Research map / Label-efficient and robust learning / Recognition and long-tail
A Unified View on Multi-class Support Vector Classification
Ürün Doğan, Tobias Glasmachers, Christian Igel
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 |
|---|---|---|---|
| use self-margin only (MMR) instead of relative margins (WW) | MMR · Vehicle | test accuracy (%): 84.28 -> 73.94 | Hurt |
| use self-margin only (MMR) for training speed | MMR · Car | training time (s): 131.7 -> 12.89 | Improved |
| use absolute margin with max aggregation (AMO) instead of WW | AMO · Abalone | test accuracy (%): 27.51 -> 20.42 | Hurt |
| use absolute margin (ATS) with sum aggregation instead of WW | ATS · Abalone | test accuracy (%): 27.51 -> 26.64 | Hurt |
| use absolute margin AMO instead of CS | AMO · Glass | test accuracy (%): 69.03 -> 69.52 | Improved |
| add relative margins (RM) versus OVA | RM · Soy bean | test accuracy (%): 89.58 -> 91.16 | Improved |
| use absolute margin machines instead of WW/CS | AMO, ATS, ATM · non-linear benchmarks | training time | Hurt |
| tighten stopping threshold to 10^-5 (ablation) | SVM solvers · non-linear benchmarks | test accuracy | No clear effect |
| use a different loss function in multi-class SVM | all SVMs · support vector fraction | sparsity | No clear effect |
Closest papers by meaning
- Multi-Class Support Vector Machine with Maximizing Minimum Margin (AAAI 2024 · similarity 0.90)
- Support Vector Machine Classifier via Soft-Margin Loss (TPAMI 2021 · similarity 0.90)
- Fast Support Vector Classification for Large-Scale Problems (TPAMI 2021 · similarity 0.90)
- GenSVM: A Generalized Multiclass Support Vector Machine (JMLR 2016 · similarity 0.89)
- Weston-Watkins Hinge Loss and Ordered Partitions (NeurIPS 2020 · similarity 0.89)
- Solving L1-regularized SVMs and Related Linear Programs: Revisiting the Effectiveness of Column and Constraint Generation (JMLR 2022 · similarity 0.89)
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.
