Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Transductive Zero-Shot Learning with Visual Structure Constraint
Read the paperCheck 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 |
|---|---|---|---|
| Add visual structure alignment to source-domain VCL using Wasserstein distance (ablation) | VCL · AwA1, AwA2, CUB, SUN72, and SUN10 SS | MCA (%): 82.0 -> 96.2; 82.5 -> 96.7; 60.1 -> 74.2; 63.8 -> 67.8; 89.6 -> 92.2 | Helped |
| Add Chamfer-distance visual structure alignment (ablation) | VCL · AwA1, AwA2, CUB, SUN72, and SUN10 SS | MCA (%): 82.0 -> 94.3; 82.5 -> 93.9; 60.1 -> 74.2; 63.8 -> 64.5; 89.6 -> 90.5 | Helped |
| Add one-to-one bipartite matching alignment (ablation) | VCL · AwA1, AwA2, CUB, SUN72, and SUN10 SS | MCA (%): 82.0 -> 95.9; 82.5 -> 96.8; 60.1 -> 73.6; 63.8 -> 66.2; 89.6 -> 91.7 | Helped |
| Use soft Wasserstein matching instead of hard one-to-one bipartite matching (ablation) | Visual structure constraint · AwA2 SS | MCA (%): 96.8 -> 96.7 | No clear effect |
| Use soft Wasserstein matching instead of hard one-to-one bipartite matching (ablation) | Visual structure constraint · CUB SS | MCA (%): 73.6 -> 74.2 | Helped |
| Use soft Wasserstein matching instead of hard one-to-one bipartite matching (ablation) | Visual structure constraint · SUN10 SS | MCA (%): 91.7 -> 92.2 | Helped |
| Filter unrelated test images before clustering and structure alignment (ablation) | WDVSc · AwA2 SS+noise and PS+noise | MCA (%): 81.3 -> 92.4; 60.8 -> 78.3 | Helped |
| Apply visual structure constraints without filtering unrelated images (ablation) | VCL · AwA2 SS+noise | MCA (%): 82.5 -> 79.7; 82.5 -> 78.3; 82.5 -> 81.3 | Hurt |
| Increase the assumed cluster number K (ablation) | BMVSc and WDVSc · AwA2 | MCA (%): 61.5 -> 81.7; 61.5 -> 87.3 | Helped |
| Add visual structure constraint to VCL (ablation) | VCL · AwA2, CUB, and SUN72 PS | MCA (%): 61.5 -> 87.3; 59.6 -> 73.4; 59.4 -> 63.4 | Helped |
Closest papers by meaning
- Improving Generalized Zero-Shot Learning by Exploring the Diverse Semantics from External Class Names (CVPR 2024 · similarity 0.91)
- ZeroDiff: Solidified Visual-semantic Correlation in Zero-Shot Learning (ICLR 2025 · similarity 0.90)
- ZeroMamba: Exploring Visual State Space Model for Zero-Shot Learning (AAAI 2025 · similarity 0.89)
- Data-Free Generalized Zero-Shot Learning (AAAI 2024 · similarity 0.89)
- Progressive Semantic-Guided Vision Transformer for Zero-Shot Learning (CVPR 2024 · similarity 0.88)
- Epsilon: Exploring Comprehensive Visual-Semantic Projection for Multi-Label Zero-Shot Learning (AAAI 2025 · similarity 0.88)
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.