Research map / Learning theory and algorithms / Prediction and decision losses
Multi-label classification: do Hamming loss and subset accuracy really conflict with each other?
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 |
|---|---|---|---|
| Optimize Hamming Loss instead of Subset Loss | Linear multi-label classifier · emotions, image, scene, enron, rcv1-subset1, bibtex | Hamming Loss: emotions: 0.224 ± 0.015 -> 0.202 ± 0.019; image: 0.214 ± 0.013 -> 0.180 ± 0.002; scene: 0.142 ± 0.009 -> 0.103 ± 0.011; | Helped |
| Optimize Hamming Loss instead of Subset Loss | Linear multi-label classifier · emotions (6 labels), image (5 labels), scene (6 labels), enron (53 labels) | Subset Accuracy: emotions: 0.240 ± 0.021 -> 0.288 ± 0.026; image: 0.396 ± 0.031 -> 0.471 ± 0.004; scene: 0.515 ± 0.032 -> 0.628 ± 0.035; | Helped |
| Optimize Hamming Loss instead of Subset Loss | Linear multi-label classifier · rcv1-subset1 (101 labels), bibtex (159 labels) | Subset Accuracy: rcv1-subset1: 0.111 ± 0.004 -> 0.079 ± 0.008; bibtex: 0.198 ± 0.001 -> 0.190 ± 0.001 | Hurt |
| Optimize Subset Loss instead of Hamming Loss | Kernel-based multi-label classifier · Subset Accuracy, small label spaces | Subset Accuracy | Hurt |
| Optimize Subset Loss instead of Hamming Loss | Kernel-based multi-label classifier · Subset Accuracy, large label spaces | Subset Accuracy | Helped |
Closest papers by meaning
- Multi-Label Learning with Stronger Consistency Guarantees (NeurIPS 2024 · similarity 0.91)
- Regret Bounds for Multilabel Classification in Sparse Label Regimes (NeurIPS 2022 · similarity 0.90)
- Rethinking and Reweighting the Univariate Losses for Multi-Label Ranking: Consistency and Generalization (NeurIPS 2021 · similarity 0.90)
- Forming Auxiliary High-confident Instance-level Loss to Promote Learning from Label Proportions (CVPR 2025 · similarity 0.86)
- Consistent algorithms for multi-label classification with macro-at-$k$ metrics (ICLR 2024 · similarity 0.86)
- Semi-supervised Multi-label Learning with Balanced Binary Angular Margin Loss (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.