Research map / Vision-language and multimodal / CLIP and zero-shot
ExpAlign: Expectation-Guided Vision–Language Alignment for Open-Vocabulary Grounding
Junyi Hu, Tian Bai, Fengyi Wu, Wenyan Li, Zhenming Peng, Yi ZHANG
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 token-level alignment (EAH) instead of mean pooling (ablation) | ExpAlign · LVIS minival | AP: 31.9 -> 37.1 | Improved |
| use token-level alignment (EAH) instead of EOT pooled token (ablation) | ExpAlign · LVIS minival | AP: 34.4 -> 37.1 | Improved |
| use token-level alignment (EAH) instead of mean pooling (ablation) | ExpAlign · LVIS minival | rare AP: 27.3 -> 36.2 | Improved |
| add semantic contrastive loss weight 0.5 (ablation) | ExpAlign · LVIS minival | AP: 35.0 -> 37.0 | Improved |
| add geometry-aware consistency (GACO) weight 0.5 to semantic loss 0.5 (ablation) | ExpAlign · LVIS minival | AP: 37.0 -> 37.8 | Improved |
| raise semantic loss weight to 1.0 with geo weight 0.5 (ablation) | ExpAlign · LVIS minival | AP: 37.8 -> 36.2 | Hurt |
| increase geo weight from 1.0 to 1.5 with sem 0.5 (ablation) | ExpAlign · LVIS minival | AP: 37.2 -> 36.7 | Hurt |
| use frozen DINOv3 backbone instead of YOLOv8 (ablation) | ExpAlign · LVIS minival | AP: 35.6 -> 37.2 | Improved |
| freeze the DINOv3 backbone (ablation) | ExpAlign · LVIS minival | AP: N/A -> 37.2 | Improved |
| add RefCOCO data to pre-training | ExpAlign · RefCOCO val | top-1 accuracy: 42.2 -> 51.6 | Improved |
| add RefCOCO data to pre-training | ExpAlign · LVIS val | AP: 30.3 -> 29.5 | Hurt |
| add RefCOCO data to pre-training | ExpAlign · LVIS minival | AP: 37.2 -> 37.1 | No clear effect |
Closest papers by meaning
- Fine-Grained Verifiers: Preference Modeling as Next-token Prediction in Vision-Language Alignment (ICLR 2025 · similarity 0.88)
- DINOv2 Meets Text: A Unified Framework for Image- and Pixel-Level Vision-Language Alignment (CVPR 2025 · similarity 0.87)
- Text as Partial Constraint: Core–Residual Alignment for Robust Vision–Language Learning (NeurIPS 2026 · similarity 0.87)
- Analyzing Fine-Grained Alignment and Enhancing Vision Understanding in Multimodal Language Models (NeurIPS 2025 · similarity 0.87)
- ExGra-Med: Extended Context Graph Alignment for Medical Vision-Language Models (NeurIPS 2025 · similarity 0.87)
- DH-Set: Improving Vision-Language Alignment with Diverse and Hybrid Set-Embeddings Learning (CVPR 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.
