Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Bayesian model selection and misspecification testing in imaging inverse problems only from noisy and partial measurements
Tom Sprunck, Marcelo Pereyra, Tobías I. Liaudat
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 Phi1 data-fission estimator instead of Bayesian residual | blur kernel selection · single shot accuracy | accuracy: 40.0% -> 86.7% | Improved |
| use Phi1 instead of EB residual | blur kernel selection · single shot accuracy | accuracy: 46.7% -> 86.7% | Improved |
| pool three measurements (few shot) instead of single shot | Phi1 · blur kernel selection | accuracy: 86.7% -> 100% | Improved |
| increase blur strength sigma from 0.5 to 5 | Phi2 OOD test · moderate OOD (Met) deblurring | power: 87% -> 60% | Hurt |
| increase blur strength sigma from 0.5 to 5 | Phi2 OOD test · strong OOD deblurring | power: 100% -> 100% | No clear effect |
| use Phi2 instead of Phi1 for OOD detection | MRI R=4 · brain vs knee power | power: 94% -> 100% | Improved |
| use Phi2 instead of Phi1 for OOD detection | MRI R=8 · brain vs knee power | power: 96% -> 100% | Improved |
| decrease splitting parameter alpha | Phi estimator · Gaussian toy model | model discrimination | Improved |
| average K noise realizations | estimator · Gaussian toy model | bias | Improved |
| decrease alpha | estimator · Gaussian toy model | computational cost | Hurt |
| use DRUNet prior instead of wavelet prior | Phi1 · galaxy denoising | Phi1: 83.38 -> 22.23 | Improved |
| use DRUNet prior instead of DnCNN | posterior mean PSNR · galaxy denoising | PSNR (dB): 29.86 -> 32.86 | Improved |
Closest papers by meaning
- Principled Probabilistic Imaging using Diffusion Models as Plug-and-Play Priors (NeurIPS 2024 · similarity 0.87)
- Divide-and-Conquer Posterior Sampling for Denoising Diffusion priors (NeurIPS 2024 · similarity 0.86)
- BUFF: Bayesian Uncertainty Guided Diffusion Probabilistic Model for Single Image Super-Resolution (AAAI 2025 · similarity 0.86)
- Plug-and-Play Posterior Sampling under Mismatched Measurement and Prior Models (ICLR 2024 · similarity 0.86)
- Provably Robust Score-Based Diffusion Posterior Sampling for Plug-and-Play Image Reconstruction (NeurIPS 2024 · similarity 0.86)
- Provable Posterior Sampling with Denoising Oracles via Tilted Transport (NeurIPS 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.
