Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Towards Affordable Semantic Searching: Zero-Shot Retrieval via Dominant Attributes
Yang Long, Li Liu, Yuming Shen, Ling Shao
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 orthogonal constraint on the shared latent space (LIAD vs NoV) (ablation) | LIAD zero-shot retrieval · CUB I2A rank 10 | hit rate: 19.50 -> 36.20 | Improved |
| add orthogonal constraint on the shared latent space (LIAD vs NoV) (ablation) | LIAD zero-shot retrieval · SUN I2A rank 10 | hit rate: 49.2 -> 68.5 | Improved |
| add orthogonal constraint (LIAD vs NoV) (ablation) | LIAD zero-shot retrieval · SUN A2I rank 10 | hit rate: 36.8 -> 61.9 | Improved |
| use augmented attributes (class names added) instead of binary attributes (ablation) | LIAD conventional ZSL · CUB | accuracy: 43.95 -> 50.2 | Improved |
| use augmented attributes instead of binary attributes (ablation) | LIAD conventional ZSL · SUN | accuracy: 82.9 -> 85.8 | Improved |
| add seen-class candidates to gallery with augmented attributes vs normal attributes (ablation) | generalised ZSR · SUN and CUB | hit rate at rank 10 | Improved |
| increase number of dominant attributes beyond 20 (ablation) | LIAD · SUN and CUB | hit rate | No clear effect |
| LIAD vs ESZSL at rank 1 on SUN A2I | zero-shot retrieval · SUN A2I rank 1 | hit rate: 18.8 -> 18.7 | Mixed |
Closest papers by meaning
- Beyond Instance-Level Image Retrieval: Leveraging Captions to Learn a Global Visual Representation for Semantic Retrieval (CVPR 2017 · similarity 0.89)
- Automatic Discovery, Association Estimation and Learning of Semantic Attributes for a Thousand Categories (CVPR 2017 · similarity 0.89)
- Highlighting What Matters: Promptable Embeddings for Attribute-Focused Image Retrieval (NeurIPS 2025 · similarity 0.87)
- Instance-Level Composed Image Retrieval (NeurIPS 2025 · similarity 0.87)
- Knowledge-Enhanced Dual-stream Zero-shot Composed Image Retrieval (CVPR 2024 · similarity 0.87)
- Context-I2W: Mapping Images to Context-Dependent Words for Accurate Zero-Shot Composed Image Retrieval (AAAI 2024 · similarity 0.86)
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.
