Research map / Learning theory and algorithms / Matrix and tensor methods
Fitting Low-Rank Tensors in Constant Time
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 |
|---|---|---|---|
| sample indices per mode (s=40 or 80) instead of full tensor for residual estimation | Tucker residual error estimation · Real datasets (8 tensors) | absolute error vs HOOI residual (averaged over ranks) | Helped |
| replace HOOI with sampling-based residual estimate (sample80) | Tucker rank ordering · visor, EEM, movie_gray | Kendall's tau vs HOOI ranking | Helped |
| sample size s=80 vs s=40 (ablation) | Sampling estimator · Real datasets | Kendall's tau and absolute error | Helped |
| sampling-based residual estimate on sparse 'wine' tensor | Tucker residual estimation · wine | absolute error vs HOOI | Mixed |
| randomized SVD replaced by sampling for residual estimation | Runtime · Real datasets | runtime vs HOOI | Helped |
| MACH sparsification (p=0.1 or 0.3) | Tucker residual estimation · Real datasets | absolute error vs HOOI | Hurt |
| small sample size s (synthetic) (ablation) | Sampling estimator · Synthetic N=100-800 | estimated residual error | Hurt |
Closest papers by meaning
- Low-Rank Tucker Decomposition of Large Tensors Using TensorSketch (NeurIPS 2018 · similarity 0.92)
- Fast Tucker Rank Reduction for Non-Negative Tensors Using Mean-Field Approximation (NeurIPS 2021 · similarity 0.90)
- Fast and accurate randomized algorithms for low-rank tensor decompositions (NeurIPS 2021 · similarity 0.90)
- Fast Tensor Completion via Approximate Richardson Iteration (ICML 2025 · similarity 0.88)
- Functional Bayesian Tucker Decomposition for Continuous-indexed Tensor Data (ICLR 2024 · similarity 0.87)
- Efficient Leverage Score Sampling for Tensor Train Decomposition (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.