Research map / Learning theory and algorithms / Matrix and tensor methods
Low-Rank Tensor Learning by Generalized Nonconvex Regularization
Sijia Xia, Michael K. Ng, Xiongjun Zhang
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 |
|---|---|---|---|
| replace convex TTNN regularizer with nonconvex LFNR regularizer (ablation) | TTNN tensor completion · Balloons (sigma=0.005, SR 0.05) | PSNR: 33.91 -> 35.65 | Improved |
| replace convex TTNN regularizer with nonconvex LFNR regularizer (ablation) | TTNN tensor completion · Lemons (sigma=0.01, SR 0.30) | PSNR: 44.90 -> 47.00 | Improved |
| replace convex TTNN regularizer with nonconvex LFNR regularizer (ablation) | TTNN tensor completion · Lemons (sigma=0.005, SR 0.05) | SSIM: 0.9265 -> 0.9554 | Improved |
| replace convex TTNN regularizer with nonconvex LFNR regularizer (ablation) | TTNN tensor completion · MRI (Brainweb, sigma=0.01) | PSNR: at least 1dB improvement | Improved |
| replace convex TTNN regularizer with nonconvex LFNR regularizer (ablation) | TTNN tensor completion · MRI (Brainweb, sigma=0.01, SR 0.10) | SSIM: 0.4567 -> 0.5136 | Improved |
| use TTNN-based nonconvex regularization instead of unfolding-based NORT (ablation) | LFNR vs NORT · MRI (Brainweb, sigma=0.01, SR 0.50) | PSNR: 26.11 -> 29.68 | Improved |
| use TTNN instead of sum of nuclear norms of unfoldings (SNN) (ablation) | TTNN vs SNN · Balloons (sigma=0.005, SR 0.05) | PSNR: 24.50 -> 33.91 | Improved |
| use TTNN instead of NORT (ablation) | TTNN vs NORT · Balloons (sigma=0.01, SR 0.05) | PSNR: 33.60 -> 33.24 | Hurt |
| use TTNN instead of TMac at low sampling (ablation) | TTNN vs TMac · MRI (Brainweb, sigma=0.01, SR 0.10) | PSNR: 19.85 -> 19.13 | Hurt |
| use TTNN instead of overlapped trace norm (LOTN) (ablation) | TTNN vs LOTN · CIFAR-10 binary (dog-frog) | test accuracy: 0.696 -> 0.747 | Improved |
| use TTNN instead of DuSK (ablation) | TTNN vs DuSK · CIFAR-10 binary (bird-deer) | test accuracy: 0.719 -> 0.715 | Hurt |
| use LFNR instead of SVM | LFNR vs SVM · CIFAR-10 binary (bird-airplane) | test accuracy: 0.623 -> 0.812 | Improved |
Closest papers by meaning
- Low-rank Tensor Learning with Nonconvex Overlapped Nuclear Norm Regularization (JMLR 2022 · similarity 0.94)
- Nonconvex Low-Rank Tensor Completion from Noisy Data (NeurIPS 2019 · similarity 0.92)
- Efficient Nonconvex Regularized Tensor Completion with Structure-aware Proximal Iterations (ICML 2019 · similarity 0.92)
- Low-Rank Tensor Completion: A Pseudo-Bayesian Learning Approach (ICCV 2017 · similarity 0.92)
- Low-Rank Tensor Completion by Approximating the Tensor Average Rank (ICCV 2021 · similarity 0.91)
- Tensor Ring Decomposition with Rank Minimization on Latent Space: An Efficient Approach for Tensor Completion (AAAI 2019 · similarity 0.90)
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.
