shifts on ImageNet | error | Helped |
| Vary the smoothness parameter E over a broad range (ablation) | PS-W · Covariate shifts | error | No clear effect |
| Use E=0 rather than a positive smoothness allowance (ablation) | PS-W · Covariate shifts | error | No clear effect |
| Use robust importance-weight intervals rather than heuristic point weights (ablation) | PS-W · Rate shifts on DomainNet | mean normalized size: 0.0257 -> 0.0047 | Mixed |
Closest papers by meaning
- PAC Prediction Sets Under Label Shift (ICLR 2024 · similarity 0.90)
- Tackling covariate shift with node-based Bayesian neural networks (ICML 2022 · similarity 0.88)
- Quantifying Uncertainty in the Presence of Distribution Shifts (NeurIPS 2025 · similarity 0.87)
- Optimal Aggregation of Prediction Intervals under Unsupervised Domain Shift (NeurIPS 2024 · similarity 0.87)
- From Matrix Inversion to Constraints: Provably Tighter Confidence Regions for Importance Weights in Label Shift (NeurIPS 2026 · similarity 0.84)
- Not all distributional shifts are equal: Fine-grained robust conformal inference (ICML 2024 · similarity 0.84)
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.