Research map / Label-efficient and robust learning / Recognition and long-tail
Maximizing Subset Accuracy with Recurrent Neural Networks in 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 |
|---|---|---|---|
| replace full-label RNN chain (RNNb) with positive-label-only RNN (RNNm) (ablation) | RNN for multi-label classification · Reuters-21578 | ACC: 0.6757 -> 0.7744 | Helped |
| use encoder-decoder with attention instead of fixed-input RNN | RNNm · Reuters-21578 | ACC: 0.7744 -> 0.8281 | Helped |
| use encoder-decoder with attention instead of fixed-input RNN | RNNm · BioASQ | ebF1: 0.4210 -> 0.5585 | Helped |
| use encoder-decoder with attention instead of fixed-input RNN | RNNm · RCV1-v2 | ACC: 0.6218 -> 0.6798 | Helped |
| reverse label ordering (rare-to-frequent or reverse topological) instead of frequent-first (ablation) | RNNm and EncDec · BioASQ | ebF1: EncDec 0.5294 (f2r) -> 0.5531 (r2f); RNNm 0.3917 -> 0.4188 | Helped |
| change label ordering (f2r, r2f, topological) (ablation) | RNNm and EncDec · RCV1-v2 | ACC | No clear effect |
| change label ordering on EncDec (ablation) | EncDec · Reuters-21578 | ACC: 0.8281 (f2r) vs 0.8261 (r2f) | No clear effect |
| use label ordering by frequency or topology vs none on label-order-sensitive methods (ablation) | PCC(NN) · Reuters-21578 | ACC: 0.7844 vs 0.7864 (f2r vs r2f) | No clear effect |
| switch from binary-relevance-style full-label RNN to positive-label RNN with EncDec on large label space | RNN-based MLC · BioASQ | ACC: 0.0001 (RNNm) vs 0.0007 (EncDec) | Mixed |
Closest papers by meaning
- Learning Context-dependent Label Permutations for Multi-label Classification (ICML 2019 · similarity 0.91)
- Recurrent Bayesian Classifier Chains for Exact Multi-Label Classification (NeurIPS 2021 · similarity 0.91)
- Orderless Recurrent Models for Multi-Label Classification (CVPR 2020 · similarity 0.90)
- Enhancing Multi-Label Classification via Dynamic Label-Order Learning (AAAI 2024 · similarity 0.88)
- Rebalancing Multi-Label Class-Incremental Learning (AAAI 2025 · similarity 0.86)
- Addressing Multi-Label Learning with Partial Labels: From Sample Selection to Label Selection (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.