Research map / Learning theory and algorithms / Matrix and tensor methods
PD-Sparse : A Primal and Dual Sparse Approach to Extreme Multiclass and Multilabel Classification
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Adding primal and dual sparsity (L1-L2 penalty with FC-BCFW) versus standard one-vs-all baselines | PD-Sparse · Multiclass and multilabel benchmarks (Tables 1-2) | test accuracy | Mixed |
| Using PD-Sparse instead of tree/low-rank methods on large multilabel data | PD-Sparse · LSHTC-wiki (K=320338) | test acc %: 89.3 vs 73.44 for SLEEC | Helped |
| Using PD-Sparse instead of tree-based FastXML on multiclass Dmoz | PD-Sparse · Dmoz (K=11947) | test acc(%): 39.58 vs 38.4 | Helped |
| Using PD-Sparse on imgNet multiclass | PD-Sparse · imgNet (K=1000) | test acc(%): 12.7 vs 15.25 for 1vsA-SVM | Hurt |
| Using PD-Sparse on EUR-Lex multilabel | PD-Sparse · EUR-Lex (K=3956) | test acc %: 76.3 vs 77.3 for 1vsA-Logi | Hurt |
| Using PD-Sparse on bibtex multilabel | PD-Sparse · bibtex (K=159) | test acc %: 64.55 vs 65.09 for SLEEC | Hurt |
Closest papers by meaning
- Sparse Extreme Multi-label Learning with Oracle Property (ICML 2019 · similarity 0.89)
- Taming the Sigmoid Bottleneck: Provably Argmaxable Sparse Multi-Label Classification (AAAI 2024 · similarity 0.89)
- On Controllable Sparse Alternatives to Softmax (NeurIPS 2018 · similarity 0.88)
- Multilabel Classification by Hierarchical Partitioning and Data-dependent Grouping (NeurIPS 2020 · similarity 0.88)
- Semi-Supervised Sparse Gaussian Classification: Provable Benefits of Unlabeled Data (NeurIPS 2024 · similarity 0.86)
- Navigating Extremes: Dynamic Sparsity in Large Output Spaces (NeurIPS 2024 · similarity 0.86)
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.