Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Geometric Autoencoder Priors for Bayesian Inversion: Learn First Observe Later
Arnaud Vadeboncoeur, Gregory Duthé, Mark Girolami, Eleni Chatzi
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 |
|---|---|---|---|
| learned geometric autoencoder prior instead of graph GP prior | GABI-ABC · Heat on rectangles (known noise) | MAE: 8.06e-2 (GP M 3/2) -> 1.58e-2 | Improved |
| GABI-ABC vs supervised direct map | GABI-ABC · Heat on rectangles (known noise) | MAE: 1.25e-2 -> 1.58e-2 | Hurt |
| GABI-ABC vs supervised direct map | GABI-ABC · Airfoil flow (pressure) | MAE: 5.35e-2 -> 6.92e-2 | Hurt |
| GABI-ABC vs supervised direct map | GABI-ABC · Car resonance, amplitude u | MAE: 6.03e-1 -> 2.32e-1 | Improved |
| GABI-ABC vs supervised direct map | GABI-ABC · Car resonance, source forcing f | MAE: 2.46e-1 -> 2.26e-3 | Improved |
| use NUTS MCMC instead of ABC sampling (ablation) | GABI · Heat on rectangles (known noise) | MAE: 1.58e-2 -> 1.11e-2; coverage 1 std 80.91% -> 66.66%; pred time 0.908s -> 410.31s | Mixed |
| swap GCN for Transformer architecture (ablation) | GABI-ABC · Heat on rectangles, architecture comparison | MAE: 1.58e-2 -> 1.19e-2 | Improved |
| estimate unknown observation noise jointly | GABI-ABC · Heat on rectangles (unknown noise) | MAE of u: 1.58e-2 -> 2.09e-2 | Mixed |
Closest papers by meaning
- GAIA: Geometry-Adaptive Operator Learning for Forward and Inverse Problems (NeurIPS 2026 · similarity 0.87)
- Uncertainty Quantification for Forward and Inverse Problems of PDEs via Latent Global Evolution (AAAI 2024 · similarity 0.85)
- 3DID: Direct 3D Inverse Design for Aerodynamics with Physics-Aware Optimization (NeurIPS 2025 · similarity 0.84)
- Geometry Aware Operator Transformer as an efficient and accurate neural surrogate for PDEs on arbitrary domains (NeurIPS 2025 · similarity 0.84)
- Variational Multi-scale Representation for Estimating Uncertainty in 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.84)
- Calibrated Physics-Informed Uncertainty Quantification (ICML 2025 · 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.
