Research map / Label-efficient and robust learning / Recognition and long-tail
Orderless Recurrent Models for Multi-Label Classification
Check 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 |
|---|---|---|---|
| dynamically align labels to predicted labels instead of using a fixed frequent-first order | CNN-RNN multi-label classifier · MS-COCO | per-class F1: 62.63 -> 64.34 | Helped |
| add attention to PLA (ablation) | CNN-RNN with PLA · MS-COCO | per-class F1: 64.34 -> 65.81 | Helped |
| use predicted-label alignment instead of minimal-loss alignment (ablation) | CNN-RNN multi-label classifier · MS-COCO | per-class F1: 64.13 -> 64.34 | Helped |
| replace fixed frequent-first ordering with predicted-label alignment | CNN-RNN multi-label classifier · MS-COCO validation | ratio of duplicates: 23.84% -> 0.04% | Helped |
| use dynamic predicted-label alignment instead of a fixed order | CNN-RNN multi-label classifier · MS-COCO | overall F1: 68.13 -> 69.66 | Helped |
| use PLA rather than the strongest reported baseline | multi-label classifier · NUS-WIDE | overall F1: 73.70 -> 72.37 | Hurt |
| use PLA rather than the strongest reported baseline | multi-label classifier · NUS-WIDE | per-class F1: 58.70 -> 56.23 | Hurt |
| use PLA rather than ACfs | multi-label classifier · WIDER Attribute | overall F1: 82.40 -> 83.13 | Helped |
| use PLA rather than ACfs | multi-label classifier · PA-100K | accuracy: 79.44 -> 79.83 | Helped |
Closest papers by meaning
- Order-Free RNN With Visual Attention for Multi-Label Classification (AAAI 2018 · similarity 0.92)
- Maximizing Subset Accuracy with Recurrent Neural Networks in Multi-label Classification (NeurIPS 2017 · similarity 0.90)
- Enhancing Multi-Label Classification via Dynamic Label-Order Learning (AAAI 2024 · similarity 0.90)
- CNN-RNN: A Unified Framework for Multi-Label Image Classification (CVPR 2016 · similarity 0.89)
- Order-Free Learning Alleviating Exposure Bias in Multi-Label Classification (AAAI 2020 · similarity 0.88)
- Neural Collapse in Multi-label Learning with Pick-all-label Loss (ICML 2024 · similarity 0.84)
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.