Research map / Probabilistic and scientific modeling / Variational and generative sampling
Uncertainty Modeling in Generative Compressed Sensing
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Add uncertainty modeling (variational latent z and MAP-inferred parameters theta with Gaussian prior) to CSGM generator, expanding its range | CSGM generative compressed sensing · MNIST, m between 100 and 400 | per-pixel l2 reconstruction loss | Helped |
| Add uncertainty modeling to CSGM generator | CSGM generative compressed sensing · CelebA, m from 20 to 104 | per-pixel l2 reconstruction loss | Helped |
| Use Sparse-Gen-style sparse deviation vector instead of CS-BGM uncertainty modeling at large measurement count | MNIST, m of 500 or larger · MNIST | per-pixel l2 reconstruction loss | No clear effect |
| Replace VI on z with MAP estimation (CS-BGM w/o MC) (ablation) | CS-BGM · MNIST and CelebA | reconstruction error | No clear effect |
| Replace Lasso baseline with CS-BGM on CelebA DCT/DWT sparse basis | Lasso · CelebA, m=1000 | qualitative reconstruction | Hurt |
Closest papers by meaning
- Sparse Bayesian Generative Modeling for Compressive Sensing (NeurIPS 2024 · similarity 0.90)
- Modeling Sparse Deviations for Compressed Sensing using Generative Models (ICML 2018 · similarity 0.89)
- Robust Compressed Sensing MRI with Deep Generative Priors (NeurIPS 2021 · similarity 0.89)
- Robust compressed sensing using generative models (NeurIPS 2020 · similarity 0.89)
- QCS-SGM+: Improved Quantized Compressed Sensing with Score-Based Generative Models (AAAI 2024 · similarity 0.85)
- Recovery Guarantees for Posterior Sampling of One-Bit Compressed Sensing (NeurIPS 2026 · 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.