Research map / Label-efficient and robust learning / Recognition and long-tail
Multi-Label Classification via Adaptive Resonance Theory-Based Clustering
Naoki Masuyama, Yusuke Nojima, Chu Kiong Loo, Hisao Ishibuchi
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 MLCA instead of MCIC for continual multi-label learning | MLCA · Synthetic 2D, 7 sequential distributions, after Dist. #7 | Exact Match: 0.334 -> 0.930 | Improved |
| use MLCA instead of MCIC with all data given at once | MLCA · Synthetic 2D stationary | Exact Match: 0.342 -> 0.972 | Improved |
| use MLCA instead of MCIC, 3 sequential distributions | MLCA · Synthetic 2D, 3 sequential distributions, after Dist. #3 | Exact Match: 0.392 -> 0.909 | Improved |
| use MLCA instead of MCIC, Hamming loss, 3 sequential distributions | MLCA · Synthetic 2D, 3 sequential distributions, after Dist. #3 | Hamming Loss: 0.212 -> 0.031 | Improved |
| increase similarity threshold V from 0.01 to 0.30 in MLCA (ablation) | MLCA · Genbase (nominal) | Average number of nodes: 192.0 -> 1.7 | Improved |
| use MLCA-C (CIM variant) instead of MLCA at large V (ablation) | MLCA-C · Genbase (nominal) | Average number of nodes: 1.7 -> 130.5 | Improved |
| raise V from 0.01 to 0.25 in MLCA-C (ablation) | MLCA-C · Emotions | Average number of nodes: 533.7 -> 244.6 | Improved |
| scale data to [0,1] for mlODM and GLOCAL (ablation) | mlODM, GLOCAL · 16 real-world multi-label datasets | rank (Nemenyi critical difference) | Improved |
| use batch learners (ML-kNN, mlODM, GLOCAL) instead of streaming learners (MCIC, MuENL, MLSA-kNN) (ablation) | comparison algorithms · 16 real-world multi-label datasets | rank (Nemenyi critical difference) | Improved |
| vary similarity threshold V (0.01 vs 0.25) in MLCA-C on nominal data (ablation) | MLCA-C · categorical datasets | rank (Nemenyi critical difference) | No clear effect |
| raise similarity threshold V for MLCA and MLCA-I on nominal data (ablation) | MLCA, MLCA-I · categorical datasets | rank (Nemenyi critical difference) | Hurt |
Closest papers by meaning
- Conditional Bernoulli Mixtures for Multi-label Classification (ICML 2016 · similarity 0.84)
- Enhancing Multi-Label Classification via Dynamic Label-Order Learning (AAAI 2024 · similarity 0.82)
- T-MDML: Triplet-based Multiple Distance Metric Learning for Multi-Instance Multi-Label Classification with Label Correlation (AAAI 2025 · similarity 0.81)
- Consistent algorithms for multi-label classification with macro-at-$k$ metrics (ICLR 2024 · similarity 0.81)
- Tensorized Label Learning Based Fast Fuzzy Clustering (AAAI 2025 · similarity 0.81)
- Decouple then Classify: A Dynamic Multi-view Labeling Strategy with Shared and Specific Information (ICML 2024 · similarity 0.81)
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.
