Research map / Learning theory and algorithms / Matrix and tensor methods
Rank-one Convexification for Sparse Regression
Alper Atamturk, Andres Gomez
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 sdp2 (rank-one on two variables) instead of sdp1 | sdp1 · auto MPG (lambda=0) | optimality gap (%): 55.8 -> 4.0 | Improved |
| use sdp2 instead of sdp1 | sdp1 · breast cancer (lambda=0) | optimality gap (%): 14.4 -> 2.4 | Improved |
| use sdp2 instead of sdp1 | sdp1 · diabetes (lambda=0) | optimality gap (%): 22.2 -> 8.2 | Improved |
| use sdp2 instead of sdp1 | sdp1 · crime (lambda=0.05) | optimality gap (%): 1.1 -> 0.5 | Improved |
| add the sdp2 relaxation versus big-M MIO on the largest dataset | big-M · crime (lambda=0) | optimality gap (%): 3.7 -> 2.7 | Improved |
| use sdp2 instead of big-M MIO for solve time | big-M · crime (lambda=0) | solve time (s): 527.03 -> 45.29 | Improved |
| use sdp2 instead of persp MIO for solve time | persp · crime (lambda=0.05) | solve time (s): 518.03 -> 43.86 | Improved |
| use sdp2 instead of sdp1 at the cost of compute | sdp1 · diabetes (lambda=0) | solve time (s): 3.58 -> 9.28 | Hurt |
| use sdpLB relaxation instead of sdp1 | sdp1 · crime (lambda=0) | optimality gap (%): 5.6 -> 16.0 | Hurt |
| raise rank-one dimension r from 2 to 3 (ablation) | sdp2 · real instances (lambda=0 and 0.05) | optimality gap | No clear effect |
| raise r above 4 when p=15, n=10, lambda=0 (ablation) | sdpr · synthetic p=15, n=10 (lambda=0) | optimality gap (%): 100% gap for r<=4 | Improved |
| increase ridge regularization lambda (ablation) | sdpr · synthetic p=15, n=10 | optimality gap | Improved |
Closest papers by meaning
- Robust Sparse Regression with Non-Isotropic Designs (NeurIPS 2024 · similarity 0.88)
- Accelerated Sparse Linear Regression via Random Projection (AAAI 2016 · similarity 0.88)
- Hardness and Algorithms for Robust and Sparse Optimization (ICML 2022 · similarity 0.88)
- On the Statistical Limits of Convex Relaxations (ICML 2016 · similarity 0.88)
- A Constructive Approach to $L_0$ Penalized Regression (JMLR 2018 · similarity 0.88)
- Sharp Oracle Inequalities for Square Root Regularization (JMLR 2017 · similarity 0.88)
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.
