Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Decoupling Primitive with Experts: Dynamic Feature Alignment for Compositional Zero-Shot Learning
Xiao Zhang, Haodong Jing, Yongqiang Ma, Nanning Zheng
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 |
|---|---|---|---|
| Add domain-expert adaption (ablation) | Baseline CLIP · MIT-States | AUC: 20.2 -> 23.0 | Improved |
| Add domain-expert adaption (ablation) | Baseline CLIP · C-GQA | AUC: 10.4 -> 17.2 | Improved |
| Add semantic variant alignment (ablation) | Baseline CLIP · C-GQA | AUC: 10.4 -> 12.1 | Improved |
| Add semantic variant alignment (ablation) | Baseline CLIP · MIT-States | AUC: 20.2 -> 22.2 | Improved |
| Combine adaption and alignment (ablation) | EVA · C-GQA | AUC: 17.2 -> 18.8 | Improved |
| Add one shared expert (1+8 vs 0+8) (ablation) | MoE adapter · C-GQA | AUC: 18.0 -> 18.8 | Improved |
| Add more shared experts (2+8) (ablation) | MoE adapter · C-GQA | AUC: 18.8 -> 18.4 | Hurt |
| Use 10 routed experts instead of 8 (ablation) | MoE adapter · C-GQA | AUC: 18.8 -> 18.0 | Hurt |
| Activate K=2 routed experts instead of 0 (ablation) | MoE adapter · C-GQA | AUC: 16.7 -> 18.8 | Improved |
| Increase activated experts from 2 to 8 (ablation) | MoE adapter · C-GQA | AUC: 18.8 -> 17.5 | Hurt |
| Add semantic isolation of state and object experts (ablation) | EVA · C-GQA | AUC: 18.8 -> 18.5 | Hurt |
| Add expert load balancing loss (ablation) | EVA · C-GQA | AUC: 18.8 -> 18.0 | Hurt |
Closest papers by meaning
- ProCC: Progressive Cross-Primitive Compatibility for Open-World Compositional Zero-Shot Learning (AAAI 2024 · similarity 0.91)
- Retrieval-Augmented Primitive Representations for Compositional Zero-Shot Learning (AAAI 2024 · similarity 0.91)
- A Dynamic Learning Method towards Realistic Compositional Zero-Shot Learning (AAAI 2024 · similarity 0.90)
- Compositional Few-Shot Class-Incremental Learning (ICML 2024 · similarity 0.89)
- Beyond Seen Primitive Concepts and Attribute-Object Compositional Learning (CVPR 2024 · similarity 0.89)
- Revealing the Proximate Long-Tail Distribution in Compositional Zero-Shot Learning (AAAI 2024 · similarity 0.89)
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.
