Research map / Learning theory and algorithms / Matrix and tensor methods
Random Matrix Analysis to Balance between Supervised and Unsupervised Learning under the Low Density Separation Assumption
Vasilii Feofanov, Malik Tiomoko, Aladin Virmaux
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 |
|---|---|---|---|
| select alpha_l and alpha_u by RMT-based theoretical error estimate instead of 10-fold CV | QLDS · books, electronics, kitchen, splice (Table 1) | classification error (%): books 26.03 vs 27.91; electronics 19.41 vs 20.1; kitchen 19.11 vs 19.98; splice 35.35 vs 37.02 | Improved |
| select alpha_l and alpha_u by RMT-based theoretical error estimate instead of 10-fold CV | QLDS · mushrooms, adult (Table 1) | classification error (%): mushrooms 8.49 vs 2.57; adult 32.88 vs 32.25 | Hurt |
| use both labeled and unlabeled terms with default alpha_l=1, alpha_u=1 instead of tuned alphas (ablation) | QLDS · books (Table 1) | classification error (%): 49.13 vs 26.03 for tuned theory-selected | Hurt |
| graph-based unlabeled-only term (alpha_l=0, alpha_u=1) vs LS-SVM labeled-only term (ablation) | QLDS · books, electronics, kitchen, splice, mushrooms | classification error (%): GB-SSL better on books (26.47 vs 37.47), electronics, kitchen, splice; LS-SVM better on adult (33.35 vs 36.28) and mushr | Mixed |
| oracle alpha selection on ground truth | QLDS · books, mushrooms, splice (Table 1) | classification error (%): books 25.7 vs 26.03 (th); mushrooms 1.75 vs 8.49 (th) | Improved |
Closest papers by meaning
- A Random Matrix Analysis and Improvement of Semi-Supervised Learning for Large Dimensional Data (JMLR 2018 · similarity 0.90)
- Minimum Density Hyperplanes (JMLR 2016 · similarity 0.86)
- Out-Of-Domain Unlabeled Data Improves Generalization (ICLR 2024 · similarity 0.85)
- Semi-Supervised Sparse Gaussian Classification: Provable Benefits of Unlabeled Data (NeurIPS 2024 · similarity 0.85)
- Continuous Partitioning for Graph-Based Semi-Supervised Learning (NeurIPS 2024 · similarity 0.84)
- Supervised Matrix Factorization: Local Landscape Analysis and Applications (ICML 2024 · 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.
