Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Learning Dynamic Alignment via Meta-Filter for Few-Shot Learning
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 |
|---|---|---|---|
| add dynamic sampling (ablation) | Dynamic Meta-filter · miniImageNet 1-shot | accuracy: 66.73 -> 67.76; 1.03% | Helped |
| add dynamic sampling (ablation) | Dynamic Meta-filter · miniImageNet 5-shot | accuracy: 81.20 -> 82.71; 1.50% | Helped |
| remove the last pooling layer (ablation) | Res-12 backbone · miniImageNet 1-shot | accuracy: 67.12 -> 67.76; 0.64% | Helped |
| remove the last pooling layer (ablation) | Res-12 backbone · miniImageNet 5-shot | accuracy: 81.54 -> 82.71; 1.16% | Helped |
| use 1 group instead of the task-specific group setting (ablation) | Dynamic Meta-filter · miniImageNet 1-shot | accuracy: 66.48 | Hurt |
| increase groups beyond 64 (ablation) | Dynamic Meta-filter · miniImageNet 1-shot | accuracy: 67.76; 67.61; 67.51; 67.43 | Hurt |
| increase groups beyond 160 (ablation) | Dynamic Meta-filter · miniImageNet 5-shot | accuracy: 82.71; 82.71; 82.34 | Mixed |
| add one dynamic meta-filter alignment instead of no alignment (ablation) | Dynamic Alignment Network · miniImageNet 1-shot | accuracy: 64.20 -> 67.26; 1.28% | Helped |
| add one dynamic meta-filter alignment instead of no alignment (ablation) | Dynamic Alignment Network · miniImageNet 5-shot | accuracy: 80.24 -> 82.14; 0.77% | Helped |
| use three fixed recursive alignments instead of one (ablation) | Dynamic Alignment Network · miniImageNet 5-shot | accuracy: 82.14 -> 81.19; 0.95% | Hurt |
| use adaptive ODE alignment instead of fixed-depth recursive alignment (ablation) | Dynamic Alignment Network · miniImageNet 5-way 1-shot | accuracy: 67.76 | Helped |
| use adaptive ODE alignment instead of fixed-depth recursive alignment (ablation) | Dynamic Alignment Network · miniImageNet 5-way 5-shot | accuracy: 82.71 | Helped |
Closest papers by meaning
- Adaptive Cross-Modal Few-shot Learning (NeurIPS 2019 · similarity 0.89)
- Simple Semantic-Aided Few-Shot Learning (CVPR 2024 · similarity 0.88)
- Pushing the Limit of Fine-Tuning for Few-Shot Learning: Where Feature Reusing Meets Cross-Scale Attention (AAAI 2024 · similarity 0.88)
- Making Large Vision Language Models to Be Good Few-Shot Learners (AAAI 2025 · similarity 0.87)
- VT-FSL: Bridging Vision and Text with LLMs for Few-Shot Learning (NeurIPS 2025 · similarity 0.87)
- FeatWalk: Enhancing Few-Shot Classification through Local View Leveraging (AAAI 2024 · 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.