Research map / Label-efficient and robust learning / Recognition and long-tail
The Missing Link: Finding Label Relations across Datasets
Jasper Uijlings, Thomas Mensink, Vittorio Ferrari
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 visual embeddings instead of pixel probabilities (ablation) | visual relation discovery with set theory · ADE20k-COCO relation type | accuracy: 47 -> 51 | Improved |
| use score asymmetry instead of set theory for relation type (ablation) | Visual Embeddings · ADE20k-COCO relation type | accuracy: 51 -> 61 | Improved |
| use score asymmetry instead of set theory for relation type (ablation) | Pixel Predictions · ADE20k-COCO relation type | accuracy: 47 -> 60 | Improved |
| combine WordNet taxonomy with visual embeddings asymmetry (ablation) | Visual Embeddings asymmetry · average over dataset pairs | accuracy: 56 -> 57 | Improved |
| combine WordNet taxonomy with visual embeddings asymmetry (ablation) | Visual Embeddings asymmetry · BDD-COCO relation type | accuracy: 51 -> 53 | Improved |
| use vision instead of language only | Visual Embeddings asymmetry vs WordNet taxonomy · average over dataset pairs | accuracy: 46 -> 56 | Improved |
| use WordNet instead of Word2Vec | language-only relation discovery · average over dataset pairs | accuracy: 38 -> 46 | Improved |
| use pixel predictions with asymmetry vs set theory on ADE20k-BDD (ablation) | Pixel Predictions · ADE20k-BDD relation type | accuracy: 56 -> 54 | No clear effect |
| add stronger label link to predict transfer gain | COCO to ADE20k fine-tuning on 1000 images · ADE20k val per-class IoU | transfer gain | Improved |
Closest papers by meaning
- Weakly-Supervised Learning of Visual Relations (ICCV 2017 · similarity 0.88)
- Large-Scale Visual Relationship Understanding (AAAI 2019 · similarity 0.87)
- Visual Relationship Detection With Internal and External Linguistic Knowledge Distillation (ICCV 2017 · similarity 0.87)
- Navigating Real-World Partial Label Learning: Unveiling Fine-Grained Images with Attributes (AAAI 2024 · similarity 0.86)
- Epsilon: Exploring Comprehensive Visual-Semantic Projection for Multi-Label Zero-Shot Learning (AAAI 2025 · similarity 0.85)
- Incomplete Multi-View Multi-label Learning via Disentangled Representation and Label Semantic Embedding (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.
