Research map / Probabilistic and scientific modeling / Variational and generative sampling
Proximal Deep Structured Models
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 |
|---|---|---|---|
| Use proximal deep structured model instead of RTF (second best) | denoising · BSDS sigma=25 | PSNR: 28.75 -> 28.79 | Helped |
| Use proposed model vs RTF in runtime | denoising · BSDS sigma=25 | time (s): 69.25 -> 0.23 | Helped |
| Run proposed model on GPU | proximal deep structured model · BSDS sigma=25 | time (s): 0.23 -> 0.011 | Helped |
| Increase filter size from 3x3 to 7x7 (64 filters) (ablation) | proximal deep structured model · BSDS denoising | PSNR: 28.49 -> 28.79 | Helped |
| Increase filters per layer from 16 to 64 (7x7) (ablation) | proximal deep structured model · BSDS denoising | PSNR: 28.68 -> 28.79 | Helped |
| Increase filter size from 3x3 to 7x7 (16 filters) (ablation) | proximal deep structured model · BSDS denoising | PSNR: 28.43 -> 28.68 | Helped |
| Use proposed model vs FilterForest on depth | depth refinement · 7 Scenes | PSNR: 35.63 -> 36.31 | Helped |
| Use proposed model vs BM3D on depth | depth refinement · 7 Scenes | PSNR: 35.46 -> 36.31 | Helped |
| Add high-order structured proximal layers on Flownet | optical flow · Flying Chairs | end-point error: 4.98 -> 4.91 | Helped |
| Add TV-l1 post-processing to Flownet | optical flow · Flying Chairs | end-point error: 4.98 -> 4.96 | No clear effect |
Closest papers by meaning
- Adaptive Proximal Gradient Methods for Structured Neural Networks (NeurIPS 2021 · similarity 0.87)
- What's in a Prior? Learned Proximal Networks for Inverse Problems (ICLR 2024 · similarity 0.87)
- Sparse Flows: Pruning Continuous-depth Models (NeurIPS 2021 · similarity 0.85)
- Non-parametric Structured Output Networks (NeurIPS 2017 · similarity 0.85)
- Proximal Mapping for Deep Regularization (NeurIPS 2020 · similarity 0.85)
- BMRS: Bayesian Model Reduction for Structured Pruning (NeurIPS 2024 · similarity 0.83)
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.