Research map / Label-efficient and robust learning / Recognition and long-tail
Learning Sparse Confidence-Weighted Classifier on Very High Dimensional Data
Mingkui Tan, Yan Yan, Li Wang, Anton Van Den Hengel, Ivor W. Tsang, Qinfeng (Javen) Shi
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 |
|---|---|---|---|
| BCWL1 instead of SCW (best second-order CW baseline) | online CW learning · covtype 450k | test error rate: 0.2382 ± 0.0003 -> 0.2402 ± 0.0003 | Improved |
| BCWL1 instead of AROW | online CW learning · covtype 450k | training time (s): 21.72 -> 0.41(0.06) | Improved |
| BCWL1 instead of AROW | online CW learning · epsilon 50k | training time (s): 3506.0 -> 5.73(0.39) | Improved |
| BCWL1 instead of AROW | online CW learning · epsilon 50k | test error rate: 0.1105 ± 0.0004 -> 0.1093 ± 0.0000 | Improved |
| BCWL1 instead of PA (no covariance) | online learning · covtype 450k | test error rate: 0.3443 ± 0.0433 -> 0.2402 ± 0.0003 | Improved |
| BCWL1 instead of PA on entire epsilon | online learning · epsilon (entire set) | test error rate: 0.1821 -> 0.1019 | Improved |
| BCWL1 instead of batch DCD | batch learning · epsilon (entire set) | training time (s): 81.52 -> 32.06(1.19) | Improved |
| BCWL1 instead of DCD | batch learning · epsilon (entire set) | test error rate: 0.1021 -> 0.1019 | No clear effect |
| BCWL2 instead of DCD | batch learning · covtype 450k | test error rate: 0.2460 -> 0.2462 ± 0.0005 | No clear effect |
| BCWL2 instead of BCWL1 (squared hinge versus hinge) | BCW · covtype 450k | test error rate: 0.2402 ± 0.0003 -> 0.2462 ± 0.0005 | Hurt |
| BCWL1 instead of DCD on epsilon 50k | batch learning · epsilon 50k | training time (s): 2.62 -> 5.73(0.39) | Hurt |
| SBCW with block covariance instead of FGM, L1SVM, sparDCA | feature selection · high-dimensional data sets | accuracy | Mixed |
Closest papers by meaning
- Adaptive Sparse Confidence-Weighted Learning for Online Feature Selection (AAAI 2019 · similarity 0.87)
- Polynomial-Time Robust Multiclass Linear Classification under Gaussian Marginals (NeurIPS 2026 · similarity 0.83)
- Confidence Calibration of Classifiers with Many Classes (NeurIPS 2024 · similarity 0.82)
- Stabilizing Linear Passive-Aggressive Online Learning with Weighted Reservoir Sampling (NeurIPS 2024 · similarity 0.82)
- Semi-Supervised Sparse Gaussian Classification: Provable Benefits of Unlabeled Data (NeurIPS 2024 · similarity 0.82)
- Improving Generalization via Meta-Learning on Hard Samples (CVPR 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.
