Research map / Label-efficient and robust learning / Recognition and long-tail
Joint Active Learning with Feature Selection via CUR Matrix Decomposition
Changsheng Li, Xiangfeng Wang, Weishan Dong, Junchi Yan, Qingshan Liu, Hongyuan Zha B
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 local linear reconstruction (ALFS-II over ALFS-I) (ablation) | ALFS · Madelon (SVM) | accuracy (%): 61.9 -> 62.7 | Improved |
| add local linear reconstruction (ablation) | ALFS · HAR (decision tree) | accuracy (%): 91.9 -> 93.1 | Improved |
| add local linear reconstruction (ablation) | ALFS · ORL (decision tree) | accuracy (%): 35.1 -> 37.9 | Improved |
| add local linear reconstruction (ablation) | ALFS · HAR (SVM) | accuracy (%): 95.3 -> 95.3 | No clear effect |
| add local linear reconstruction (ablation) | ALFS · CLL_SUB_111 (SVM) | accuracy (%): 55.0 -> 55.9 | Improved |
| joint sample and feature selection instead of ALNR (SVM, Madelon, 1200 samples) (ablation) | ALFS-I · Madelon (SVM) | relative improvement: 14.8% relative improvement over ALNR | Improved |
| joint sample and feature selection instead of ALNR (decision tree, Madelon) (ablation) | ALFS-II · Madelon (decision tree) | relative improvement: 12.3% relative improvement over ALNR | Improved |
| select 50 features before active learning instead of using all 5748 features (ablation) | ALFS-II · TOX-171 (SVM, 10 queries) | accuracy (%): 40.8 -> 49.0 | Improved |
| select features before active learning instead of all 5748 features (ablation) | ALFS-II · TOX-171 (decision tree, 70 queries) | accuracy (%): 65.8 -> 64.8 | Hurt |
| jointly select features with samples instead of Laplacian+TED pipeline (ablation) | ALFS-II · TOX-171 (SVM, 50 features) | accuracy (%): 60.8 -> 69.3 | Improved |
| use ALFS optimization instead of randomized R-CUR (ablation) | ALFS-II · Musk (SVM, 10 features) | accuracy (%): 71.9 -> 77.4 | Improved |
Closest papers by meaning
- Active Feature Selection for the Mutual Information Criterion (AAAI 2021 · similarity 0.87)
- Unsupervised Active Learning via Subspace Learning (AAAI 2021 · similarity 0.87)
- Joint Representative Selection and Feature Learning: A Semi-Supervised Approach (CVPR 2019 · similarity 0.87)
- Joint Out-of-Distribution Filtering and Data Discovery Active Learning (CVPR 2025 · similarity 0.85)
- Towards a statistical theory of data selection under weak supervision (ICLR 2024 · similarity 0.84)
- Enhanced Sample Selection with Confidence Tracking: Identifying Correctly Labeled Yet Hard-to-Learn Samples in Noisy Data (AAAI 2025 · 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.
