Research map / Learning theory and algorithms / Matrix and tensor methods
More Efficient Sampling for Tensor Decomposition With Worst-Case Guarantees
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 CP-ALS-ES instead of CP-ARLS-LEV | CP decomposition feature extraction · COIL-100 | Time (s): 28.7 -> 27.9 | Helped |
| Use CP-ALS-ES instead of CPD-MINF | CP decomposition feature extraction · COIL-100 | Acc. (%): 99.7 -> 98.3 | Hurt |
| Use TR-ALS-ES instead of TR-ALS-Sampled | Tensor ring feature extraction · COIL-100 | Time (s): 9.9 -> 28.5 | Hurt |
| Use TR-ALS-ES instead of TR-ALS-Sampled | Tensor ring feature extraction · COIL-100 | Acc. (%): 98.5 -> 98.0 | Hurt |
| Replace CP-ARLS-LEV sampling with CP-ALS-ES approximate leverage sampling at J1 = 1e+3 | CP least-squares sampling · Cat tensor, CP-ALS least squares, rank 10 | KL-divergence: 0.2342 -> 0.0151 | Helped |
| Replace TR-ALS-Sampled sampling with TR-ALS-ES approximate leverage sampling at J1 = 1e+3 | Tensor ring least-squares sampling · Cat tensor, TR-ALS least squares, rank 3 | KL-divergence: 0.3087 -> 0.0076 | Helped |
| Use CP-ALS-ES instead of CP-ARLS-LEV | CP decomposition · Synthetic 10-way tensor, rank 4 | Time (s): 350 -> 4.8 | Helped |
| Increase recursive sketch size to J1 = 1e+4 (ablation) | CP-ALS-ES · Cat tensor, CP-ALS least squares, rank 10 | KL-divergence: 0.0151 -> 0.0005 | Helped |
| Reduce recursive sketch size to J1 = 1e+2 (ablation) | CP-ALS-ES · Cat tensor, CP-ALS least squares, rank 10 | KL-divergence: 0.0151 -> 0.1416 | Hurt |
| Increase recursive sketch size to J1 = 1e+4 (ablation) | TR-ALS-ES · Cat tensor, TR-ALS least squares, rank 3 | KL-divergence: 0.0076 -> 0.0005 | Helped |
| Reduce recursive sketch size to J1 = 1e+2 (ablation) | TR-ALS-ES · Cat tensor, TR-ALS least squares, rank 3 | KL-divergence: 0.0076 -> 0.1565 | Hurt |
Closest papers by meaning
- A Sampling-Based Method for Tensor Ring Decomposition (ICML 2021 · similarity 0.94)
- SPALS: Fast Alternating Least Squares via Implicit Leverage Scores Sampling (NeurIPS 2016 · similarity 0.92)
- Efficient Leverage Score Sampling for Tensor Train Decomposition (NeurIPS 2024 · similarity 0.91)
- Fused Orthogonal Alternating Least Squares for Tensor Clustering (NeurIPS 2022 · similarity 0.90)
- Towards Multi-Mode Outlier Robust Tensor Ring Decomposition (AAAI 2024 · similarity 0.89)
- Fast Tensor Completion via Approximate Richardson Iteration (ICML 2025 · 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.