Research map / Vision-language and multimodal / CLIP and zero-shot
Test-Time Retrieval-Augmented Adaptation for Vision-Language Models
Xinqi Fan, Xueli Chen, Luoxiao Yang, Chuin Hong Yap, Rizwan Qureshi, Qi Dou, Moi Hoon Yap, Mubarak Shah
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 SMGD module to the TDA one-shot baseline (ablation) | TDA baseline · Cross-domain benchmark, ViT-B/16 | top-1 accuracy: 67.04 -> 68.46 | Improved |
| add SMGD module to the TDA one-shot baseline (ablation) | TDA baseline · Cross-domain benchmark, ResNet-50 | top-1 accuracy: 59.82 -> 61.03 | Improved |
| add MRA module to the TDA one-shot baseline (ablation) | TDA baseline · Cross-domain benchmark, ViT-B/16 | top-1 accuracy: 67.04 -> 67.49 | Improved |
| add MRA module to the TDA one-shot baseline (ablation) | TDA baseline · Cross-domain benchmark, ResNet-50 | top-1 accuracy: 59.82 -> 60.02 | Improved |
| combine SMGD and MRA over the baseline (ablation) | TDA baseline · Cross-domain benchmark, ViT-B/16 | top-1 accuracy: 67.04 -> 68.76 | Improved |
| combine SMGD and MRA over the baseline (ablation) | TDA baseline · Cross-domain benchmark, ResNet-50 | top-1 accuracy: 59.82 -> 61.25 | Improved |
| use a nonzero update coefficient eta (0.65) instead of eta = 0 (ablation) | TT-RAA · UCF101 | top-1 accuracy: eta = 0 gives poor adaptation; best eta = 0.65 | Improved |
| raise update coefficient eta beyond 0.65 (ablation) | TT-RAA · UCF101 | top-1 accuracy: performance deteriorates beyond 0.65 | Hurt |
| build the Gaussian database from the target training set (TT-RAA-A) (ablation) | TT-RAA · Cross-domain benchmark (Fig. 5b) | top-1 accuracy | Improved |
| use TT-RAA instead of TDA | TDA · EuroSAT | top-1 accuracy: 58.00 -> 66.12 | Improved |
| use TT-RAA instead of TDA | TDA · UCF101 | top-1 accuracy: 70.66 -> 71.29 | Improved |
Closest papers by meaning
- The Illusion of Progress? A Critical Look at Test-Time Adaptation for Vision-Language Models (NeurIPS 2025 · similarity 0.93)
- Frustratingly Easy Test-Time Adaptation of Vision-Language Models (NeurIPS 2024 · similarity 0.93)
- BoostAdapter: Improving Vision-Language Test-Time Adaptation via Regional Bootstrapping (NeurIPS 2024 · similarity 0.92)
- Efficient Test-time Adaptation through Candidate Verification and Divergence Shifts (NeurIPS 2026 · similarity 0.91)
- Understanding Retrieval-Augmented Task Adaptation for Vision-Language Models (ICML 2024 · similarity 0.91)
- Realistic Test-Time Adaptation of Vision-Language Models (CVPR 2025 · similarity 0.91)
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.
