Research map / Learning theory and algorithms / Matrix and tensor methods
Learning-Based Low-Rank Approximations
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 |
|---|---|---|---|
| Train the sparse sketch matrix on examples from the same dataset | SCW low-rank approximation · Logo, Eagle, Friends, Hyper, and Tech (k=10, m=20) | test error: 2.09 -> 0.10; 4.31 -> 0.18; 4.11 -> 0.22; 2.92 -> 0.52; 7.99 -> 2.95 | Helped |
| Use a mixed training set from Logo, Eagle, and Friends instead of training only on Logo | Learned sketch for Logo test matrices · Logo | test error: 0.27 -> 0.67 | Hurt |
| Replace a random sketch with a half-trained, half-random sketch | SCW low-rank approximation · Logo, Hyper, and Tech (k=10, m=20) | test error: 2.09 -> 0.20; 2.92 -> 0.78; 7.99 -> 3.73 | Helped |
| Train one half of the sketch separately, then append a random half | SCW low-rank approximation · Logo, Hyper, and Tech (k=10, m=20) | test error: 2.09 -> 0.24; 2.92 -> 0.87; 7.99 -> 3.69 | Helped |
Closest papers by meaning
- Few-Shot Data-Driven Algorithms for Low Rank Approximation (NeurIPS 2021 · similarity 0.93)
- Additive Error Guarantees for Weighted Low Rank Approximation (ICML 2021 · similarity 0.91)
- Approximation Algorithms for $\ell_0$-Low Rank Approximation (NeurIPS 2017 · similarity 0.91)
- Efficient Alternating Minimization with Applications to Weighted Low Rank Approximation (ICLR 2025 · similarity 0.89)
- Low Rank Matrix Completion via Robust Alternating Minimization in Nearly Linear Time (ICLR 2024 · similarity 0.89)
- Entrywise error bounds for low-rank approximations of kernel matrices (NeurIPS 2024 · similarity 0.87)
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.