Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Spot and Learn: A Maximum-Entropy Patch Sampler for Few-Shot Image Classification
Wen-Hsuan Chu, Yu-Jhe Li, Jing-Cheng Chang, Yu-Chiang Frank Wang
Check 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 maximum-entropy patch sampling with Ours-FC instead of the fully connected baseline | Conv-4 few-shot classifier · Omniglot 5-Way 1-Shot | accuracy: 5.5% | Improved |
| Use maximum-entropy patch sampling with Ours-CS instead of the cosine-similarity baseline | Conv-4 few-shot classifier · Omniglot 5-Way 1-Shot | accuracy: 4.2% | Improved |
| Use maximum-entropy patch sampling with Ours-FC instead of the fully connected baseline | Conv-4 few-shot classifier · Omniglot 5-Way 5-Shot | accuracy: 0.5% | Improved |
| Use maximum-entropy patch sampling with Ours-CS instead of the cosine-similarity baseline | Conv-4 few-shot classifier · Omniglot 5-Way 5-Shot | accuracy: 0.3% | Improved |
| Use maximum-entropy patch sampling with Ours-FC instead of the fully connected baseline | Conv-4 few-shot classifier · miniImagenet 5-Way 1-Shot | accuracy: 5.1% | Improved |
| Use maximum-entropy patch sampling with Ours-CS instead of the cosine-similarity baseline | Conv-4 few-shot classifier · miniImagenet 5-Way 1-Shot | accuracy: 4.2% | Improved |
| Use maximum-entropy patch sampling with Ours-FC instead of the fully connected baseline | Conv-4 few-shot classifier · miniImagenet 5-Way 5-Shot | accuracy: 4.9% | Improved |
| Use maximum-entropy patch sampling with Ours-CS instead of the cosine-similarity baseline | Conv-4 few-shot classifier · miniImagenet 5-Way 5-Shot | accuracy: 3.8% | Improved |
| Add cosine-similarity classification instead of a fully connected classifier | Baseline few-shot models · Omniglot 5-Way | accuracy | Improved |
| Add cosine-similarity classification instead of a fully connected classifier | Baseline few-shot models · miniImagenet 5-Way | accuracy | Improved |
| Add soft voting over predictions, increasing votes from N=1 to N=3 (ablation) | Ours on miniImagenet · miniImagenet 5-Way 5-Shot | accuracy: 3% to 4% | Improved |
| Add soft voting over predictions, increasing votes from N=1 to N=3 (ablation) | Ours on miniImagenet · miniImagenet 5-Way 1-Shot | accuracy: 2% to 3% | Improved |
Closest papers by meaning
- Variational Patch Gating for Training-Free Few-Shot Classification (ECCV 2026 · similarity 0.87)
- A Baseline for Few-Shot Image Classification (ICLR 2020 · similarity 0.87)
- Rethinking Generalization in Few-Shot Classification (NeurIPS 2022 · similarity 0.87)
- Pushing the Limits of Simple Pipelines for Few-Shot Learning: External Data and Fine-Tuning Make a Difference (CVPR 2022 · similarity 0.87)
- Pushing the Limit of Fine-Tuning for Few-Shot Learning: Where Feature Reusing Meets Cross-Scale Attention (AAAI 2024 · similarity 0.86)
- Provably Improving Generalization of Few-shot models with Synthetic Data (ICML 2025 · 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.
