Research map / Learning theory and algorithms / Matrix and tensor methods
Error Analysis of Tensor-Train Cross Approximation
Zhen Qin, Alexander Lidiak, Zhexuan Gong, Gongguo Tang, Michael Wakin, Zhihui Zhu
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 |
|---|---|---|---|
| truncated pseudo-inverse (tau>0) instead of exact pseudo-inverse when r'>r (ablation) | TT cross approximation · Synthetic TT tensors, N=10, d=2, r=2 | MSE (dB) | Improved |
| overestimate rank r' > r with exact pseudo-inverse (ablation) | TT cross approximation · Synthetic TT tensors, N=10, d=2, r=2 | MSE (dB) | Hurt |
| use exact rank r'=r (ablation) | TT cross approximation · Synthetic TT tensors, N=10, d=2, r=2 | MSE (dB) | No clear effect |
| larger tau relative to singular values (ablation) | TT cross approximation · Synthetic TT tensors, N=10, d=2, r=2 | MSE (dB) | Hurt |
| larger r' with truncated pseudo-inverse (ablation) | TT cross approximation · Synthetic TT tensors, N=10, d=2, r=2 | MSE (dB) | Improved |
| increase tensor order N from 6 to 22 with r'>r and exact pseudo-inverse (ablation) | TT cross approximation · Synthetic TT tensors, d=2, r=2 | MSE (dB) | Hurt |
| increase tensor order N from 6 to 22 with r'=r (ablation) | TT cross approximation · Synthetic TT tensors, d=2, r=2 | MSE (dB) | No clear effect |
| increase deviation eta from TT format (ablation) | TT cross approximation with truncated pseudo-inverse · Synthetic TT tensors, r'=3 to 5 | MSE (dB) | No clear effect |
| increase measurement noise mu (ablation) | TT cross approximation with truncated pseudo-inverse · Synthetic TT tensors, r'=3 to 5 | MSE (dB) | No clear effect |
| increase true rank r (ablation) | TT cross approximation with truncated pseudo-inverse · Synthetic TT tensors, r'=8 to 10 | MSE (dB) | Hurt |
| increase tensor order N from 5 to 50 on exact TT tensors (ablation) | TT cross approximation · Exact TT tensors, r'=r=2 | MSE (dB) | Hurt |
Closest papers by meaning
- Efficient Leverage Score Sampling for Tensor Train Decomposition (NeurIPS 2024 · similarity 0.88)
- Many-body Approximation for Non-negative Tensors (NeurIPS 2023 · similarity 0.87)
- Expressive power of tensor-network factorizations for probabilistic modeling (NeurIPS 2019 · similarity 0.87)
- Guaranteed Nonconvex Factorization Approach for Tensor Train Recovery (JMLR 2024 · similarity 0.86)
- Statistical Estimation in the Spiked Tensor Model via the Quantum Approximate Optimization Algorithm (NeurIPS 2024 · similarity 0.84)
- Higher-Rank Irreducible Cartesian Tensors for Equivariant Message Passing (NeurIPS 2024 · similarity 0.82)
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.
