Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Generative Multi-Label Zero-Shot Learning
Akshita Gupta, Sanath Narayan, Salman Hameed Khan, Fahad Shahbaz Khan, Ling Shao, Joost van de Weijer
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 cross-level feature fusion (CLF) instead of ALF (ablation) | f-VAEGAN · NUS-WIDE (925 seen / 81 unseen) ZSL | mAP: 22.9 -> 25.7 | Improved |
| use CLF instead of FLF (ablation) | f-VAEGAN · NUS-WIDE (925 seen / 81 unseen) ZSL | mAP: 23.3 -> 25.7 | Improved |
| use CLF instead of ALF (ablation) | f-CLSWGAN · NUS-WIDE (925 seen / 81 unseen) ZSL | mAP: 22.1 -> 23.7 | Improved |
| use CLF instead of ALF (ablation) | f-VAEGAN · NUS-WIDE (925 seen / 81 unseen) GZSL | F1 (K=3): 17.7 -> 18.9 | Improved |
| use CLF instead of AVG fusion (ablation) | f-VAEGAN · NUS-WIDE (925 seen / 81 unseen) ZSL | mAP: 23.2 -> 25.7 | Improved |
| use CLF instead of CONCAT fusion (ablation) | f-VAEGAN · NUS-WIDE (925 seen / 81 unseen) ZSL | mAP: 23.6 -> 25.7 | Improved |
| use nearest-unseen class replacement instead of random combinations (ablation) | f-VAEGAN · NUS-WIDE (925 seen / 81 unseen) ZSL | mAP: 21.0 -> 25.7 | Improved |
| use nearest-unseen class replacement instead of random combinations (ablation) | f-VAEGAN · NUS-WIDE (925 seen / 81 unseen) GZSL | mAP: 6.1 -> 8.9 | Improved |
| use CLF with f-VAEGAN instead of f-CLSWGAN (ablation) | CLF · NUS-WIDE (925 seen / 81 unseen) ZSL | mAP: 23.7 -> 25.7 | Improved |
| swap VGG backbone for DINO ResNet-50 features (ablation) | CLF · NUS-WIDE (925 seen / 81 unseen) ZSL | mAP: 25.7 -> 27.1 | Improved |
| swap VGG backbone for DINO ResNet-50 features (ablation) | CLF · NUS-WIDE (925 seen / 81 unseen) GZSL | mAP: 8.9 -> 10.4 | Improved |
| swap VGG backbone for Places365 ResNet-50 features (ablation) | CLF · NUS-WIDE (925 seen / 81 unseen) ZSL | mAP: 25.7 -> 26.0 | Improved |
Closest papers by meaning
- Generative Dual Adversarial Network for Generalized Zero-Shot Learning (CVPR 2019 · similarity 0.91)
- Feature Generating Networks for Zero-Shot Learning (CVPR 2018 · similarity 0.90)
- Leveraging the Invariant Side of Generative Zero-Shot Learning (CVPR 2019 · similarity 0.89)
- Adaptive and Generative Zero-Shot Learning (ICLR 2021 · similarity 0.89)
- Semi-Supervised Multi-View Multi-Label Learning with View-Specific Transformer and Enhanced Pseudo-Label (AAAI 2025 · similarity 0.87)
- Epsilon: Exploring Comprehensive Visual-Semantic Projection for Multi-Label Zero-Shot Learning (AAAI 2025 · similarity 0.87)
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.
