Research map / Label-efficient and robust learning / Recognition and long-tail
Learning Context-dependent Label Permutations for Multi-label 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 |
|---|---|---|---|
| hybrid AC + MbLP dynamic label permutation learning (CLP-RNN) vs static/fixed orderings | RNN multi-label classifier · Mediamill, Delicious, EURLex | Example F1 and ranking measures | Helped |
| adding AC policy-gradient term (alpha>0) on top of MbLP (ablation) | CLP-RNN · Delicious | Example F1: 34.43 -> 35.80 | Helped |
| adding AC policy-gradient term (alpha>0) on top of MbLP (ablation) | CLP-RNN · EURlex | Example F1: 51.04 -> 53.61 | Helped |
| adding AC policy-gradient term (alpha>0) on top of MbLP (ablation) | CLP-RNN · Mediamill | macro F1: 43.33 (rare2freq) vs 36.49 (α=0.6); α=0 gives 38.75 | Hurt |
| adding AC term (alpha>0) on top of MbLP (ablation) | CLP-RNN · Mediamill | nDCG@5: 78.92 -> 80.52 | Helped |
| dynamic model-based label permutation (MbLP) vs fixed random ordering (ablation) | RNN · Delicious | nDCG@5: 36.96 -> 58.80 | Helped |
| frequency-based rare-first ordering (rare2freq) vs frequent-first | RNN · Mediamill | ranking measures (nDCG@5): 79.22-level freq2rare 78.74 vs rare2freq 63.46 | Hurt |
| always-rnd random dynamic permutation vs static orderings | RNN · Mediamill | nDCG@5 | Mixed |
| hyperparameter sensitivity: alpha=0.9 with gamma>0.9 (ablation) | CLP-RNN · Delicious | Example F1 | Hurt |
| CLP-RNN training cost vs freq2rare | CLP-RNN · Mediamill, Delicious, EURLex | wall-clock seconds per epoch: Mediamill 10.42 -> 23.78 s | Hurt |
Closest papers by meaning
- Enhancing Multi-Label Classification via Dynamic Label-Order Learning (AAAI 2024 · similarity 0.91)
- Maximizing Subset Accuracy with Recurrent Neural Networks in Multi-label Classification (NeurIPS 2017 · similarity 0.91)
- Order-Free Learning Alleviating Exposure Bias in Multi-Label Classification (AAAI 2020 · similarity 0.88)
- Addressing Multi-Label Learning with Partial Labels: From Sample Selection to Label Selection (AAAI 2025 · similarity 0.86)
- Batch Selection for Multi-Label Classification Guided by Uncertainty and Dynamic Label Correlations (AAAI 2025 · similarity 0.85)
- Rebalancing Multi-Label Class-Incremental Learning (AAAI 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.