Research map / Label-efficient and robust learning / Recognition and long-tail
Learning from Weak-Label Data: A Deep Forest Expedition
Qian-Wei Wang, Liang Yang, Yu-Feng Li
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 LCForest (cascade forest with label complement) instead of WELL, ML-kNN, RF-PCT, DBPNN | LCForest · yeast, TMC, scene, medical weak-label data | Micro-F1: 0.742 vs 0.555 (medical 20%); 0.186 vs 0.262 (medical 50%, ML-kNN) | Mixed |
| LCForest vs baselines on TMC | LCForest · TMC | Micro-F1, Hamming Loss: 0.630 vs 0.482 (DBPNN, 20%) | Improved |
| LCForest vs WELL on Macro-F1 at low incomplete label ratio | LCForest · yeast and scene, I.L. 20% | Macro-F1: 0.407 vs 0.476 (yeast, WELL) | Hurt |
| increase incomplete label ratio from 20% to 50% | WELL · yeast | Micro-F1: 0.591 -> 0.015 | Hurt |
Closest papers by meaning
- Deep Generative Models for Weakly-Supervised Multi-Label Classification (ECCV 2018 · similarity 0.87)
- Addressing Imbalance in Multi-Label Classification Using Structured Hellinger Forests (AAAI 2017 · similarity 0.87)
- Learning From Semi-Supervised Weak-Label Data (AAAI 2018 · similarity 0.87)
- Learning from weak labelers as constraints (ICLR 2025 · similarity 0.86)
- Limited-Supervised Multi-Label Learning with Dependency Noise (AAAI 2024 · similarity 0.85)
- Forming Auxiliary High-confident Instance-level Loss to Promote Learning from Label Proportions (CVPR 2025 · similarity 0.85)
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.
