Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Proxy Zero-Shot Hashing with Multimodal Fusion via Stable Diffusion
Hui Zhang, Weikang Gao, Tao Yang, Yuan Cao
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 unseen-class synthetic images from Stable Diffusion (ablation) | PZSH · AWA2 128 bits | mAP@all: 0.6183 -> 0.6477 | Improved |
| add unseen-class synthetic images (ablation) | PZSH · CUB 64 bits | mAP@all: 0.1915 -> 0.2651 | Improved |
| add multimodal fusion generator and feature alignment module (ablation) | PZSH · AWA2 64 bits | mAP@all: 0.5561 -> 0.6297 | Improved |
| add multimodal fusion generator and feature alignment module (ablation) | PZSH · CUB 64 bits | mAP@all: 0.1599 -> 0.2651 | Improved |
| use BLIP instead of CLIP encoder (ablation) | PZSH · AWA2 64 bits | mAP@all: 0.5942 -> 0.6297 | Improved |
| use BLIP instead of CLIP encoder (ablation) | PZSH · CUB 128 bits | mAP@all: 0.2626 -> 0.2775 | Improved |
| add pair/proxy loss (ablation) | PZSH · AWA2 64 bits | mAP@all: 0.5618 -> 0.6297 | Improved |
| add pair/proxy loss (ablation) | PZSH · CUB 64 bits | mAP@all: 0.2375 -> 0.2651 | Improved |
| add quantization loss (ablation) | PZSH · AWA2 64 bits | mAP@all: 0.6206 -> 0.6297 | Improved |
| add quantization loss (ablation) | PZSH · CUB 128 bits | mAP@all: 0.2613 -> 0.2775 | Improved |
| PZSH vs PIXEL | PZSH · CUB 64 bits | mAP@all: 0.1994 -> 0.2651 | Improved |
| vary alpha and beta loss weights (ablation) | PZSH · AWA2 64 bits | mAP@all | No clear effect |
Closest papers by meaning
- Zero-Shot Sketch-Image Hashing (CVPR 2018 · similarity 0.88)
- Self-guided Semantic Inspection for Zero-Shot Composed Image Retrieval (CVPR 2026 · similarity 0.88)
- ZeroDiff: Solidified Visual-semantic Correlation in Zero-Shot Learning (ICLR 2025 · similarity 0.87)
- STiTch: Semantic Transition and Transportation in Collaboration for Training-Free Zero-Shot Composed Image Retrieval (CVPR 2026 · similarity 0.87)
- Context-I2W: Mapping Images to Context-Dependent Words for Accurate Zero-Shot Composed Image Retrieval (AAAI 2024 · similarity 0.87)
- TOMCAT: Test-time Comprehensive Knowledge Accumulation for Compositional Zero-Shot Learning (NeurIPS 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.
