Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
TapNet: Neural Network Augmented with Task-Adaptive Projection for Few-Shot Learning
Read the paperCheck 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 task-adaptive null-space projection and learned class references | TapNet · 20-way Omniglot, 1-shot | accuracy: 97.64\% -> 98.07\% | Helped |
| Add task-adaptive null-space projection and learned class references | TapNet · 20-way Omniglot, 5-shot | accuracy: 99.36\% -> 99.49\% | Helped |
| Add task-adaptive null-space projection and learned class references | TapNet · 5-way miniImageNet, 1-shot | accuracy: 58.5 \pm 0.3\% -> 61.65 \pm 0.15\% | Helped |
| Add task-adaptive null-space projection and learned class references | TapNet · 5-way miniImageNet, 5-shot | accuracy: 76.7 \pm 0.3\% vs 76.36 \pm 0.10\% | No clear effect |
| Add task-adaptive null-space projection and learned class references | TapNet · 5-way tieredImageNet, 1-shot | accuracy: 59.91 \pm 0.94\% -> 63.08 \pm 0.15\% | Helped |
| Add task-adaptive null-space projection and learned class references | TapNet · 5-way tieredImageNet, 5-shot | accuracy: 73.30 \pm 0.75\% -> 80.26 \pm 0.12\% | Helped |
| Reduce projection dimension from full dimension toward 50 (ablation) | TapNet · 5-way miniImageNet, 5-shot | accuracy | Helped |
| Vary projection dimension around 200 (ablation) | TapNet · 5-way miniImageNet, 5-shot | accuracy | Mixed |
Closest papers by meaning
- LGM-Net: Learning to Generate Matching Networks for Few-Shot Learning (ICML 2019 · similarity 0.87)
- Task Adaptive Parameter Sharing for Multi-Task Learning (CVPR 2022 · similarity 0.87)
- Cross-Domain Few-Shot Learning With Task-Specific Adapters (CVPR 2022 · similarity 0.87)
- XtarNet: Learning to Extract Task-Adaptive Representation for Incremental Few-Shot Learning (ICML 2020 · similarity 0.87)
- Finding Task-Relevant Features for Few-Shot Learning by Category Traversal (CVPR 2019 · similarity 0.87)
- Looking Wider for Better Adaptive Representation in Few-Shot Learning (AAAI 2021 · similarity 0.86)
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.