Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Variational Inference for Deblending Crowded Starfields
Runjing Liu, Jon D. McAuliffe, Jeffrey Regier, The LSST Dark Energy Science Collaboration
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 StarNet instead of PCAT | StarNet · M2 globular cluster subregion A (100x100, HST ground truth) | F1 score: 0.44 -> 0.50 | Improved |
| use StarNet instead of PCAT | StarNet · M2 globular cluster subregion A (100x100, HST ground truth) | PPV: 0.37 -> 0.48 | Improved |
| use StarNet instead of DAOPHOT | StarNet · M2 globular cluster subregion A (100x100, HST ground truth) | F1 score: 0.31 -> 0.50 | Improved |
| use StarNet instead of DAOPHOT on dense core region C | StarNet · M2 subregion C (center) | F1 score: 0.007 -> 0.13 | Improved |
| train with expected forward KL instead of ELBO with reparameterized gradients (ablation) | StarNet inference network · simulated 20x20 image | convergence across restarts | Improved |
| use reparameterized gradients instead of REINFORCE for ELBO (ablation) | ELBO optimization · simulated 20x20 image | optimization convergence | Improved |
| use 3 pixels of tile padding instead of 1 (ablation) | StarNet · tile boundary star counts | overestimation of sources near tile edges | Improved |
| constrain stars to be at least 0.1 pixels from tile edges (ablation) | StarNet posterior · synthetic 100x100 images | credible interval coverage | Improved |
| vary prior mean mu from 0.08 to 0.20 (ablation) | StarNet · M2 subregion A | F1 score: 0.50 -> 0.49 (range 0.48 to 0.51) | No clear effect |
| vary flux slope alpha from 0.25 to 1.0 (ablation) | StarNet · M2 subregion A | F1 score: 0.51 -> 0.51 | No clear effect |
| random refits of StarNet (ten seeds) (ablation) | StarNet · M2 subregion A | F1 score: 0.50 to 0.55 | No clear effect |
Closest papers by meaning
- A New Statistical Model of Star Speckles for Learning to Detect and Characterize Exoplanets in Direct Imaging Observations (CVPR 2025 · similarity 0.83)
- Variational Inference with Coverage Guarantees in Simulation-Based Inference (ICML 2024 · similarity 0.83)
- A Data-Driven Prism: Multi-View Source Separation with Diffusion Model Priors (NeurIPS 2025 · similarity 0.82)
- A systematic evaluation of vision-language models for observational astronomical reasoning tasks (NeurIPS 2026 · similarity 0.82)
- The Multimodal Universe: Enabling Large-Scale Machine Learning with 100 TB of Astronomical Scientific Data (NeurIPS 2024 · similarity 0.82)
- Unified Framework for Diffusion Generative Models in SO(3): Applications in Computer Vision and Astrophysics (AAAI 2024 · similarity 0.82)
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.
