Research map / Vision-language and multimodal / CLIP and zero-shot
Link the Head to the "Beak": Zero Shot Learning From Noisy Text Description at Part Precision
Check 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 |
|---|---|---|---|
| use part-based classifiers predicted from text with detected bird parts instead of the compared whole-image or text-based methods | zero-shot bird recognition · CUB2011, SCS (easy) | Top-1 accuracy (%): 33.5 -> 37.2 | Helped |
| use part-based classifiers predicted from text instead of the compared baselines | zero-shot bird recognition · CUB2011, SCE (hard) | Top-1 accuracy (%): 8.6 -> 9.7 | Helped |
| use part-based classifiers predicted from text instead of the compared baselines | zero-shot bird recognition · NABirds, SCS (easy) | Top-1 accuracy (%): 24.5 -> 30.3 | Helped |
| use part-based classifiers predicted from text instead of the compared baselines | zero-shot bird recognition · NABirds, SCE (hard) | Top-1 accuracy (%): 6.8 -> 8.1 | Helped |
| use ground-truth part annotations at test time instead of detected parts (ablation) | Ours-DET · CUB2011, SCS | Accuracy: 37.2 -> 43.6 | Helped |
| use detected parts at test time instead of ground-truth part annotations (ablation) | Ours-ATN · CUB2011, SCS | Accuracy: 43.6 -> 37.2 | Hurt |
Closest papers by meaning
- Improved Zero-Shot Classification by Adapting VLMs with Text Descriptions (CVPR 2024 · similarity 0.89)
- Learning Deep Representations of Fine-Grained Visual Descriptions (CVPR 2016 · similarity 0.88)
- Follow-Up Differential Descriptions: Language Models Resolve Ambiguities for Image Classification (ICLR 2024 · similarity 0.86)
- Interpreting and Analysing CLIP's Zero-Shot Image Classification via Mutual Knowledge (NeurIPS 2024 · similarity 0.85)
- A Simple Framework for Open-Vocabulary Zero-Shot Segmentation (ICLR 2025 · similarity 0.85)
- S3A: Towards Realistic Zero-Shot Classification via Self Structural Semantic Alignment (AAAI 2024 · 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.