Research map / Label-efficient and robust learning / Recognition and long-tail
Does Head Label Help for Long-Tailed Multi-Label Text 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 |
|---|---|---|---|
| Transfer meta-knowledge learned from head labels to tail-label classifiers | HTTN · EUR-Lex, l_tail=768 | F1-score: 53.72 vs 53.28 | Helped |
| Ensemble multiple sampled tail classifiers (ablation) | HTTN · AAPD, l_tail=18 | F1-score: 67.71 -> 68.84 | Helped |
| Ensemble multiple sampled tail classifiers (ablation) | HTTN · AAPD, l_tail=9 | F1-score: 68.14 -> 69.25 | Helped |
| Ensemble multiple sampled tail classifiers (ablation) | HTTN · RCV1, l_tail=28 | F1-score: 76.09 -> 76.92 | Helped |
| Ensemble multiple sampled tail classifiers (ablation) | HTTN · RCV1, l_tail=14 | F1-score: 76.86 -> 77.72 | Helped |
| Add tail-label attention using head-label prototypes (ablation) | HTTN without attention or fine-tuning (H-F-A) · AAPD and RCV1 | F1-score | Helped |
| Add fine-tuning (ablation) | HTTN without fine-tuning (H-F) · AAPD and RCV1 | F1-score | Helped |
| Increase transfer-learner prototype sampling frequency from 5 to 20 (ablation) | HTTN · AAPD and RCV1 | F1-score | Helped |
| Use fewer tail labels, leaving more head labels for learning the transfer mapping | HTTN · AAPD and RCV1 | F1-score: 68.14 -> 68.14; 76.09 -> 76.86 | Helped |
| Use multiple sampled tail classifiers rather than a single sampled classifier (ablation) | HTTN · AAPD and RCV1 | F1-score | Helped |
Closest papers by meaning
- MLC-NC: Long-Tailed Multi-Label Image Classification Through the Lens of Neural Collapse (AAAI 2025 · similarity 0.86)
- Long-Tailed Partial Label Learning by Head Classifier and Tail Classifier Cooperation (AAAI 2024 · similarity 0.85)
- Keep It on a Leash: Controllable Pseudo-label Generation Towards Realistic Long-Tailed Semi-Supervised Learning (NeurIPS 2025 · similarity 0.83)
- Self-Paced Unified Representation Learning for Hierarchical Multi-Label Classification (AAAI 2024 · similarity 0.83)
- Multi-Label Node Classification with Label Influence Propagation (ICLR 2025 · similarity 0.83)
- Addressing Multi-Label Learning with Partial Labels: From Sample Selection to Label Selection (AAAI 2025 · similarity 0.83)
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.