Research map / Optimization / Nonconvex and smooth optimization
A Unified Optimization Framework for Low-Rank Inducing Penalties
Check 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 |
|---|---|---|---|
| Combine rank-threshold parameters b with singular-value shrinkage parameters a | Unified regularizer Rh · Synthetic missing data, Uniform 0% | distance to ground truth (normalized): 0.0229 -> 0.0199 | Helped |
| Combine rank-threshold parameters b with singular-value shrinkage parameters a | Unified regularizer Rh · Synthetic missing data, Uniform 20% | distance to ground truth (normalized): 0.0233 -> 0.0198 | Helped |
| Use the unified regularizer rather than the strongest listed baseline | Synthetic missing data · Uniform 80% | distance to ground truth (normalized): 0.2305 -> 0.3117 | Hurt |
| Use the unified regularizer rather than the strongest listed baseline | Synthetic missing data · Tracking 50% | distance to ground truth (normalized): 0.2847 -> 0.3128 | Hurt |
| Increase C from 10⁻² to 1, favoring the weighted nuclear norm penalty (ablation) | Unified regularizer on the book sequence · Non-rigid deformation with missing data | datafit: 0.8354 -> 0.4485 | Helped |
| Increase C from 10⁻² to 1, favoring the weighted nuclear norm penalty (ablation) | Unified regularizer on the book sequence · Non-rigid deformation with missing data | reconstruction of missing points | Hurt |
| Increase C from 1 to 100 (ablation) | Unified regularizer on the book sequence · Non-rigid deformation with missing data | datafit: 0.4485 -> 6.5221 | Hurt |
| Replace the soft-rank penalty with the WNNM special case (bi ≡ 0) (ablation) | Prior-free NRSfM · CMU MOCAP sequences | distance to ground truth | Mixed |
Closest papers by meaning
- Bilinear Parameterization for Non-Separable Singular Value Penalties (CVPR 2021 · similarity 0.88)
- An Alternating Proximal Splitting Method with Global Convergence for Nonconvex Structured Sparsity Optimization (AAAI 2016 · similarity 0.87)
- Accelerated Cyclic Coordinate Dual Averaging with Extrapolation for Composite Convex Optimization (ICML 2023 · similarity 0.86)
- Inertial Block Proximal Methods for Non-Convex Non-Smooth Optimization (ICML 2020 · similarity 0.86)
- Low-Rank Extragradient Method for Nonsmooth and Low-Rank Matrix Optimization Problems (NeurIPS 2021 · similarity 0.86)
- Cauchy-Schwarz Regularizers (ICLR 2025 · similarity 0.85)
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.