Research map / Image generation / Fast diffusion sampling
Mapping Dark-Matter Clusters via Physics-Guided Diffusion Models
Diego Royo, Brandon Zhao, Adolfo Muñoz, Diego Gutierrez, Katherine 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 physics-guided diffusion posterior sampling instead of photometry-only UNet | UNet · DarkClusters-15k TNG test (z_L=0.5) | PSNR: 31.1 -> 33.3 | Improved |
| use physics-guided diffusion posterior sampling instead of photometry-only UNet | UNet · DarkClusters-15k SIMBA test (z_L=0.5) | PSNR: 29.6 -> 30.9 | Improved |
| use diffusion method instead of Napier et al. light-traces-mass baseline | Napier et al. · DarkClusters-15k TNG test (z_L=0.5) | PSNR: 26.0 -> 33.3 | Improved |
| use diffusion method instead of Napier et al. | Napier et al. · DarkClusters-15k SIMBA test (z_L=0.5) | PCC: 0.616 -> 0.930 | Improved |
| use diffusion method instead of UNet | UNet · DarkClusters-15k TNG test (z_L=0.5) | PCC: 0.924 -> 0.975 | Improved |
| use diffusion method instead of UNet | UNet · DarkClusters-15k TNG test (z_L=0.5) | held-out kappa RMSE: 0.0067 -> 0.0058 | Improved |
| use diffusion method instead of UNet | UNet · DarkClusters-15k TNG test (z_L=0.5) | held-out alpha RMSE: 0.411 -> 0.530 | Hurt |
| use diffusion method instead of UNet | UNet · DarkClusters-15k SIMBA test (z_L=0.5) | held-out alpha RMSE: 1.597 -> 1.485 | Improved |
| add strong and weak lensing constraints to photometry-only reconstruction (ablation) | Ours (photometry only) · DarkClusters-15k ablation (WL SNR 0.01) | PSNR: improving the computed PSNR by more than five points | Improved |
| use diffusion method instead of Kaiser-Squires weak-lensing inversion | Kaiser-Squires · DarkClusters-15k TNG test (z_L=0.5) | PSNR: 20.9 -> 33.3 | Improved |
| replace DDPM++ with Diffusion Transformer prior backbone (ablation) | Ours · DarkClusters-15k | accuracy | No clear effect |
| apply to real MACS 1206 cluster versus expert-tuned Lenstool | Caminha et al. reconstruction · MACS 1206 (real cluster) | multiple-image RMS scatter (arcsec): 0.44 -> 0.56 | Hurt |
Closest papers by meaning
- CosmoBench: A Multiscale, Multiview, Multitask Cosmology Benchmark for Geometric Deep Learning (NeurIPS 2025 · similarity 0.85)
- Revealing the 3D Cosmic Web through Gravitationally Constrained Neural Fields (ICLR 2025 · similarity 0.85)
- Unified Framework for Diffusion Generative Models in SO(3): Applications in Computer Vision and Astrophysics (AAAI 2024 · similarity 0.83)
- VESTA: Visual Exploration with Statistical Tool Agents (NeurIPS 2026 · similarity 0.82)
- A Data-Driven Prism: Multi-View Source Separation with Diffusion Model Priors (NeurIPS 2025 · similarity 0.81)
- Galaxy Walker: Geometry-aware VLMs For Galaxy-scale Understanding (CVPR 2025 · 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.
