Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Bayesian Calibration of Imperfect Computer Models using Physics-Informed Priors
Michail Spitieris, Ingelin Steinsland
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 BCPI instead of KOH | Bayesian calibration of WK2 · WK3-simulated data | prediction RMSE: 1.91 -> 1.61 | Improved |
| use BCPI instead of KOH | Bayesian calibration of WK2 · WK3-simulated data | runtime (sec): 483 -> 23 | Improved |
| add model discrepancy GP and Bayesian calibration to PI opt | physics-informed GP prior · WK3-simulated data | prediction RMSE: 6.85 -> 1.61 | Improved |
| add a GP model discrepancy term delta(t) to the WK2 physics-informed prior (ablation) | WK2 model · WK3-simulated data | posterior of R and C | Improved |
| add a GP model discrepancy term delta(t) to the WK2 prior (ablation) | WK2 model · real blood pressure and flow data | noise parameters sigma_P, sigma_Q | Improved |
| add a GP sensor bias term to the heat equation prior (ablation) | heat equation physics-informed prior · Heat equation with biased sensors | diffusion parameter alpha | Improved |
| add a sensor bias term to the heat equation prior (ablation) | heat equation physics-informed prior · Heat equation with biased sensors | prediction uncertainty | Hurt |
| use a periodic kernel instead of SE or RQ (ablation) | WK2 physics-informed prior · WK2 simulated data | posterior uncertainty of R | Improved |
| fix noise parameters to point estimates from independent GPs at prediction time (ablation) | physics-informed VFE approximation · WK2 simulated data | prediction accuracy | Improved |
| increase discrepancy R1 from 0.03 to 0.08 without modelling discrepancy (ablation) | WK2 model · WK3-simulated data | bias of R | Hurt |
Closest papers by meaning
- Bayesian Adaptive Calibration and Optimal Design (NeurIPS 2024 · similarity 0.83)
- Online Bayesian Calibration under Gradual and Abrupt System Changes (NeurIPS 2026 · similarity 0.81)
- Smooth ECE: Principled Reliability Diagrams via Kernel Smoothing (ICLR 2024 · similarity 0.81)
- Addressing Misspecification in Simulation-based Inference through Data-driven Calibration (ICML 2025 · similarity 0.81)
- Provable Uncertainty Decomposition via Higher-Order Calibration (ICLR 2025 · similarity 0.80)
- Testing Calibration in Nearly-Linear Time (NeurIPS 2024 · similarity 0.80)
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.
