Research map / 3D vision / 3D generation
Generative Diffusion Priors for 3D Mapping of the Dark Universe
Brandon Zhao, Diana Scognamiglio, Olivier Doré, Katherine L. Bouman
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 diffusion posterior sampling with learned Conicus3D prior instead of Neural Ensemble | 3D WL mass mapping · Mock lightcone vols 1-3 | 3D blurred Pearson correlation: 0.79 -> 0.83 (vol 1); 0.80 -> 0.83 (vol 2); 0.86 -> 0.92 (vol 3) | Improved |
| use diffusion posterior sampling instead of Wiener filter | 3D WL mass mapping · Mock lightcone vols 1-3 | 3D blurred Pearson correlation: 0.71 -> 0.83 (vol 1); 0.72 -> 0.83 (vol 2); 0.84 -> 0.92 (vol 3) | Improved |
| use diffusion posterior sampling instead of Neural Ensemble for 2D convergence maps | 2D mass mapping · Mock lightcone vols 1-3 | 2D Pearson correlation: 0.86 -> 0.87; 0.87 -> 0.88; 0.96 -> 0.98 | Improved |
| use diffusion posterior sampling instead of Neural Ensemble on unblurred full-resolution 3D fields | 3D WL mass mapping · Mock lightcone vols 1-3 | 3D Pearson correlation: 0.21 -> 0.23; 0.21 -> 0.27; 0.09 -> 0.18 | Mixed |
| use diffusion posterior sampling instead of Wiener filter on unblurred 3D fields | 3D WL mass mapping · Mock lightcone vols 1-3 | 3D Pearson correlation: 0.21 -> 0.23; 0.23 -> 0.27; 0.13 -> 0.18; vol 1 tied with Neural at 0.21 | Mixed |
| apply method under out-of-distribution cosmology (massless neutrinos) | diffusion posterior sampler · Vol 3 cosmology mismatch | Pearson correlation: 2D 0.98, blurred 3D 0.92 vs baselines 0.96/0.86 and 0.92/0.84 | No clear effect |
| use likelihood-guided posterior sampling with Fourier-diagonal covariance (power spectrum) in DAPS | diffusion posterior sampler · Angular power spectrum of kappa | small-scale power fidelity | Improved |
| use diffusion sampler instead of Neural Ensemble for uncertainty | uncertainty calibration · Sample std vs MAE | correlation r: 0.90 -> 0.92 | No clear effect |
Closest papers by meaning
- Revealing the 3D Cosmic Web through Gravitationally Constrained Neural Fields (ICLR 2025 · similarity 0.91)
- Unified Framework for Diffusion Generative Models in SO(3): Applications in Computer Vision and Astrophysics (AAAI 2024 · similarity 0.88)
- A Data-Driven Prism: Multi-View Source Separation with Diffusion Model Priors (NeurIPS 2025 · similarity 0.86)
- CosmoBench: A Multiscale, Multiview, Multitask Cosmology Benchmark for Geometric Deep Learning (NeurIPS 2025 · similarity 0.85)
- Listening to the noise: Blind Denoising with Gibbs Diffusion (ICML 2024 · similarity 0.85)
- Inflationary Flows: Calibrated Bayesian Inference with Diffusion-Based Models (NeurIPS 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.
