Research map / Label-efficient and robust learning / Self-supervised and contrastive
ATD: Augmenting CP Tensor Decomposition by Self Supervision
Chaoqi Yang, Cheng Qian, Navjot Singh, Cao Xiao, M Brandon Westover, Edgar Solomonik, Jimeng Sun
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 |
|---|---|---|---|
| Add the self-supervised alignment loss to tensor decomposition (ablation) | ATD · Sleep-EDF | accuracy: 84.19 ± 0.221 -> 85.01 ± 0.224 | Improved |
| Add the self-supervised alignment loss to tensor decomposition (ablation) | ATD · HAR | accuracy: 92.41 ± 0.391 -> 93.35 ± 0.357 | Improved |
| Add the self-supervised alignment loss to tensor decomposition (ablation) | ATD · PTB-XL | accuracy: 69.38 ± 0.612 -> 70.26 ± 0.523 | Improved |
| Add the self-supervised alignment loss to tensor decomposition (ablation) | ATD · MGH | accuracy: 72.78 ± 0.522 -> 74.15 ± 0.431 | Improved |
| Use two iterative-rule rounds instead of one (ablation) | ATD · HAR | accuracy: 93.30 ± 0.413 -> 93.35 ± 0.172 | No clear effect |
| Use two iterative-rule rounds instead of one (ablation) | ATD · Sleep-EDF | accuracy: 85.25 ± 0.209 -> 85.33 ± 0.173 | No clear effect |
| Use more diverse augmentation methods (ablation) | ATD · MGH Sleep | accuracy: 72.53 ± 0.652 -> 74.15 ± 0.203 | Improved |
| Combine jittering, bandpass filtering, and 3D position rotation rather than use one augmentation (ablation) | ATD · HAR | accuracy: 92.06 ± 0.621 -> 93.35 ± 0.357 | Improved |
| Increase decomposition rank from R=8 to R=64 (ablation) | ATD · HAR | accuracy: 71.91 ± 0.253 -> 93.43 ± 0.411 | Improved |
| Increase decomposition rank from R=64 to R=128 (ablation) | ATD · HAR | accuracy: 93.43 ± 0.411 -> 92.97 ± 0.273 | Hurt |
| Increase decomposition rank from R=8 to R=128 (ablation) | ATD · Sleep-EDF | accuracy: 82.49 ± 0.464 -> 85.32 ± 0.305 | Improved |
| Use increasingly unrealistic jittering noise (ablation) | ATD · MGH | accuracy: 74.18 ± 0.326 -> 72.18 ± 0.676 | Hurt |
Closest papers by meaning
- Computational and Statistical Guarantees for Tensor-on-Tensor Regression With Tensor Train Decomposition (TPAMI 2025 · similarity 0.86)
- Guaranteed Nonconvex Low-Rank Tensor Estimation via Scaled Gradient Descent (JMLR 2026 · similarity 0.86)
- Factor Augmented Tensor-on-Tensor Neural Networks (AAAI 2025 · similarity 0.86)
- High-Dimensional Tensor Regression With Oracle Properties (ICML 2025 · similarity 0.84)
- Over-parameterized Student Model via Tensor Decomposition Boosted Knowledge Distillation (NeurIPS 2024 · similarity 0.84)
- Factor Augmented High-Dimensional SGD (NeurIPS 2026 · similarity 0.84)
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.
